View Full Version : LAV Filters - DirectShow Media Splitter and Decoders


Pages : 1 2 3 4 5 6 7 8 [9] 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

hissatsu
11th January 2012, 00:21
I'm having an issue with seamless blu-ray playback, using every version of LAV Filters after 0.39. When it switches from one m2ts in the mpls playlist to the next, the screen goes black, but audio continues playing. To get the video back, I have to seek. If I skip chapters forward or back to the point where the m2ts starts, this does not occur; it only happens during normal playback.

I'm using MPC-HC with MadVR. I've tried every version of LAV Filters through 0.44, and I've tried every version of MadVR for the past few months and many builds of MPC-HC, along with other renderers (EVR-CP, Haali), and no setting makes any difference other than going back to LAV Filters 0.39 or earlier. I'd really like to use a newer version, but something appears to have changed with the splitter after that point.

Andy o
11th January 2012, 01:00
yes my cpu can handle xvid playback fine, i just thought i would test it to see if it works. i though the lavcuvid had mpeg4-asp hardware decode working.

could be my graphics driver, im using 275.33. may need to update not sure.

The old LAV CUVID used to, but it was dropped, so don't bother trying other drivers if it's just to test this.

blexley
11th January 2012, 01:15
Newbie question

Is it best to use MPC-HC with LAV Filters. ?

rica
11th January 2012, 02:13
For video: If you have an nVidia; yes. However, if you have an Ati no; since no DXVA support exists on LavVideo yet.
For audio: If you pass-through LPCM or pass-through decoded HD audio; yes. No cons, but pros, I mean.
As for splitter: My favorite, (And don't forget; this project is based on the splitter, mainly.)

shimaflarex
11th January 2012, 03:48
I've noticed more glitched audio issues due to high CPU usage on LAV Splitter than with Haali.
Is there a way to change the priorities? Maybe lowering the video decoder priority?
Using Reclock with WASAPI.

hubblec4
11th January 2012, 05:15
...
As for splitter: My favorite, (And don't forget; this project is based on the splitter, mainly.)

i like to see the enhancement of the splitter but i cant use it at the moment, cause the splitter doesent support ordered chapters.
its on the todolist but with Low prio.

i hope Nev will implement this feature soon.

the tray icon feature is on the todolist too.

blexley
11th January 2012, 06:45
For video: If you have an nVidia; yes. However, if you have an Ati no; since no DXVA support exists on LavVideo yet.
For audio: If you pass-through LPCM or pass-through decoded HD audio; yes. No cons, but pros, I mean.
As for splitter: My favorite, (And don't forget; this project is based on the splitter, mainly.)

Thanks

It's only for an Nvidia card on a desktop computer with stereo speakers but i wasn't sure if MPC-HC was best or another player such as MSplayer.

Mangix
11th January 2012, 08:54
Newbie question

Is it best to use MPC-HC with LAV Filters. ?

i do and i love it. just make sure to disable all the internal filters or just disable the ones you don't want.

blexley
11th January 2012, 11:11
Thanks Mangix :)

Guys i've found a newbies MPC-HC LAV Filters install guide with pictures.

If i paste it up can you check through it to see if it is accurate to follow as i'm unsure about it suggesting Haali’s Media Splitter instead of LAV Splitter and such.Plus it seems biased towards HD content and i wish to be able to play as many formats and containers of SD and HD as possible.

Thanks.

blexley
11th January 2012, 11:11
Setup guide for MPC-HC + madVR

This method will result in significantly more image quality and performance than using CCCP (or anything ffdshow-based, really), so if you’re struggling to play back Hi10P anime, or even regular 1080p, consider trying this. MadVR is updated constantly and uses a very new version of libav for decoding, giving it much more speed than the ridiculously outdated ffdshow-tryouts project, as well as the even more ridiculously CCCP codec pack that is based on it.

Step Zero: Prerequisites

1. Uninstall ALL instances of MPC-HC, CCCP, ffdshow, madVR, Haali, K-lite, CoreAVC etc. you might have on your system.This is a clean guide that requires no other prerequisites to function, and ideally should have none installed either.

Step One: Downloading and Installation

1. Download and install the latest version of MPC-HC (you want the .exe version for an installer) Note: Get the 32-bit version ONLY! The 64-bit version is incompatible with madVR and will not function at all.
http://mpc-hc.sourceforge.net/download-media-player-classic-hc.html

2. Install the LAV Filters. These are necessary for decoding video and audio. Take care while installing to deselect the LAV Splitter:

http://i.lulzimg.com/90f8afc109.png

3. Obtain and install the latest version of Haali’s Media Splitter as LAV Splitter v0.42 still doesn’t support segment linking yet, so playing series releases that have OP/ED as seperate video files won’t playback in each of the episodes.
http://haali.su/mkv/MatroskaSplitter.exe

4. Note: You can skip this step if your PC isn’t very powerful!
Download the latest version of madVR and extract it anywhere you like. Run install.bat inside the folder. DO NOT DELETE THE FOLDER AFTERWARDS, madVR lives inside it and does not copy itself anywhere else. Move it somewhere sensible before installing, for example C:\Program Files\madVR.
Note: Make sure you’re logged in as an Administrator when running this! Do not right click and “run as admin”, log in as admin and run it normally!
http://madshi.net/madVR.zip

Step Two: Configuration

1. Open up MPC-HC’s options menu. Under “Internal Filters”, disable everything. You can leave some of the ones on the left active, but make sure you’ve disabled all of the ones not selected here:

http://i.lulzimg.com/5c1da2c562.png

2. Under Playback, enable “Auto-load subtitles”

http://i.lulzimg.com/826f2efd95.png

3. Under Subtitles, make sure “Allow animation when buffering” is enabled, and “Maximum texture resolution” is set to “Desktop”:

http://i.lulzimg.com/1ebc2b9564.png

4. Under Output, choose “madVR” as renderer.
Note: If you skipped madVR, or you’re experiencing lagging, frame drops or poor performance in general, set this to “EVR Custom Pres.”, “Haali Renderer” or “VMR-9 (renderless)” instead – try them in that order until you find one that works.

http://i.lulzimg.com/9d3f945def.png

5. Make sure the LAV Video is selected as default decoder. If in doubt, go to external filters, choose “Add Filter”, select LAV Video Decoder and switch it to “Prefer”:

http://i.lulzimg.com/8972c925d3.png

6. After making the above changes, restart MPC-HC and play back any file. You should notice the madVR icon in the system tray. Right click this to access the settings:

http://i.lulzimg.com/90038636ef.png

Once inside, disable the “fullscreen exclusive mode” under Rendering -> General Settings. I recommend leaving this off unless you are having playback problems, because it prevents you from taking screenshots and makes the transition to fullscreen very ugly (It also messes up MPC-HC’s interface):

http://i.lulzimg.com/fd74b8ba08.png

Step Three: Confirmation

1. Play back a Hi10P file and use Ctrl+J to enter the OSD: (If it doesn’t show up or looks completely different, doublecheck output configuration)

Correct

http://i.lulzimg.com/c22162c90a.png

Incorrect

http://i.lulzimg.com/c35c898b03.png

2. Finally, check the “Filters” list (in the right click menu of MPC-HC during playback) and confirm that LAV Video is being used:

http://i.lulzimg.com/1fbec71e48.png


Bonus: Adding ffdshow (Optional)

Adding ffdshow will allow you to use its audio decoder (for filters + a more customizable mixer), as well as using it for formats other than AVC1 (H.264), VC1 and MPEG-2. Especially for older 480p and worse content (which isn’t likely to be encoded in AVC1), you can use the ffdshow deblocking or debanding filters to attempt to improve image quality this way.

1. Install the latest ffdshow-tryouts.
http://sourceforge.net/projects/ffdshow-tryout/files/SVN%20builds%20by%20clsid/generic%20builds/

2. Open ffdshow’s settings dialog (Start -> All Programs -> ffdshow -> Video decoder configuration), under “DirectShow control” set the Merit to normal:

http://i.lulzimg.com/b218361575.png


3. Go back to the internal filters menu of MPC-HC and disable all of the remaining transform filters:

http://i.lulzimg.com/966a50142c.png

4. Double check Step Three to make sure ffdshow doesn’t touch your H.264 video.

Superb
11th January 2012, 13:29
What a crappy guide.
Why use Haali and not LAV Splitter?
Why not install the x64 filters?
Why leave MPC-HC's internal source filters on? (LAV Splitter does a better job)
Why bother w/ ffdshow when you have LAV Video/Audio installed? For the audio mixer? That's it? For video post processing? Newbies don't need it.
Why turn off madVR automatic exclusive full screen mode?

SamuriHL
11th January 2012, 13:48
What the hell was that? Other than completely unnecessary and annoying to those of us on a mobile device. Thanks SO much for that. :rolleyes:

Sent from my Xoom using Tapatalk

blexley
11th January 2012, 13:56
Thanks for your constructive posts on what a newbie needs to do as a step by step guide of what to adjust and alter when using MPC-HC with LAV Filters.

I'm sure every new user now has better clarification by your comment.

Superb
11th January 2012, 14:08
Instead of being sarcastic and following guides w/o knowing what each step does, simply read about the features of each project and its options. All information is available online.
There's also a "newbies" forum: http://forum.doom9.org/forumdisplay.php?f=6

SamuriHL
11th January 2012, 14:12
A newbie simply needs to ask and will get help. Spamming the forum with a half a page of large images simply to ask a question is seriously unnecessary. Especially since most of it had nothing to do with lav filters in the first place. To add lav filters to mpc-hc all you need to do is run the lav filters installer, go into options in mpc-hc, disable all the internal filters, go to external filters, add lav splitter, lav video, and lav audio. Set to preferred. Done. No massive post of images necessary.

Sent from my Xoom using Tapatalk

blexley
11th January 2012, 14:16
How else am i supposed to ask if the guide is accurate without pasting it up for it to be read through.

Telepathy. ?

Because pasting it up got the result i was looking for that it is not accurate , though the abbrasive attitude towards a newbie that doesn't have a clue could have been left out to start with.

A newbie simply needs to ask and will get help. Spamming the forum with a half a page of large images simply to ask a question is seriously unnecessary. Especially since most of it had nothing to do with lav filters in the first place. To add lav filters to mpc-hc all you need to do is run the lav filters installer, go into options in mpc-hc, disable all the internal filters, go to external filters, add lav splitter, lav video, and lav audio. Set to preferred. Done. No massive post of images necessary.

Sent from my Xoom using Tapatalk

And if that is all you need to do why is there page after page of also running MadVR and FFDshow and AC3filter. ?

It would be nice if nevcairiel added a step by step guide to the first post or to his webpage.


.

SamuriHL
11th January 2012, 14:28
How else am i supposed to ask if the guide is accurate without pasting it up for it to be read through.

Telepathy. ?

Because pasting it up got the result i was looking for that it is not accurate , though the abbrasive attitude towards a newbie that doesn't have a clue could have been left out to start with.



And if that is all you need to do why is there page after page of also running MadVR and FFDshow and AC3filter. ?

It would be nice if nevcairiel added a step by step guide to the first post or to his webpage.


.

Um by linking to it instead of reposting the whole thing??? Madvr is a renderer so that is fine to add. And not hard. Run the madvr install script and select it in the mpc-hc options. Ffdshow and ac3filter are most definitely not needed unless you're doing post processing...and that is not a newbie thing to be doing. See you're getting information simply by discussion rather than half a page of images. That was my point. You started with the premise that you needed a pretty guide but that isn't really needed. A simple "I'm new and need help setting up mpc-hc with lav filters" would have been sufficient. :)

Sent from my Xoom using Tapatalk

fastplayer
11th January 2012, 14:30
How else am i supposed to ask if the guide is accurate without pasting it up for it to be read through.
Where's the link to the source? Do you have permission of the original author of the guide? You're copying and pasting content without naming the source and giving credit. :rolleyes:

sneaker_ger
11th January 2012, 14:33
Half of that is unnecessary. Setting LAV video (and ffdshow) to preferred in the external filters options is sufficient, lowering the merit of ffdshow and deactivating the internal transform filters is not required. What is important though is deactivating the transform filters for e.g. Matroska. Also it hardly offers "significantly more image quality" than CCCP. If you want improved subtitle performance I'd rather recommend one of the new vsfilter forks like xy-vsfilter (no DS libass filter yet, sadly, only in mplayer2, VLC and whatnot).
Also why did you go out of your way to say that madVR has up-to-date libav decoders, when your guide explains using LAV Video for that?

blexley
11th January 2012, 14:36
Um by linking to it instead of reposting the whole thing??? Madvr is a renderer so that is fine to add. And not hard. Run the madvr install script and select it in the mpc-hc options. Ffdshow and ac3filter are most definitely not needed unless you're doing post processing...and that is not a newbie thing to be doing. See you're getting information simply by discussion rather than half a page of images. That was my point. You started with the premise that you needed a pretty guide but that isn't really needed. A simple "I'm new and need help setting up mpc-hc with lav filters" would have been sufficient. :)

Sent from my Xoom using Tapatalk

Thanks i now have a better understanding what i need to do. :)

I didn't link to the orginal guide as it is up on a couple of bitorrent sites that shares Hi10P anime and assumed links to a forum sharing would be against the rules.

Thing that is confusing me now is if i add the MadVR render doesn't that cancel out LAV filter. ?


Half of that is unnecessary. Setting LAV video (and ffdshow) to preferred in the external filters options is sufficient, lowering the merit of ffdshow and deactivating the internal transform filters is not required. What is important though is deactivating the transform filters for e.g. Matroska. Also it hardly offers "significantly more image quality" than CCCP. If you want improved subtitle performance I'd rather recommend one of the new vsfilter forks like xy-vsfilter (no DS libass filter yet, sadly, only in mplayer2, VLC and whatnot).
Also why did you go out of your way to say that madVR has up-to-date libav decoders, when your guide explains using LAV Video for that?

Thanks sneaker_ger

I'm not sure why they said that about MadVR as i just copied and pasted the only step by step guide i could find at google.

All i want to do is to be able to install a mediaplyer with LAV filters to be able to play as many formats and containers as possible without all these Klite all in one codec type solutions that makes a right mess.

.

SamuriHL
11th January 2012, 14:39
Thanks i now have a better understanding what i need to do. :)

I didn't link to the orginal guide as it is up on a couple of bitorrent sites that shares Hi10P anime and assumed links to a forum sharing woud be against the rules.

Thing that is confusing me now is if i add the MadVR render doesn't that cancel out LAV filter. ?

No. Lav filters don't include a renderer. The choice of madvr is a good one and what most of us use. Lav consists of a splitter and a decoder for video and a decoder for audio. You still need renderers for video and audio and that's where madvr fits in (for video). Audio rendering has some options, as well. Some like reclock so you can use wasapi exclusive mode and bypass the windows audio stuff.

Sent from my Xoom using Tapatalk

blexley
11th January 2012, 14:47
I keep finding old posts that say the brightness control doesn't work in MadVR is this still the case. ?

sneaker_ger
11th January 2012, 14:49
It does not have a brightness control in the first place. But direct your madVR questions to the madVR thread.

blexley
11th January 2012, 15:21
There are 500 posts in this 400 page thread asking and talking about using MadVR with LAV filters but if i ask i need to go to another thread. ?

sneaker_ger
11th January 2012, 15:29
Many of them are about the interaction of those two, many others are off-topic as well. Yours was only about madVR and had no connection to LAV Filters.

nevcairiel
11th January 2012, 19:14
Hi,

i have a testing request to people using ATI cards.

http://files.1f0.de/lavf/LAVFilters-0.44-dxva2-debug.zip

Install that, set the configuration to use the DXVA2 decoder, and play a H264 file for a few seconds (preferably a 1920x1080 file to maximize the load)
It will create a LAVVideo.txt log file on your desktop. Please paste that file to pastebin.com and link it here.

Note that this version is probably unstable, definitely slow, and still a work in progress. Just want to know if the basic idea is going in the right direction.

Thanks!

fastplayer
11th January 2012, 19:18
Log for Radeon HD4670:
http://pastebin.com/Xnmfk3fg

Log with 1920x1088 source:
http://pastebin.com/HvyJm9st

fairchild
11th January 2012, 19:33
HD 5830 running latest 11.12 drivers

http://pastebin.com/AxLHTTp2

First impression, everything seems to be running smoothly and I get no frame drops/delays/glitches with MadVR. I verified that the GPU is changing to 400/900 gpu/mem speeds so that is the correct DXVA/2 UVD speeds. If you can get this running this would be E-mazing.

If you need any more logs please let me know. I also have a 5450 I could test on if you need.

HeadlessCow
11th January 2012, 19:46
What a crappy guide.
Why use Haali and not LAV Splitter?
Why not install the x64 filters?
Why leave MPC-HC's internal source filters on? (LAV Splitter does a better job)
Why bother w/ ffdshow when you have LAV Video/Audio installed? For the audio mixer? That's it? For video post processing? Newbies don't need it.
Why turn off madVR automatic exclusive full screen mode?

Not that I'm defending the copy/pasting of the guide here, but every one of these was explicitly answered in the text.

1. "LAV Splitter v0.42 still doesn’t support segment linking yet, so playing series releases that have OP/ED as seperate video files won’t playback in each of the episodes"
2. "The 64-bit version is incompatible with madVR"
3. "Under “Internal Filters”, disable everything. You can leave some of the ones on the left active"
4. "Bonus: Adding ffdshow (Optional) ... Adding ffdshow will allow you to use its audio decoder (for filters + a more customizable mixer), ... Especially for older 480p and worse content (which isn’t likely to be encoded in AVC1), you can use the ffdshow deblocking or debanding filters to attempt to improve image quality this way."
5. "I recommend leaving this off unless you are having playback problems, because it prevents you from taking screenshots and makes the transition to fullscreen very ugly (It also messes up MPC-HC’s interface)"

You obviously read enough to make a snarky, unhelpful reply... would it have been much more work to try to comprehend anything that was written there instead of just being a jerk?

hoborg
11th January 2012, 19:50
Hi,

i have a testing request to people using ATI cards.

http://files.1f0.de/lavf/LAVFilters-0.44-dxva2-debug.zip

Install that, set the configuration to use the DXVA2 decoder, and play a H264 file for a few seconds (preferably a 1920x1080 file to maximize the load)
It will create a LAVVideo.txt log file on your desktop. Please paste that file to pastebin.com and link it here.

Note that this version is probably unstable, definitely slow, and still a work in progress. Just want to know if the basic idea is going in the right direction.

Thanks!
Hi.
Not sure why, but when i try to register LAVSplitter.ax, i got the failed to load module error.
Registering video/audio decoders working, but when i try to play a video by using DXVA2, graphstudio crash.
.44 working just fine.

sneaker_ger
11th January 2012, 19:53
HD 5850 Catalyst 11.12

http://pastebin.com/PVjAvU3u

Running with madVR, queue filling up rather slowly, some drops depending on file, no artifacts.

wanezhiling
11th January 2012, 20:13
Mobility HD4650, Catalyst 12.1a preview, win7 x86
http://pastebin.com/a4eE3UX6

:)I'm so excited for this.

VipZ
11th January 2012, 20:31
Tested DXVA on my shiny new 7970, works very well. Good seeks, no corruption or dropped frames in the few 1080p files of mine.

I have also tested some 1080i streams and they working very well.

One weirdness, with NV12 CPU use is at 0-1%, with RGB32 it goes to 15-20% on interlaced.

http://pastebin.com/mYmTbaXQ

Let me know if you want any other testing done.

nevcairiel
11th January 2012, 20:33
The high CPU load on RGB is because its a debug build, its very slow. :p

I have enough tests btw, just wanted to see if it was working smoothly at all.
Looking at those numbers, AMD definitely did something right with the 7970, the locking and copying time are nicely low. :)

SamuriHL
11th January 2012, 20:34
Looks like everyone is doing all the testing already. Sweet. I have a 5870 I can test on later if you need more feedback. I'm running 12.1a preview, as well.

VipZ
11th January 2012, 20:36
the clip you asked us to test last time, Test_clip_avc.1080i59.94.ac3.5.1.mkv also works without any issue with this build.

EDIT: MPEG2 and VC1 1080p also seem to work without issue and 10bit h264 crashes.


Looking at those numbers, AMD definitely did something right with the 7970, the locking and copying time are nicely low. :)

I was hoping GCN was going to do good with this :)

Andy o
11th January 2012, 21:24
Man I keep reading good things about GCN, can't wait to get me a 7800 series.

hoborg
11th January 2012, 21:32
Hi.
Not sure why, but when i try to register LAVSplitter.ax, i got the failed to load module error.

No one have this problem? What am i missing/doing wrong? :/
http://hobring.esero.net/saf/lavf/snap1.png

adam777
11th January 2012, 21:40
No one have this problem? What am i missing/doing wrong? :/
http://hobring.esero.net/saf/lavf/snap1.png

Same here.

SamuriHL
11th January 2012, 21:44
It's a debug build so I'm guessing you don't have a debug version of the runtime installed. Just a guess, though.

hoborg
11th January 2012, 21:47
It's a debug build so I'm guessing you don't have a debug version of the runtime installed. Just a guess, though.
Link?

SamuriHL
11th January 2012, 21:53
Don't have one offhand. I have MSVC installed so, I kinda don't need them.:)

Snowknight26
11th January 2012, 22:16
There seems to be an exception in avcodec-lav-53.dll when using the filter without registering it: http://pastebin.com/DQGu1whV

When it does work though: http://pastebin.com/qyXdgLem

Edit: On a side note, the tool tip for multi-threaded decoding is missing punctuation in codec names here and there (mostly periods and dashes).

noee
11th January 2012, 22:58
...
Looking at those numbers, AMD definitely did something right with the 7970, the locking and copying time are nicely low. :)

Cool. But is it the card or the driver?

Boltron
12th January 2012, 00:00
Two samples, both high bit rate bluray rips.

http://pastebin.com/vHqjsB6F
http://pastebin.com/hjm00gLP

Windows 7 x64 (using 32 bit versions for all video stuff)

1920x1080, 5670, ATI 12.1

MPC-HC, madVR

PPPPark
12th January 2012, 00:33
Hey~
I have some DVD image(ISO,img), LAV can play, but the image has "JACKET_P" folder LAV can not play, you can try, hope can be resolved.
Ordinary DVD only "AUDIO_TS" and "VIDEO_TS"

madshi
12th January 2012, 02:40
Tested DXVA on my shiny new 7970
Cool! Would you mind running this test, too?

http://madshi.net/madNV12Test.zip

SamuriHL
12th January 2012, 02:44
Knock it off....I can't buy a new card right now! :p

apgood
12th January 2012, 04:18
What version intel driver are people using with QS?

I've tried both 2509 & 2559 and they both stutter in hatdware and software mode. The only one that seems to work at all is 2372 in software mode only. While hardware mode on 2372 does that wierd frame flicker.

Specs are: i3 2100 on asrock z68m/usb3 using win7 64bit. With dsplayer in xbmc (hardware & software) and mc17 (software only) to verify it's not just the player.

skingery
12th January 2012, 06:41
What version intel driver are people using with QS?

I've tried both 2509 & 2559 and they both stutter in hatdware and software mode. The only one that seems to work at all is 2372 in software mode only. While hardware mode on 2372 does that wierd frame flicker.

Specs are: i3 2100 on asrock z68m/usb3 using win7 64bit. With dsplayer in xbmc (hardware & software) and mc17 (software only) to verify it's not just the player.
I have this same issue. Currently running the 2509 driver. i3 2100T. In my case I am running Win 7 32bit and mpc-hc. Works fine in software mode but serious stuttering/screen flashing in hardware mode.

wanezhiling
12th January 2012, 07:23
1. As a ATI user, Thx nev first.:)

2. I simply tested all 1080p and 1080i I have(BD files .m2ts).

3. My M hd4650(UVD2.0) supports H.264_VLD, VC1(WMV3)_VLD, MPEG2_IDCT, the project for ATI only works on H.264 and VC1(WMV3), MPEG2 failed, just like MPC-HC whose DXVA is only based on VLD(Variable Length Decoding).:)

4. Compared with H.264, VC1 files perform poor,not very smooth(Same is CUVID).. Yep, 1080p's playback is better than 1080i too.

5. Everything(except VC1) seems to be running smoothly, I only found one 1080i/H.264 clip(Image is broken (http://img165.poco.cn/mypoco/myphoto/20120112/13/5905971020120112131401080.jpg)) .Maybe you can try it here (http://www.gokuai.com/f/2igXjZBs6v96Rbt0).:)



Mobility HD4650, Catalyst 12.1a preview, win7 x86
H.264 1080I (http://pastebin.com/FXNcJbmw)(Image broken)

VC1 (http://pastebin.com/g5pJk1L3)

WMV3 (http://pastebin.com/1rbzyeLk)

MPEG2 (http://pastebin.com/nqG5g96Q)

nevcairiel
12th January 2012, 07:46
From all these numbers it looks like anything up to 60i will most likely play properly, however 60p might be too slow, somewhat similar to my own results.

We'll see if i can find a better way to lock and retrieve the video frame. I didn't even optimize the memory copying yet because all the time is spent in LockRect, not copying the data.
I must be doing somethign wrong still, because even on my NVIDIA the performance isn't close to what i would expect. :)

Does anyone have any experience with copying the data from the GPU?
Right now i'm just LockRect'ing the DXVA surface and copying the data over - but the LockRect can quite frequently take a long time. I read that an alternative is using GetRenderTargetData to copy the data to system memory, and because that method is async in the background, i could potentially gain some benefits from it?

PS:
Partial acceleration is annoyingly complicated to support for only a small gain, which is why its not supported and never will be. Even Microsoft got the hint, and any new formats don't have a partial acceleration type (and won't have).

Superb
12th January 2012, 10:13
Not that I'm defending the copy/pasting of the guide here, but every one of these was explicitly answered in the text.

1. "LAV Splitter v0.42 still doesn’t support segment linking yet, so playing series releases that have OP/ED as seperate video files won’t playback in each of the episodes"
2. "The 64-bit version is incompatible with madVR"
3. "Under “Internal Filters”, disable everything. You can leave some of the ones on the left active"
4. "Bonus: Adding ffdshow (Optional) ... Adding ffdshow will allow you to use its audio decoder (for filters + a more customizable mixer), ... Especially for older 480p and worse content (which isn’t likely to be encoded in AVC1), you can use the ffdshow deblocking or debanding filters to attempt to improve image quality this way."
5. "I recommend leaving this off unless you are having playback problems, because it prevents you from taking screenshots and makes the transition to fullscreen very ugly (It also messes up MPC-HC’s interface)"

You obviously read enough to make a snarky, unhelpful reply... would it have been much more work to try to comprehend anything that was written there instead of just being a jerk?1. 99% of the releases don't use this feature. And LAV Splitter supports MANY other containers much much better than Haali or MPC-HC's filters.
2. The 64 bit version of MPC-HC. Yes. We all know that 64bit player cannot run a 32bit renderer. I was asking: "why not install the 64bits components of LAV Filters?" (so that they will be available for other [64bit] programs)
3. But why? LAV Splitter does a much better job.
4. Well, wasn't this a "newbie" guide? Adding ffdshow to the filter chain would be be useless for 99.999% of the users, since they don't even know what video post processing is or how to setup it properly.
5. It doesn't mess up the MPC-HC GUI. The move to exclusive full screen is just fine (does the quick "flash" really bother anyone?). They must have forgotten exclusive mode results in a much smoother playback (which is one of the best features of madVR).

blexley
12th January 2012, 10:13
LOL :p

I spent a week going around in circles trying to find an easy step by step install guide and i've just found it at the very page i first started at, nevcairiel own HomePage. :D :D :D

http://1f0.de/lav-cuvid/guide/

nevcairiel is there any chance you could make this guide more obvious in the first post of this thread and on the start of your webpage so newbies and people with low computer skills such as myself can avoid the obnoxious comments from users when asking for guidance as if LAV filters is some how for the elite and you need to search and read about codecs and splitters before you can use it.As all we need to know is what to adjust and tick/untick so we can just get on and use it and your guide does that prefectly.

Plus the more accessible it is and the quicker we can get going the quicker newbies can hit the paypal donate. :)

Thanks

madshi
12th January 2012, 10:15
We'll see if i can find a better way to lock and retrieve the video frame. I didn't even optimize the memory copying yet because all the time is spent in LockRect, not copying the data.
I must be doing somethign wrong still, because even on my NVIDIA the performance isn't close to what i would expect. :)

Does anyone have any experience with copying the data from the GPU?
Right now i'm just LockRect'ing the DXVA surface and copying the data over - but the LockRect can quite frequently take a long time. I read that an alternative is using GetRenderTargetData to copy the data to system memory, and because that method is async in the background, i could potentially gain some benefits from it?
Using LockRect is the same method I'm using in madNV12Test. GetRenderTargetData() is what madNV12Test lists as "trick download" and it usually fails for NV12 surfaces for me simply because it would require me to create an NV12 surface in system memory and neither CreateOffscreenPlainSurface() nor IDirectXVideoAcceleratorService::CreateSurface() seem to be willing to do that on my PC. FWIW, with NVidia I'm getting higher download speeds by allocating the NV12 decoding surface with IDirectXVideoAcceleratorService::CreateSurface() instead of CreateOffscreenPlainSurface().

nevcairiel
12th January 2012, 10:21
FWIW, with NVidia I'm getting higher download speeds by allocating the NV12 decoding surface with IDirectXVideoAcceleratorService::CreateSurface() instead of CreateOffscreenPlainSurface().

Thats what i'm using, well technically IDirectXVideoDecoderService, but that one just inherits from IDirectXVideoAccelerationService. The copy speed is fine, its just the LockRect operation that takes its time.
The weird thing is that the locking speed is not consistent, its going like 20ms, 1ms, 1ms, 20ms, 1ms, 1ms .. as if its still decoding a frame there and waiting.

I'll see about running more tests later.

cengizhan
12th January 2012, 10:29
nevcairiel,

i have an ATI 6850. new dxva works ok with mpc-hc but there is a problem with smartdvb:

http://postimage.org/image/93aj13v8r/

nevcairiel
12th January 2012, 10:36
nevcairiel is there any chance you could make this guide more obvious in the first post of this thread

No, because the guide is outdated, and LAV CUVID has been deprecated.

blexley
12th January 2012, 10:45
No, because the guide is outdated, and LAV CUVID has been deprecated.

If we are not supposed to use that guide then will you be making a new one for MPC-HC and LAV filters when you have time. ?

Thanks

nevcairiel
12th January 2012, 11:03
If we are not supposed to use that guide then will you be making a new one for MPC-HC and LAV filters when you have time. ?


I write code, not guides.

blexley
12th January 2012, 11:09
Then why have the guide on your webpage if it's beneath you as a coder. ?

What is it with the abrasive rude attitude towards newbies in this thread , do you wan't peple to use your filter and to donate money or not. ?

madshi
12th January 2012, 11:22
@blexley, the reason people are being rude to you has nothing to do with you being a newbie. Rudeness in these forums is usually just a reaction to annoying posts/behaviour. That said, I don't see nevcairiel's post as rude at all. He's just clearly saying that he's not interested in writing guides. What's rude about that? Nothing. He's just stating his preferences. However, insinuating that nevcairiel would feel that something were "beneath him" is quite rude in itself. So from my point of view, you're the one being rude here. And furthermore you're being rude to a developer who spends LOADS of his time to create free software for you. You should reconsider your style of posting if you don't want to end up on the ignore lists of all the most active developers in this forum.

blexley
12th January 2012, 11:43
What you have just posted is quite frankly ridiculous and illogical as how is a newbie supposed to enjoy all this so called time spent creating free software if there isn't a guide on how to use it correctly let alone donate money. ?

Are people supposed to be grateful and donate money for something they haven't even used yet and can't use until they know what exact settings to alter. ?


.

Koepi
12th January 2012, 11:57
blexley,

if a coder just wants to code, just let him do that. Website maintenance isn't real fun either, and as a starting point the old guide may be still of use for someone.

As you are that demanding - why don't _you_, who got 'a free lunch', take the time and write a guide and offer that?

Demanding is simple (and annoying, I can tell). Doing the real work however, is much more satisfying for everyone. So instead of just demand updates of documentation and website, write that guide. Which will bring you in a much better position to demand anything then, by the way.

nevcairiel
12th January 2012, 11:58
You seem to think we all do this so people donate money, as you keep bringing up that point. But you're dead wrong.

I do this because i consider working on it fun.
Because of that, i only do things that are fun for me, and writing guides isn't one of these things, and dealing with obnoxious people isn't fun either - so i just ignore them.

I only put up the donation button because people asked if/how they could donate, and who am i to say no if they offer. :)
It always feels good getting a donation, as a recognition for my work, however i would still continue if i never got any donations.

There are plenty guides out there, and since using it "correctly" is rather subjective and depends on your goals, just pick one that suits your needs and go with it.
Just don't copy/paste some random guide into this thread or ask people to hand-hold you through the setup, this isn't the place for that.

Also, if you're such a "newbie", then use a codec pack, there are plenty of them that come with my filters, pre-configured with all the settings you need.

blexley
12th January 2012, 12:09
blexley,

if a coder just wants to code, just let him do that. Website maintenance isn't real fun either, and as a starting point the old guide may be still of use for someone.

As you are that demanding - why don't _you_, who got 'a free lunch', take the time and write a guide and offer that?

Demanding is simple (and annoying, I can tell). Doing the real work however, is much more satisfying for everyone. So instead of just demand updates of documentation and website, write that guide. Which will bring you in a much better position to demand anything then, by the way.

Is this some sort of joke. ?

How is somebody with low computer skills who can't work out what settings are supposed to altered with MPC-HC and if i'm supposed to run MadVR aswell write the very step by step guide that i'm asking for. ?



There are plenty guides out there, and since using it "correctly" is rather subjective and depends on your goals, just pick one that suits your needs and go with it. Just don't copy/paste some random guide into this thread or ask people to hand-hold you through the setup, this isn't the place for that.

Also, if you're such a "newbie", then use a codec pack, there are plenty of them that come with my filters, pre-configured with all the settings you need.

So you say use a guide without actually pointing to one and the one i do paste up and ask if it is accurate and can be used is hand holding. ?

The ridiculous logic and reasoning and contradictions to these posts instead of just helping is beyond belief.

entrecour
12th January 2012, 12:20
blexley, Here is IMHO a pretty comprehensive guide

Watching H.264 videos using Compute Unified Device Architecture (CUDA) (http://imouto.my/watching-h264-videos-using-compute-unified-device-architecture-cuda/)

To simplify the setup further you can easily leave out halli, madflac, and ffdshow unless you have a specfic need for them.

blexley
12th January 2012, 12:24
Thanks entrecour :thanks:

You sir are a gent. :)

PeQuE
12th January 2012, 12:33
nev, how is lav video decoder designed to handle acceleration of vc-1 on Nvidia cards that support only VDPAU feature set A. My GTX 295 only supports Set A which according to wikipedia is as follows

Feature Set A
Complete acceleration for H.264
Partial acceleration for MPEG-1, MPEG-2, VC-1/WMV9

When I watch vc-1, with cuda enabled all is fine for the most part, until certain scenes where the image will appear blocky for a second or so.

Of course I would love to leave cuda on for vc1 so in my case should I manually disable it, or is there anything that can be done from the lav video side to allow me to use this "partial acceleration" and still see a valid image? Cheers

Its a bug in the driver, seems to have appeared a while ago. You could try different drivers, otherwise there isn't anything that can be done.

258.96 is the last ver to work right. And I think its actually hardware deinterlacing related. cause if i set CUVID to weave with newer drivers the blocking goes away.

It still is somehow related to partial acceleration, because it doesn't happen on cards with full acceleration capabilities. It might just be coincidence and depend on the series of the card, but somewhere is another factor in there. :)

I quote some messages that I consider have something to do with what I want to say...

Yesterday I upgraded my nVidia GT240 drivers from 258.96 to 285.62... Well, this drivers have broken in some way (framedrops, irregular fps and render times) the CUDA hardware acceleration of H264 interlaced material (HD livetv mainly).

- Interlaced MPEG2 is ok.
- Progressive H264 is ok.
- Switched Lav Video (0.44) to Cyberlink PowerDVD 11 H264 (so no CUDA), ok.
- Downgrade to Lav Video 0.43, still nok.
- Downgrade nVidia drivers to 280.26, ok (but I would say not as stable decoding as 258.96 anyway...)

ty, you're right. If anyones interested, heres some nvidia drivers I tried, showing vc1 decoding was broken somewhere between 280.36 and 285.27 on these cards at least. (I tried using the working cuda related dlls with the newer drivers but that doesn't seem to work very well :P )

Nvidia Driver - Cuda dll version - Status
285.62 - 8.17.12.8562 - broken
285.38 - 8.17.12.8538 - broken
285.27 - 8.17.12.8527 - broken
280.36 - 8.17.12.8036 - working
280.26 - 8.17.12.8026 - working

Definitely, 285.xx series drivers have some problems...

Nev, I don't ask you to do anything, as is clearly driver related. I only want to leave this comment in case anyone else finds the same, knows where's the problem.


bye!

madshi
12th January 2012, 12:45
Finally I can put this forum's ignore list feature to good use.

Alabanda
12th January 2012, 12:59
nevcairiel, you're awesome! Finally, the possibility of having quality hardware decoding in my HTPC-intended setup. I'm running an AMD Fusion board, an E350.

(pastebin was under heavy load) http://dl.dropbox.com/u/26229491/LAVVideo.txt

Again, thanks a lot!

SamuriHL
12th January 2012, 13:44
Finally I can put this forum's ignore list feature to good use.

Seriously unreal. I honestly don't understand people sometimes. Just know that the majority of us appreciate what you guys do.

Sent from my Xoom using Tapatalk

VipZ
12th January 2012, 13:48
Cool! Would you mind running this test, too?

http://madshi.net/madNV12Test.zip

Will do when I get home.

Bit OT, but may be useful for others to know. HDMI bit steaming on the 7970 negotiates pretty much instantly for all formats, where it used to take up to 3secs on my 5850. And on pausing it doesn't drop the stream immediately but keeps it showing as no channels on the AVR for around 2 secs before dropping.

STaRGaZeR
12th January 2012, 15:41
DXVA? Sounds nice :D

Let's hope drivers don't ruin the party.

CruNcher
12th January 2012, 15:50
Now a real DXVA (GPU independent, no frame copy) and DXVA support for MadVR is left ;)

goldie
12th January 2012, 16:38
Sorry for my late feedback. :p
LAVFilters 0.44 debug log for ATI Radeo HD 6850 with Catalyst 11.12 is here.
http://pastebin.com/RgdMrTcH

Execution environment:
Win 7 Pro 64bit
MPC-HC 1.5.3.3958 (32bit)
madVR 0.80

Playing a H264 L4.2 1080p60 clip was not so smooth confirmed,
but I'm so excited at the news about hw decoding for ATI users!
Good job, and :thanks: nev.

CruNcher
12th January 2012, 16:51
could you test vs the Performance of Cyberlink HAM it's the same :)

goldie
12th January 2012, 17:15
I had test Cyberlink HAM video decoder before,
but I think it always fallback to software decoding.
(I saw the CPU utilization didn't come down with HAM mode compared to software mode.)
So I uninstall it few months ago,
maybe I don't know how to set it up properly. :o

CruNcher
12th January 2012, 17:26
hmm does potplayers copy decoder works then ?

Not Ati but ;)

1080p @60 fps

IMSDK = 60 fps
Nev DXVA2 = 40 fps

Not that bad for the start :)

nevcairiel
12th January 2012, 17:30
Good News everyone!

I managed to increase the performance quite a bit, and its now close to the CUVID decoder on my system.
Just to figure out why one of the first frames is always corrupted now... :)

madshi
12th January 2012, 17:37
I managed to increase the performance quite a bit
How? (If I may ask...) :D

fastplayer
12th January 2012, 17:38
When it comes to DXVA, always assume it's the driver's fault! :D

nevcairiel
12th January 2012, 17:40
How? (If I may ask...) :D

I remembered something which i did for the CUVID decoder waaaay back.

I just store the surface in a queue, and only if the queue is full, i start processing them, one at a time. This gives the GPU time to finish rendering to the surface before i access it. A queue of 2 frames makes all the difference, just a tiny bit of delay to give the GPU some breathing room.
Speed is now 99% that of CUVID. :)

SamuriHL
12th January 2012, 17:46
I remembered something which i did for the CUVID decoder waaaay back.

I just store the surface in a queue, and only if the queue is full, i start processing them, one at a time. This gives the GPU time to finish rendering to the surface before i access it. A queue of 2 frames makes all the difference, just a tiny bit of delay to give the GPU some breathing room.
Speed is now 99% that of CUVID. :)

Well done!!

madshi
12th January 2012, 17:50
I remembered something which i did for the CUVID decoder waaaay back.

I just store the surface in a queue, and only if the queue is full, i start processing them, one at a time. This gives the GPU time to finish rendering to the surface before i access it. A queue of 2 frames makes all the difference, just a tiny bit of delay to give the GPU some breathing room.
Speed is now 99% that of CUVID. :)
So how many FPS do you with ATI cards? I'm wondering why I got so miserable results with my madNV12Test tool. All I did there was LockRect + memcpy in a loop, with no real decoding going on...

nevcairiel
12th January 2012, 17:51
So how many FPS do you with ATI cards? I'm wondering why I got so miserable results with my madNV12Test tool. All I did there was LockRect + memcpy in a loop, with no real decoding going on...

I have no ATI card, but when i post a first real test version, i'm sure people will be glad to benchmark it.

SamuriHL
12th January 2012, 17:54
I'll be able to test it later for sure. Very interested in this.

madshi
12th January 2012, 17:56
Ah sorry, misunderstood. Thought you had 99% of CUVID performance with ATI DXVA!

SamuriHL
12th January 2012, 17:57
He might. :D That's what we need to test. ;)

goldie
12th January 2012, 18:00
I'll be able to test it later for sure. Very interested in this.
+1 :thanks:

CruNcher
12th January 2012, 18:09
@ nev
ehh how does it compare with the IMSDK layer (Quicksync Decoder overhead) ? do you think it's not really needed anymore if this runs generically fine on every GPU :) or do you meant you currently reach the same performance as CUVID in speed (on your Nvidia VPx) which would be still less then with Intels MSDK DXVA2 implementation ?

nevcairiel
12th January 2012, 18:25
Intel has some issues with the "generic" DXVA2, i would have to figure out what settings exactly it needs to run properly - so instead just use the MSDK? :)

SamuriHL
12th January 2012, 18:27
Is this stuff checked in to the repository yet? if so I can start building it and take a look.

nevcairiel
12th January 2012, 18:36
Ok, here is a more polished version without debugging.

http://files.1f0.de/lavf/LAVFilters-0.44-dxva2-perftest.zip

It would be great if some people could benchmark this using following clip:

http://xhmikosr.1f0.de/samples/2160p/CrowdRun/CrowdRun_1080p50.x264.CRF23.mkv

I would recommend to benchmark with GraphStudio, mostly because its so easy. ;)
It would also be great if you could mention which kind of CPU you're running, so i know which kind of memory copy is being used.

As a reference, my NVIDIA with a VP4 decoder does around 73fps, both in DXVA2 and CUVID.

Disclaimer:
- This version is generally not all that much tested, and might blow up (and possibly take your PC with it)
- DXVA2 is a Vista/7 tech, don't expect it to work on XP
- Fallback to software decoding is still unfinished, and will most likely crash.
- Seeking in VC-1 still is somewhat rough, H264 seems to work better, however.
- VC-1 interlaced decoding is not yet implemented (if i can pull it off to finish it)

hoborg
12th January 2012, 18:48
Radeon HD 6750: (it stutter a little)
http://hobring.esero.net/saf/lavf/dxva2_perf.png

Sebastiii
12th January 2012, 18:48
Amazing :) Thanks

SamuriHL
12th January 2012, 18:59
@hoborg....what is that? And where can I find it? :)

nevcairiel
12th January 2012, 19:00
I would guess thats GraphStudio Next (http://code.google.com/p/graph-studio-next/)

CruNcher
12th January 2012, 19:00
Yup it stresses User as well as Kernel time a little more and thus isn't as efficient as the Quicksync Implementation on Intel Hardware but not that bad @ all for 1080p 60 fps the difference was 2%/4%/12%/17%/25% (includes render overhead) for Cyberlink and Arcsoft DXVA/CoreAVC DXVA/Intel(Egur/Nev) DXVA2/Generic DXVA2(Nev)/Libav(Nev) @ playback (not raw benchmark performance) :)

sneaker_ger
12th January 2012, 19:06
Radeon HD 5850
~50.3 fps

hoborg
12th January 2012, 19:08
@hoborg....what is that? And where can I find it? :)
Yes it is graphstudionext - View -> Decoder performance.

SamuriHL
12th January 2012, 19:10
Brilliant. Thanks for that. 51-52fps on average with my Quad Core 8200, 5870 12.1a preview, W7 SP1 X64, using 32 bit versions of everything.

nevcairiel
12th January 2012, 19:15
Too bad, i would've hoped for faster.

Here is a second try:
http://files.1f0.de/lavf/LAVFilters-0.44-dxva2-perftest2.zip

You need the first version, and just replace the .ax file

I increased the number of buffers from 2 to 8, apparently on my Intel GPU this gives me a significant performance boost on some samples, so who knows, maybe it does something for ATI.
Otherwise, i'm not sure i know what else to try without an ATI card at hand - maybe its just everything we can get out of them...

HeadlessCow
12th January 2012, 19:18
1. 99% of the releases don't use this feature. And LAV Splitter supports MANY other containers much much better than Haali or MPC-HC's filters.
2. The 64 bit version of MPC-HC. Yes. We all know that 64bit player cannot run a 32bit renderer. I was asking: "why not install the 64bits components of LAV Filters?" (so that they will be available for other [64bit] programs)
3. But why? LAV Splitter does a much better job.
4. Well, wasn't this a "newbie" guide? Adding ffdshow to the filter chain would be be useless for 99.999% of the users, since they don't even know what video post processing is or how to setup it properly.
5. It doesn't mess up the MPC-HC GUI. The move to exclusive full screen is just fine (does the quick "flash" really bother anyone?). They must have forgotten exclusive mode results in a much smoother playback (which is one of the best features of madVR).

1. (As pointed out by blexley later) Because it came from a page about anime fansubs where it is (relatively) common.
2. Because it's a guide about getting MPC-HC working with madvr... The 64-bit stuff is clearly outside its target.
3. "disable everything" <--- repeated for your benefit. It then goes on to show the ones that could be enabled without breaking the setup that the guide is trying to give you.
4. "Here's a simple guide... and if you feel like doing something slightly more advanced, here's how to configure that as well." This isn't really that complicated to understand.
5. The "quick" flash is much, much slower on some hardware (or drivers) than others. Starting a video and then having 5 seconds of black screen before it actually starts to play can be annoying.


Still though (and I don't think you understand this), I'm not arguing that the guide is good/bad. Just that you were an jerk for no particularly good reason. And that you didn't actually read the post before trying to mock the poster.

hoborg
12th January 2012, 19:18
Nothing changed here, still ~51FPS

clsid
12th January 2012, 19:19
Will there be a direct rendering mode in the future? For older ATI and Intel cards.
Any plans to implement some kind of smart fallback mode that switches back to software decoding if the GPU isn't fast enough to keep up with the fps of the video?


Another question, do you happen to know the minimum required version of nvcuvid.dll that is compatible with current LAV Video? Or at least the lowest version you know that is compatible.

adam777
12th January 2012, 19:20
Thanks, nev.

Everyhthing is done on a mobility radeon 4570, 512MB.

http://i.imgur.com/7tybi.png

Also did a (very) quick real world testing.

1080p50 (test sample) and 1080p60, both MKV, both completely unwatchable under Zoom Player and madVR.

1080p24 seems to work fine, running smooth, though madVR's queues are a bit on the lowish side (around 7-8/12).

Downloaded YouTube 1080p clip completely unwatchable (MP4).

Let me know when will you need more detailed reports, I'll be glad to help.

* EDIT *
second test build gives similar results.
Oh, I suppose it's possible the GPU simply isn't powerful enough to handle madVR at 1080p60.

nevcairiel
12th January 2012, 19:20
Nothing changed here, still ~51FPS

Too bad.
If anyone feels like testing alot, you could grab VLC and test its DXVA on such a 50p or 60p clip, and see if it works better? I doubt it would, but maybe AMD hacked up the driver to do something special when VLC is detected - its certainly possible.


do you happen to know the minimum required version of nvcuvid.dll that is compatible with current LAV Video? Or at least the lowest version you know that is compatible.

Not of the dll, but it should work with all CUDA 2.0 and upwards version, which is around driver 190 or so.

SamuriHL
12th January 2012, 19:22
Same for me, as well. Statistically maybe SLIGHTLY slower, but, 52 on average.

HeadlessCow
12th January 2012, 19:24
As a final point, there's something important to remember about web forums...

While blexley might be the 100th in a long line of newbies asking stupid questions on doom9, it's the first time he's been on doom9 and asked a stupid question. Treating him(or anyone new) as if he is asking a stupid question for the 100th time is just being mean.

CruNcher
12th January 2012, 19:39
A better listing includes GPU utilization as well (Perftest 1) :)

Progressive 8 bit Playback efficiency 1080p 60 fps (4 girl H.264)

Arcsoft DXVA = ~2% CPU; GPU = ~20%
Cyberlink DXVA = ~2% CPU; GPU = ~20%
CoreAVC DXVA = ~4% CPU; GPU = ~20%
Intel Quicksync DXVA2 = ~12% CPU; GPU = ~23%
Nev Generic DXVA2 = ~17% CPU; GPU = ~24%
Nev Lav Video Libav = ~25% CPU; GPU = ~25%


All DXVA2 and Software results (Lav Video). Results include Parser (Lav Splitter MKV)/Audio(Lav Audio)/Renderer (EVR Custom Experimental) Overhead

fairchild
12th January 2012, 19:51
My results on my HD 5830 using 11.12 drivers and system in my sig benched with graphstudionext (avg 3 runs to null renderer):

Lav Video DXVA2: 52.2 fps

for reference on same system:

Lav Video software: 143 fps

Edit: with your perftest2, same exact results as above.

CruNcher
12th January 2012, 19:51
Too bad, i would've hoped for faster.

Here is a second try:
http://files.1f0.de/lavf/LAVFilters-0.44-dxva2-perftest2.zip

You need the first version, and just replace the .ax file

I increased the number of buffers from 2 to 8, apparently on my Intel GPU this gives me a significant performance boost on some samples, so who knows, maybe it does something for ATI.
Otherwise, i'm not sure i know what else to try without an ATI card at hand - maybe its just everything we can get out of them...

It's always better to restart the host Application it could hold up the old dll still in memory better is to first make sure the host aplication doesn't do this, and if your not sure better restart it.

VipZ
12th January 2012, 20:05
Here's results from a 7970

Here's the correct results,
http://dl.dropbox.com/u/19338638/Decoder%20Perf.JPG

I guess the 280fps was too good to be true :p

@ madshi, here's the results for your test, http://dl.dropbox.com/u/19338638/madnvtest.JPG

nevcairiel
12th January 2012, 20:06
Well, a slight increase, bringing it closer to NVIDIAs performance - its unsure if this may just be the decoder limiting now, though.

Interesting that madshi's test app still only shows 4 fps NV12 download though. :p

SamuriHL
12th January 2012, 20:08
Knock it off. I don't have money for a new video card! :D

CruNcher
12th January 2012, 20:10
Though if you compare these results with Nvidias old cards it looks not really fast those results would be in the range of a 460 GTX if i remember right ;)

SamuriHL
12th January 2012, 20:11
Yea, well, I also game on my machine, so, you can compare the video all you want....but once you start looking looking at gaming numbers things might be slightly....different. :D

STaRGaZeR
12th January 2012, 20:12
~52fps with both builds on 1080p50 Crowdrun.

CPU usage lower than the QS decoder.

1080p60 unwatchable. 1080p50 starts with a few seconds of stuttering, then it's smooth with a few drops below 50fps here and there. But it's not LAV Video's fault, Microsoft's decoder fails too, the stuttering is a bit less pronounced but that's it. The hardware decoder itself is shit, nothing to be done about that.

PS: Parkjoy is harder to decode than Crowdrun. Better benchmark material :D

sneaker_ger
12th January 2012, 20:13
Nothing changed here with the second version, still around 50 fps on the HD 5850.

VipZ
12th January 2012, 20:13
Knock it off. I don't have money for a new video card! :D

hehe. Though as much as DXVA has been improved on the 7970, the power consumption is through the roof, 2600k is massively more efficient.

CPU is about 340w from the wall for my stuff and DXVA copy back is about 390w from the wall, this is for 720p, old DXVA test was just bit over 400w.

CruNcher
12th January 2012, 20:14
~52fps with both builds on 1080p50 Crowdrun.

Power consumption and CPU usage lower than the QS decoder.

1080p60 unwatchable. 1080p50 starts with a few seconds of stuttering, then it's smooth with a few drops below 50fps here and there. But it's not LAV Video's fault, Microsoft's decoder fails too, the stuttering is a bit less pronounced but that's it. The hardware decoder itself is shit, nothing to be done about that.

Hmm impossible i guess you forgot that the QS Decoder doesn't get used for QFHD resolution (not performant enough) so Nev disabled it and it fallsback to Libav ;)

Are you sure you took the 1080p results ?

nevcairiel
12th January 2012, 20:15
I'm not sure if there is anymore to be gotten out of the AMD HW decoder, maybe i should just get a cheap card like a 5450 and see if i can improve it somehow.


Hmm impossible i guess you forgot that the QS Decoder doesn't get used for QFHD resolution as Nev disabled it and it fallsback to Libav ;)
This is all 1080p.

SamuriHL
12th January 2012, 20:18
hehe. Though as much as DXVA has been improved on the 7970, the power consumption is through the roof, 2600k is massively more efficient.

CPU is about 340w from the wall for my stuff and DXVA copy back is about 390w from the wall, this is for 720p, old DXVA test was just bit over 400w.

I couldn't care less about power consumption tbh. :D Heat is what I'm more concerned with. I have a pretty sweet Antec case so even that's marginal. My rig is over 2 years old now and will probably get replaced once it hits the 3 year mark in November. I need to behave until then, however. :)

SamuriHL
12th January 2012, 20:20
I'm not sure if there is anymore to be gotten out of the AMD HW decoder, maybe i should just get a cheap card like a 5450 and see if i can improve it somehow.


Those of us on AMD hardware would certainly appreciate it. I run both AMD and nVidia in the same house, but, I know the AMD crowd is itching for hardware assisted decoding to use with madVR.

STaRGaZeR
12th January 2012, 20:22
There seems to be nothing more in the hardware. Pure DXVA solutions are marginally faster.

CruNcher
12th January 2012, 20:23
I'm not sure if there is anymore to be gotten out of the AMD HW decoder, maybe i should just get a cheap card like a 5450 and see if i can improve it somehow.



This is all 1080p.

yep sorry added that though it's strange doesn't fit to my QS Decoder results @ all (which from the top to bottom make absolute sense) though different bitstream.

nevcairiel
12th January 2012, 20:29
There seems to be nothing more in the hardware. Pure DXVA solutions are marginally faster.

Luckily, there also isn't a hell of a lot content that wouldn't work. Blu-ray maximum is 60i, which should play just fine (its essentially 30p)

Anyhow, i gotta fix my HTPC now, the damn thing somehow died, wonder WTF that thing has now.
And i was hoping for a smooth ride until i replace it with IVB in a few month

I think its the stupid SSD in it, meh.

Snowknight26
12th January 2012, 20:31
6870 @ stock
perftest1 = 55.15fps (every time, 20 tries)
perftest2 = 55.15fps (every time, 20 tries)

This was over remote desktop no less. Surprisingly, GPU usage was at a steady 35%. Guess it's limited by the memory speed. I'd try overclocking it but the CCC doesn't open over RDP. :\

CruNcher
12th January 2012, 20:57
Playback and Null Rendering Efficiency Results are conform (except Nev and Lav you would expect Nev DXVA2 to be faster then Libav)

Lav Video Intel Quicksync DXVA2:

http://img7.imageshack.us/img7/188/lavvideointel.png

Lav Video Libav:

http://img27.imageshack.us/img27/9109/lavvideolibav.png

Lav Video Nev DXVA2:

http://img252.imageshack.us/img252/3248/lavvideonevdxva2.png

madshi
12th January 2012, 21:06
@ madshi, here's the results for your test
Thanks!

Interesting that madshi's test app still only shows 4 fps NV12 download though. :p
Yes, quite interesting! I don't really know why this happens. Will have to check that. Maybe the memcpy implementation I'm using is making problems. Funny enough, it doesn't harm with NVidia GPUs.

STaRGaZeR
12th January 2012, 21:14
Luckily, there also isn't a hell of a lot content that wouldn't work. Blu-ray maximum is 60i, which should play just fine (its essentially 30p)

Anyhow, i gotta fix my HTPC now, the damn thing somehow died, wonder WTF that thing has now.
And i was hoping for a smooth ride until i replace it with IVB in a few month

I think its the stupid SSD in it, meh.

Indeed. I also have doubts about the supposedly ultra fast Intel decoder. When you benchmark it, the power consumption that the CPU reports for the GPU part of the chip increases to enormous levels compared to normal 1080i30 playback. I'm talking about sub 1W vs 7W and such. That doesn't seem like a normal load PC for a little fixed function hardware decoder. The GPU also enters 3D mode during the benchmark, and not while normal playing.

Out of curiosity, is that SSD Intel by any chance?

Guess it's limited by the memory speed. I'd try overclocking it but the CCC doesn't open over RDP. :\

On my 5850 playing with GPU clocks (300-800) and memory clocks (300-1000) resulted in similar perfomance.

STaRGaZeR
12th January 2012, 21:22
...

Slow!

QS decoder:

http://img638.imageshack.us/img638/595/43569286.png

Software:

http://img205.imageshack.us/img205/614/softg.png

Okay, I cheat, 4500MHz on the CPU, 1500MHz on the GPU ;)

nevcairiel
12th January 2012, 21:22
Out of curiosity, is that SSD Intel by any chance?

Nah, OCZ.
Guess i'll just put a good ol' magnetic back in there and get a new one when i upgrade the HTPC

But i think something else is going on, too, maybe the GPU broke as well, something truely odd is going on...
Meh, put in my GT430 that i had sitting around somewhere and it boots again, f'ing great. SSD and GPU gone boom, wonder if there was a power spike or something.

King Kong
12th January 2012, 21:38
http://www.abload.de/img/lavfilters-0.44-dxva2e5rem.png (http://www.abload.de/image.php?img=lavfilters-0.44-dxva2e5rem.png)

Tested w/ Intel i7-2600K & AMD Radeon HD 6950 2GB.
"Crowd run" test file stutters every 1-2 secs.

Not using Intel QSD due to limitations of P67 chipset.

bjd
12th January 2012, 21:57
@Nev

Great work on the DXVA2 (copy back), I can even use this hardware acceleration on my 4 yeaVista based AMD Turion TL-60 Laptop with a Radeon X1250

PeQuE
12th January 2012, 21:59
Is it my fault, or Dolby Digital Plus (EAC3) bitstreaming in Lav Audio is broken in 0.44?

In Graphstudio I can connect Lav Splitter audio pin to Lav Audio, but then it refuses to connect to audio renderer.

EDIT: it only affects if bitstreaming is selected... If not, no problem

bjd
12th January 2012, 22:03
@Nev

Great work on the DXVA2 (copy back), I can even use this to play 1080p24 on my 4 year old Vista based AMD Turion TL-60 Laptop with a Radeon X1250 !!

egur
12th January 2012, 22:20
Using ffdshow QS with background (MT) frame copy:
http://img638.imageshack.us/img638/8950/ffdshowqs23crowdrun1080.png
i7-2600 stock speed DDR3@1333MHz

CruNcher
12th January 2012, 23:25
@ Eric
Right i forgot its always good to check against ffdshow-quicksync efficiency wise :)

and indeed its still better if the benchmark result is really valid (see comment bellow) :)


1st. FFdshow Quicksync:

http://img97.imageshack.us/img97/9573/ffdshowquicksyncres.png

2nd. Lav Video Quicksync:

http://img7.imageshack.us/img7/188/lavvideointel.png

Either that or the result is false because of the different timestamp handling (also rate and real time doesn't get shown, could indicate a problem getting the FPS header) though this shouldn't influence the benchmark shouldn't it (lav splitter) ?
Though these stuff is one reason i dont like this benchmark splitter and decoder differences can result in very weird behavior and benchmark results same for DXVA Checker (Elecard filter for example benching is impossible, they send back a 00000 result for everything same for Mainconcept with lav splitter).


Rechecking the Playback results i see no improvement though compared to Lav Videos Quicksync (only in the benchmark)

Progressive 8 bit Playback efficiency 1080p 60 fps (4 girl H.264)

Arcsoft DXVA = ~2% CPU; GPU = ~20%
Cyberlink DXVA = ~2% CPU; GPU = ~20%
CoreAVC DXVA = ~4% CPU; GPU = ~20%
Lav Quicksync DXVA2 = ~12% CPU; GPU = ~23%
ffdshow-quicksync DXVA2 = ~12% CPU; GPU = ~23%
Nev Generic DXVA2 = ~17% CPU; GPU = ~24%
Nev Lav Video Libav = ~25% CPU; GPU = ~25%

Software results (Lav Video). Results include Parser (Lav Splitter MKV)/Audio(Lav Audio)/Renderer (EVR Custom Experimental) Overhead

Andy o
13th January 2012, 01:37
I know the AMD crowd is itching for hardware assisted decoding to use with madVR.

Is that true anymore, for most? I know it's not for me, deinterlacing was the last piece of the software-decoding puzzle and madVR has fixed that beautifully. It's been a few years since H.264 progressive was very demanding on current CPUs.

SamuriHL
13th January 2012, 01:39
Is that true anymore, for most? I know it's not for me, deinterlacing was the last piece of the software-decoding puzzle and madVR has fixed that beautifully. It's been a few years since H.264 progressive was very demanding on current CPUs.

Yes. It's one of the most asked for things. Does it TRULY matter to most? Probably not. For my laptop it'd make a big difference. For my HTPC, not really. For a lot of people, it's something they want. Notice, I didn't say NEED. :)

Andy o
13th January 2012, 01:47
Oh, I don't doubt it's a good thing, just wondering if people that think they need it (or even want it) are a bit misguided, cause for instance, system efficiency (was evidenced in VipZ's for instance) can suffer in many cases, contrary to popular belief that GPU decoding is always more efficient. You don't need a super powerful GPU for that to be true, either.

SamuriHL
13th January 2012, 01:52
My laptop and my bedroom HTPC would not be able to play video without GPU decoding. My laptop I've been able to work around the issue as it's faster than my bedroom HTPC, but, it definitely benefits from GPU decoding. Any modern CPU, however, is more than sufficient, yes. Some day I'll be getting a new laptop and rebuiding both HTPC's, but, until that happens......

zero9999
13th January 2012, 02:17
Too bad, i would've hoped for faster.
Here is a second try:
http://files.1f0.de/lavf/LAVFilters-0.44-dxva2-perftest2.zip

ATI HD 5750: ~51,2 fps (CrowdRun 1080p50)
Also, crashes with Hi10P material (instead of falling back to software decoding)

DragonQ
13th January 2012, 10:58
Nev, if you get DXVA2 working for AMD cards, will it work for nVidia cards too? If so, what are the advantages/disadvantages compared to CUVID?


By the way I get 57 FPS on average with software decoding on a Core 2 E8400 with CPU usage at ~95%. When using DXVA2 on an AMD Radeon HD3400, I get 59 FPS but CPU usage is still at ~95%. Such a poor GPU (this is an office PC) is probably useless for DXVA2.

wanezhiling
13th January 2012, 11:12
Mobility HD4650 (http://we.pcinlife.com/data/attachment/forum/201201/13/174540sy8yaxcyy8spu84y.png)
AVG 52.3325fps

nevcairiel
13th January 2012, 11:12
Nev, if you get DXVA2 working for AMD cards, will it work for nVidia cards too? If so, what are the advantages/disadvantages compared to CUVID?

Compared DXVA2 to CUVID:

Pro:
- Uses potentially less power because CUVID forces max performance mode

Con
- No deinterlacing
- Marginally slower
- Requires D3D interaction (cannot use it with a secondary card)
- No VC-1 interlaced support (unless i manage to fix that)

Note that i also have plans for a "native" DXVA2 mode for people running EVR, which would behave similar to all other DXVA decoders out there.

DragonQ
13th January 2012, 11:17
Compared DXVA2 to CUVID:

Pro:
- Uses potentially less power because CUVID forces max performance mode

Con
- No deinterlacing
- Marginally slower
- Requires D3D interaction (cannot use it with a secondary card)
- No VC-1 interlaced support (unless i manage to fix that)

Note that i also have plans for a "native" DXVA2 mode for people running EVR, which would behave similar to all other DXVA decoders out there.
Thanks. Lower power consumption would be nice, especially for an HTPC, but hardware deinterlacing is a must IMO.

nevcairiel
13th January 2012, 11:18
If you use EVR or madVR, just let them deinterlace.
It was more important when madVR did not do deinterlacing yet.

DragonQ
13th January 2012, 12:24
What deinterlacing does EVR use though? It famously has no options in most media players and the highest quality vector adaptive deinterlacing within CUDA is easily the best I've tried.

nevcairiel
13th January 2012, 13:16
It should use exactly the same. Its easy enough to compare with some artificial tests like the Cheese Slice test. Differences in the algorithms are clearly visible there.

wanezhiling
13th January 2012, 20:14
GT240 CUVID (http://we.pcinlife.com/data/attachment/forum/201201/14/031003e013h6sjjqyygs7c.png)
72.2765

GT240 DXVA2 (http://we.pcinlife.com/data/attachment/forum/201201/14/031003vi4bsjdnl6ot6svj.png)
71.8401

M HD4650 (http://we.pcinlife.com/data/attachment/forum/201201/13/174540sy8yaxcyy8spu84y.png)
52.3325


http://forum.doom9.org/showthread.php?p=1548287#post1548287
;)

CruNcher
13th January 2012, 20:18
It should use exactly the same. Its easy enough to compare with some artificial tests like the Cheese Slice test. Differences in the algorithms are clearly visible there.

Nev already a idea what goes wrong with dirac playback (slowmo) on the promos ?

Btw i find it nice to have now the possibility to switch between Yadif and Render Deinterlacing on the fly :) seeing the different in Realtime is cool how the GPU utilization goes up when the renderer does the deinterlacing and how it goes down when Yadif does it (and cpu therfore goes up) also comparing is cool this way doing everything Realtime now has such great possibilities for Encoding i love it :)
Though the Interface to the options is still to hard to reach some should start implementing those possibilities directly in the Player GUIs without needing to click 1000 times to get to them :P
Also using Lav Video + Av Splitter has some great potential it mostly has the same bugs though as Lav Splitter (Mp4 playback) but it allows to switch channels on the fly (.ts) and seems to work equally good with Lav Video (also it has the tray icon control many crave for, though personally i find it useless better is direct Player integration instead of this tray icon control stuff) :)

nevcairiel
13th January 2012, 22:12
Good news everyone!

I managed to implement VC-1 interlaced DXVA decoding.
That sure was a bitch, and locked up my GPU driver quite often while testing (thank god Win7 can recover from that), but now it seems to work just fine with both Frame Interlaced and Field Interlaced content!

hoborg
13th January 2012, 22:14
Nev, you rule! :D

SamuriHL
13th January 2012, 22:17
Damn, man, that's awesome. :)

Sebastiii
13th January 2012, 22:28
Yeah :) you are the best ;)

Reino
13th January 2012, 23:09
Is there a reason you skipped DXVA1 for DXVA2?

DragonQ
13th January 2012, 23:12
Well, DXVA is only useful for XP users since XP doesn't support DXVA2. Otherwise, DXVA2 is superior, so why not use it?

Supporting a ten year old OS is silly, IMO.

BatKnight
13th January 2012, 23:28
Good news everyone!

I managed to implement VC-1 interlaced DXVA decoding.
That sure was a bitch, and locked up my GPU driver quite often while testing (thank god Win7 can recover from that), but now it seems to work just fine with both Frame Interlaced and Field Interlaced content!
Nice!
Will it be possible to select DXVA for interlaced VC-1 and CUVID for H264 or MPEG-2? Now one can only select either DXVA or CUVID for everything.

Bat

wo0zy
14th January 2012, 00:50
Good news everyone!

I managed to implement VC-1 interlaced DXVA decoding.
That sure was a bitch, and locked up my GPU driver quite often while testing (thank god Win7 can recover from that), but now it seems to work just fine with both Frame Interlaced and Field Interlaced content!

Sweet!:D

I don't say it often enough so :thanks: for all your work!

This thread has everything! I follow it every day. Sometimes it makes me laugh, sometimes I cringe at the responses some people get (justified or not :) ) but mostly I feel educated by reading and look forward to testing something new!

Thanks Nev!

Wo0zy

nevcairiel
14th January 2012, 07:22
Is there a reason you skipped DXVA1 for DXVA2?

Several.

- DXVA1 doesnt allow frame copyback
- DXVA1 is only for Windows XP
- FFMPEG only supports DXVA2

nevcairiel
14th January 2012, 07:23
Nice!
Will it be possible to select DXVA for interlaced VC-1 and CUVID for H264 or MPEG-2? Now one can only select either DXVA or CUVID for everything.

Bat

Why would you want that? CUVID is superior.

Andy o
14th January 2012, 07:36
nev, do you know if HD MPEG2 software decoding handles errors differently than DXVA decoding? I've just started testing the 1080i and 720p wtv from OTA channels (I'm in the US so ATSC) recorded in Win 7's Media Center. They play relatively well, but with LAV Video's MPEG2 I get heavy gray blocking. With ffdshow, and the MS DTV-DVD decoder however, I don't get blocking, but I get slight color (reddish-greenish) and black level shifts. I don't get either when I just switch to EVR and a full DXVA chain, it plays just perfectly like on MC itself. So I still don't know for sure what's going on here, that's why I'm asking if errors from the capture could be the culprit, and somehow DXVA has better error resilience or something like that.

If it's not as simple as that, I could upload samples if anyone is interested to confirm.

robpdotcom
14th January 2012, 07:45
@ Andy o:

Are you using LAVSplitter? LAVvideo doesn't seem to like Microsoft's .wtv spltter (StreamBufferSource). If you use mpc-hc, block StreamBufferSource - you lose continuous live TV playback, but LAVSplitter works better for everything else, and gives you the ability to select audio streams.

Andy o
14th January 2012, 08:17
Thanks, you're right. I owe you a sandwich.

NikosD
14th January 2012, 11:34
Benchmark results - Signature system - Video clip requested - LAV 0.44 latest


GraphStudio 0.3.2/ GraphStudio Next 0.4.5.5 (same results - 0.1fps difference)

3 passes - Null Renderer


5750 50,4 fps - GPU load 48% - 50% - CPU Load Max frequency ~ 25% average

(6)750 56,2 fps - GPU load 52% - 53% - CPU Load Max frequency ~ 30% average


DXVA Checker v2.6.2


LAV Video DXVA2

(6)750 EVR 53/54/54 - GPU Load 0/33/74 - CPU load max frequency 22/23/27
(6)750 VMR 48/54/56 - GPU Load 0/33/74 - CPU load max frequency 22/24/27


MS DS (vanilla DXVA2)

A lot of artifacts - Not sure about valid speed


(6)750 EVR 58/66/69 - GPU Load 0/3/3 - CPU load LOWEST frequency@1.6 GHz 9/10/11


MS MFT(vanilla DXVA2)

Perfect decoding


(6)750 EVR 68/71/72- GPU Load 0/1/7 - CPU load LOWEST frequency@1.6 GHz 0/10/10


Nice try - I'll skip because of:

1) High GPU load

2) High CPU load

3) Low decoding performance

4) Absolutely no need for anything else than EVR renderer

Waiting for your DXVA vanilla solution.

UPDATE:
According to AMD and other sources over the Net, the only way to use Frame Copy with zero penalty is through OpenCL.
If you want to implement a fast DXVA Frame copy solution for ATI hardware, you have to use OVD API (OpenVideo Decode).
It's like NVCUVID and Intel's MSDK in its basic principle.
No other way. Sorry.

blexley
14th January 2012, 11:37
If you are running nothing but MPC-HC with LAV Filters do you only uncheck all the Transform Filters or do you uncheck all the Source Filters aswell. ?

All the guides seem to contradict each other with some even saying "Transform filters (Prefer) in 'External filters' override transform filters in 'Internal Filters' (usually). So there is no reason to disable all transform filters." :confused:

Midzuki
14th January 2012, 11:42
@ blexley:

Source Filters: use only for Audio CD and (S)VCD

Transform Filters: use none.

blexley
14th January 2012, 11:47
Awsome thanks alot Midzuki.

Think i'll just leave it as MPC-HC with LAV Filters as all this other stuff is confusing. :)

IanD
14th January 2012, 12:27
Just exploring the latest LAV package with MPC and MadVR on XP SP2, Nvidia 8400GS with 285.58 driver.

However, with HW decoding (Nvidia CUVID) enabled in the LAV Video Decoder, 720p mkv files are incredibly pixelated and torn: if I disable HW decoding, it is smooth and unpixellated. Divx avi is fine. Doesn't change if I use EVR instead of MadVR.

Is my 8400GS (the updated version, not the original) inadequate to support CUVID, or is something else wrong with my setup?

Mercury_22
14th January 2012, 12:53
@ Nev the latest x64 version ( "Revision 870a5b4451 dxva2: implement software fallback " ) it's just crashing (MPC-HC) while the x86 version it's not crashing (working with VC1i too :thanks:) but also has no DXVA just software

nevcairiel
14th January 2012, 12:55
@ Nev the latest x64 version ( "Revision 870a5b4451 dxva2: implement software fallback " ) it's just crashing (MPC-HC) while the x86 version it's not crashing (working with VC1i too :thanks:) but also has no DXVA just software

If VC1i is working, then you obviously have DXVA, because the software decoder doesn't decode it. :p
Note that MPC-HCs DXVA status display doesn't work with this decoder.

Mercury_22
14th January 2012, 12:58
If VC1i is working, then you obviously have DXVA, because the software decoder doesn't decode it. :p
Note that MPC-HCs DXVA display doesn't work with this decoder.
ok then ! waiting for the x64 fix :thanks:

Mixer73
14th January 2012, 13:22
However, with HW decoding (Nvidia CUVID) enabled in the LAV Video Decoder, 720p mkv files are incredibly pixelated and torn: if I disable HW decoding, it is smooth and unpixellated. Divx avi is fine. Doesn't change if I use EVR instead of MadVR.

Sounds to me like your files may not have been encoded in a DXVA compliant fashion. Perhaps post up MediaInfo for an offending file to confirm.


Are you using LAVSplitter? LAVvideo doesn't seem to like Microsoft's .wtv spltter (StreamBufferSource). If you use mpc-hc, block StreamBufferSource - you lose continuous live TV playback, but LAVSplitter works better for everything else, and gives you the ability to select audio streams.

Oh man, I have this symptom too, buuuut I am using DVR-MS, in this case just with MPEG2/AC3 inside, which LAV Splitter isn't set up for.

Nev: I believe everything DVR-MS is a small subset of WTV container, is it at all possible to support DVR-MS in LAV Splitter?

IanD
14th January 2012, 13:44
Sounds to me like your files may not have been encoded in a DXVA compliant fashion. Perhaps post up MediaInfo for an offending file to confirm.
Format : Matroska
Format version : Version 2
File size : 1.28 GiB
Duration : 41mn 53s
Overall bit rate : 4 363 Kbps
Encoded date : UTC 2012-01-03 12:05:59
Writing application : mkvmerge v5.2.1 ('A Far Off Place') built on Jan 2 2012 23:21:10
Writing library : libebml v1.2.3 + libmatroska v1.3.0

Video
ID : 1
Format : AVC
Format/Info : Advanced Video Codec
Format profile : High@L3.1
Format settings, CABAC : No
Format settings, ReFrames : 2 frames
Codec ID : V_MPEG4/ISO/AVC
Duration : 41mn 53s
Bit rate : 3 892 Kbps
Width : 1 280 pixels
Height : 720 pixels
Display aspect ratio : 16:9
Frame rate : 23.976 fps
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Progressive
Bits/(Pixel*Frame) : 0.176
Stream size : 1.14 GiB (89%)
Language : English
Default : Yes
Forced : No
Color primaries : BT.709-5, BT.1361, IEC 61966-2-4, SMPTE RP177
Transfer characteristics : BT.709-5, BT.1361
Matrix coefficients : BT.709-5, BT.1361, IEC 61966-2-4 709, SMPTE RP177

Audio
ID : 2
Format : AC-3
Format/Info : Audio Coding 3
Mode extension : CM (complete main)
Codec ID : A_AC3
Duration : 41mn 53s
Bit rate mode : Constant
Bit rate : 384 Kbps
Channel(s) : 6 channels
Channel positions : Front: L C R, Side: L R, LFE
Sampling rate : 48.0 KHz
Bit depth : 16 bits
Compression mode : Lossy
Stream size : 115 MiB (9%)
Language : English
Default : Yes
Forced : No

Also discovered it pixelates badly on Youtube flv files.

nevcairiel
14th January 2012, 13:58
ok then ! waiting for the x64 fix :thanks:

Should be fixed.

nevcairiel
14th January 2012, 14:17
Nev: I believe everything DVR-MS is a small subset of WTV container, is it at all possible to support DVR-MS in LAV Splitter?

DVR-MS works just fine with LAV Splitter, its just not a option in the installer. Just block the StreamBufferSource in MPC-HC and it should just work.

Mercury_22
14th January 2012, 14:20
Should be fixed.

Yes it is :thanks:
But maybe you can do something about the "DXVA" message cause even in the MPC-HC's stats (ctrl+J) it's saying "DXVA: Not using DXVA" ! :confused: and it's very confusing especially now with all this testing

nevcairiel
14th January 2012, 14:58
But maybe you can do something about the "DXVA" message cause even in the MPC-HC's stats (ctrl+J) it's saying "DXVA: Not using DXVA" ! :confused: and it's very confusing especially now with all this testing

I cannot.
I don't even know how that message worked, and frankly, i don't care. Complain to MPC-HC. :P

I've been thinking if i can change the name of the decoder itself in the filter list depending on which decoder is used, wonder if that would break anything.

SamuriHL
14th January 2012, 15:02
That message is notoriously unreliable.

Sent from my Xoom using Tapatalk

Kotik
14th January 2012, 15:07
Hiho,

Nev do you own an ATI card?

Cause i remember you mentioned that you don't have one.

Thanx once more for all your time.

P.S. waiting for a test build with VC-1i!!! :)

noee
14th January 2012, 15:19
I've been thinking if i can change the name of the decoder itself in the filter list depending on which decoder is used, wonder if that would break anything.

Yeah, probably, but it seems like a reasonable solution.

Running on an HD6570 here with a very underclocked/undervolted Propus, pretty solid. Though I don't have VC1i material, everything else I've thrown at it works well. I'm testing with madVR and MC17 (custom RO) and I built the latest pull.

Software fallback appears to be working fine now too.

**Edit**
Some quick and dirty power measurements, Kill-a-watt at the wall, Propus with K10Stat with P3 @ 700Mhz, .75Vcore
Machine idles at 78W (Seasonic SII 460, two HDs, HD6570 fanless, 4GB RAM, Win7 Ult)

Playback with MC17/LAV/madVR, using the same loop for each source, smooth, perfect playback all sources

LAV DXVA2

AVCHD 1080i60 (home video) madVR Deint On: 96W
MPEG2 NTSC DVD Rip Using madVR Deint On IVTC(3:2): 85W
VC1 BD rip (2001 Space Odyssey): 88W
VC1 BD rip (Blade Runner): 91W

One core bumps up to 1100Mhz occasionally, the others stick @ 700Mhz

LAV Software

AVCHD 1080i60 (home video) madVR Deint On: 111W (2 cores at P0:2300Mhz, 2 at P3)
MPEG2 NTSC DVD Rip Using madVR Deint On IVTC(3:2): 85W (all cores at P3)
VC1 BD rip (2001 Space Odyssey): 100W (1 core at P0, the other 3 jumping around a bit)
VC1 BD rip (Blade Runner): 105W (had to lock all 4 cores at P0:2300Mhz for watchable playback)

robpdotcom
14th January 2012, 15:46
DVR-MS works just fine with LAV Splitter, its just not a option in the installer. Just block the StreamBufferSource in MPC-HC and it should just work.

I also have to block WM ASF Reader.

[ReX]
14th January 2012, 16:04
It's just crashing with my Mobile Intel GM45.
I know it's bad, but I have to stick with it. :rolleyes:


http://i.imgur.com/6x4ni.jpg

This means it supports DXVA2, right?
LAV Video say so, too.

Superb
14th January 2012, 16:44
nev: Maybe you could provide some custom interface so that MPC-HC could query the information (regarding which decoder is active) when it finds LAV Video in the graph? Or do you already provide one?

nevcairiel
14th January 2012, 16:49
There already is an interface to get that information, but i don't understand why its so important. Set it up once, and enjoy the movie.

Anyhow, i accidentally fixed the DXVA display thingy when moving some logic around, don't ask me what triggers it and what doesnt..

nevcairiel
14th January 2012, 16:56
Here is a new test version:

http://files.1f0.de/lavf/LAVFilters-0.44-dxva2-test3.zip

New is DXVA2 VC-1 interlaced support (only in DXVA, not software), software fallback should now be functional, as well as some general fixings and improvements.

Here is also a second test version for users of older Intel GPUs, it may or may not improve things when using DXVA2:
http://files.1f0.de/lavf/LAVFilters-0.44-dxva2-test-gma.zip

I'm only going by some documentation i found, i don't actually have such a GPU. :(

Mercury_22
14th January 2012, 18:06
There already is an interface to get that information, but i don't understand why its so important. Set it up once, and enjoy the movie.

Anyhow, i accidentally fixed the DXVA display thingy when moving some logic around, don't ask me what triggers it and what doesnt..

Yes the "DXVA display thingy" it's showing now for all: mpeg2, H264 and VC-1 (i&P):thanks:

Now I only need an x64 DTS-HD decoder so I can get rid of the all x86 players & decoders for good :p

BatKnight
14th January 2012, 18:47
Why would you want that? CUVID is superior.

I was guessing DXVA could handle interlaced VC-1 better than CUVID?
I still can't get smooth playback on my Nine Inch Nails - Beside You In Time Blu-ray. (It's interlaced VC-1)
Audio gets desynced after a while.

Bat

nevcairiel
14th January 2012, 18:50
A video decoder would (theoretically) only cause desync if its too slow, but that would then also show as dropped frames in the renderer.
Its possible that something else is going wrong, but who knows.

BatKnight
14th January 2012, 19:44
A video decoder would (theoretically) only cause desync if its too slow, but that would then also show as dropped frames in the renderer.
Its possible that something else is going wrong, but who knows.
I'll try to gather more data/info on this and let you know later.

Bat

[ReX]
14th January 2012, 20:19
Wow, you're fast!

It seems to work very well.
Thank you so much for this, I really appreciate it.

nevcairiel
14th January 2012, 20:20
;1551594']Wow, you're fast!

It seems to work very well.
Thank you so much for this, I really appreciate it.

You tried the normal or the special intel version?

[ReX]
14th January 2012, 20:30
You tried the normal or the special intel version?

The Intel version.
But, weirdly enough, I've just tried the new normal version from the same post and it didn't crash.

CruNcher
14th January 2012, 21:57
i get a black screen on both VC-1 test samples (nev dxva2) mc.ts as well as cd.ts with quicksync decoder it works though with the old known decoding errors on mc.ts, cd.ts is absolute fine. Arcsoft and Cyberlinks Decoder have neither problems with both (all feed from lav splitter)

nevcairiel
14th January 2012, 22:03
DXVA2 doesnt support VC-1 on Intel, their VC-1 DXVA is non-standard, and i cba to figure out whats different.

CruNcher
14th January 2012, 22:04
also

Video: DRAC 640x360 25.00fps [Video]
Audio: MPEG Audio 48000Hz stereo 256kbps [Audio]

is still crawling @ 8 fps (droping frames like crazy did you miss to compile some important asm code ?) :(

nevcairiel
14th January 2012, 22:07
Sample, mine works fine.

CruNcher
14th January 2012, 22:09
allready posted it some pages ago http://forum.doom9.org/showpost.php?p=1550121&postcount=7974 :(

nevcairiel
14th January 2012, 22:19
Fixed

egur
14th January 2012, 22:37
I released rev28 (v0.23 with ffdshow r4251) today.
It has multithreaded a decode option and all around optimized MT code. The VC1 seek fix caused the aspect ratio fix to become disfunctional, so this version fixes both issues as well.
All new MT features have a chicken bit (disable).

Pat357
14th January 2012, 23:13
Playback and Null Rendering Efficiency Results are conform (except Nev and Lav you would expect Nev DXVA2 to be faster then Libav)
Just a quick question about the version of GraphStudio that you're using : I've just checked for a newer version, but the version on the web page is v0.3.2.0-Beta (14/June/2009).
This also the one I use now, but yours seems to have a different layout for benchmarking.:D
Like I see only a column "worker time" and "FPS" on my version; nothing about rates, etc, while yours is showing more columns with info. (more is always better;) )
Do you use a customized build or so ?
I'm interested in a more recent version, but I don't have the tools nor the time to build a version myself.
Can the version you use somewhere be downloaded ?
It seems the original author doesn't provide recent binaries anymore. (yes, I'm aware it's OS)
Thanks in advance.

wo0zy
14th January 2012, 23:18
Just a quick question about the version of GraphStudio that you're using : I've just checked for a newer version, but the version on the web page is v0.3.2.0-Beta (14/June/2009).
This also the one I use now, but yours seems to have a different layout for benchmarking.:D
Like I see only a column "worker time" and "FPS" on my version; nothing about rates, etc, while yours is showing more columns with info. (more is always better;) )
Do you use a customized build or so ?
I'm interested in a more recent version, but I don't have the tools nor the time to build a version myself.
Can the version you use somewhere be downloaded ?
It seems the original author doesn't provide recent binaries anymore. (yes, I'm aware it's OS)
Thanks in advance.

http://code.google.com/p/graph-studio-next/downloads/list

I think this is what you want.

Wo0zy

CruNcher
14th January 2012, 23:50
Fixed

Was it a splitter or decoder issue ?

nevcairiel
15th January 2012, 00:07
Was it a splitter or decoder issue ?

Decoder didn't deal with B-Frame timestamps correctly.

Aleksoid1978
15th January 2012, 00:27
http://www.megaupload.com/?d=GTJZPNL0 - bad playback in DXVA mode, good in CUVID mode. MPEG2.
Also - some MPEG2, the usual is .VOB from DVD - at start show green screen 1-2 seconds.

Keiyakusha
15th January 2012, 01:03
Did I missed something? Since when LAV has DXVA? I don't see it. Not that I need it, but still interesting.

SamuriHL
15th January 2012, 01:06
In the last 48 hours. :D

Mixer73
15th January 2012, 01:37
DVR-MS works just fine with LAV Splitter, its just not a option in the installer. Just block the StreamBufferSource in MPC-HC and it should just work.

Hrm OK, didn't work for me, looked like PotP couldn't find a splitter so defaulted to internal splitter, and fouled it up...

I also have to block WM ASF Reader.

I'll try this too.

EDIT:

No go. PotPlayer doesn't register LAVSplitter at all, video plays but no audio. No change even with LAVSplitter added as a registered filter specifically as a splitter for ASF files.

MPC-HC does register LAVSplitter, but video plays stupidly fast and no audio.

Any suggestions?

Pat357
15th January 2012, 02:31
Hrm OK, didn't work for me, looked like PotP couldn't find a splitter so defaulted to internal splitter, and fouled it up...
I'll try this too.

EDIT:
No go. PotPlayer doesn't register LAVSplitter at all, video plays but no audio. No change even with LAVSplitter added as a registered filter specifically as a splitter for ASF files.
MPC-HC does register LAVSplitter, but video plays stupidly fast and no audio.
Any suggestions?

Did you block StreamBufferSource + ASF reader in MPC-HC and set Lav-splitter, Lav-video and Lav-audio as "prefered" ?
(All internal MPC stuff as disabled).
What splitter got used when you open a .wtv file ?
Did you enable WTV when installing the LAV-filters ? If not sure, just reinstall Lav-filters using the installer (must use installer) and make sure WTV is "checked".

It works this way for me in MPC-HC ; I haven't this tried PotP though.

Snowknight26
15th January 2012, 02:45
The GM45 on a ThinkPad x200 isn't fast enough to do LAV Video's DXVA mode, sadly.

Unsurprisingly, the Google laptop's (Cr-48's) chipset doesn't even support it. Guess the Intel GMA 3150 is junk.

Mixer73
15th January 2012, 03:28
Did you block StreamBufferSource + ASF reader in MPC-HC and set Lav-splitter, Lav-video and Lav-audio as "prefered" ?(All internal MPC stuff as disabled).

You're right, this works in MPC-HC.

What splitter got used when you open a .wtv file ? Did you enable WTV when installing the LAV-filters ? If not sure, just reinstall Lav-filters using the installer (must use installer) and make sure WTV is "checked".

Not sure about WTV as I don't have any handy, definitely installed LAV Filters with all enabled.

PotPlayer just will NOT let me get LAV Splitter as the splitter for ASF. I've even deleted everything out of the PotP folder and blocked the other 2. It seems to be using something internally but it doesn't show in the filter list.

Oh well wonder if I can get MPC-HC to play without stutters, which is the reason I moved to PotP.

EDIT: Err, this combination results in playback out of sync. Sigh.

IanD
15th January 2012, 03:54
Tried the new test3 version of LAV filters with my pixellated mkv/flv files on 8400GS XP SP2 setup, but no difference: HW decoding with Nvidia CUVID enabled is completely pixellated, whilst software decoding is okay.

CTRL-J says "DXVA not used" regardless of the HW decoding option I select.

DXVA Checker has the following to say about the 8400GS, but unfortunately I can't insert an image as Doom9 says any image I use is invalid:

NVIDIA GeForce 8400 GS
ModeMPEG2_C: DXVA1, 720x480 / 1280x720 / 1920x1080
ModeMPEG2_D: DXVA1, 720x480 / 1280x720 / 1920x1080
ModeMPEG2and1_VLD: DXVA1, 720x480 / 1280x720 / 1920x1080
ModeMPEG2_VLD: DXVA1, 720x480 / 1280x720 / 1920x1080
ModeMPEG2_IDCT: DXVA2, 720x480 / 1280x720 / 1920x1080
ModeMPEG1_VLD: DXVA1, 720x480 / 1280x720 / 1920x1080
ModeVC1_VLD2010: DXVA1, 720x480 / 1280x720 / 1920x1080
ModeVC1_VLD: DXVA1/2, 720x480 / 1280x720 / 1920x1080
ModeVC1_IDCT: DXVA1/2, 720x480 / 1280x720 / 1920x1080
ModeWMV9_IDCT: DXVA1/2, 720x480 / 1280x720 / 1920x1080
ModeH264_VLD_NoFGT: DXVA1, 720x480 / 1280x720 / 1920x1080
212DC724-3235-44A4-BD29-E1652BBCC71C: DXVA1, (Skipped)

Anyone have any ideas why I'm getting pixellated playback with 720p mkv and Nvidia CUVID?

Does LAVFilters-0.44.exe include LAVCUVID-0.13.exe or are both required to be installed? It's not exactly clear in the thread.

NikosD
15th January 2012, 06:40
DXVA2 doesnt support VC-1 on Intel, their VC-1 DXVA is non-standard, and i cba to figure out whats different.

Maybe because Intel only supports the Advanced profile of VC-1 implementation - not Simple and Main profiles.
It's the same reason you couldn't get WMV3 acceleration with Intel's hardware.
Of course they could implement a std DXVA VC-1 even for advanced profile only, I think.

Anyway, if they reply you, it would be useful to post it here.
Thanks.

blexley
15th January 2012, 10:40
I released rev28 (v0.23 with ffdshow r4251) today.
It has multithreaded a decode option and all around optimized MT code. The VC1 seek fix caused the aspect ratio fix to become disfunctional, so this version fixes both issues as well.
All new MT features have a chicken bit (disable).

Amazing not one thanks for egur but if Nev should posts that he has had a cup of coffee then SamuriHL is straight on here declaring his Love for him. :D

nevcairiel
15th January 2012, 11:21
http://www.megaupload.com/?d=GTJZPNL0 - bad playback in DXVA mode, good in CUVID mode. MPEG2.
Also - some MPEG2, the usual is .VOB from DVD - at start show green screen 1-2 seconds.

I fixed the missing slices at the bottom and tweaked some other parameters. There is still corruption at start of that file that CUVID manages to hide, but i'm not sure thats so easy. CUVID is just better then DXVA. :)
MPC-HCs DXVA decoder seems to just crash on that file .. so no comparison there .. too bad.

nevcairiel
15th January 2012, 11:29
Anyone have any ideas why I'm getting pixellated playback with 720p mkv and Nvidia CUVID?

You wouldn't happen to have "Use HQ processing" checked, would you? If so, turn it off, it doesn't work on XP.

IanD
15th January 2012, 12:12
You wouldn't happen to have "Use HQ processing" checked, would you? If so, turn it off, it doesn't work on XP.
I did and now it's turned off all seems to be well. Thanks.

Are there any other XP gotchas?

wanezhiling
15th January 2012, 14:43
nev, this's PotPlayer's DXVA Renderless mode(Vista/Win7) (http://i.imgur.com/3u0z1.png).

CPU is about AVG 10% when using "LAV DXVA2 or DXVA Renderless" and 3% by "PotPlayer's pure DXVA mode" on my PC.

So, is LAV DXVA2 somewhere similar to DXVA Renderless? :)

SamuriHL
15th January 2012, 14:53
Amazing not one thanks for egur but if Nev should posts that he has had a cup of coffee then SamuriHL is straight on here declaring his Love for him. :D

Excuse me? I don't have an Intel device so I don't use egur's work. If I did I'd thank him just like I do all we've who's work I use. I think it's time to follow madshi's advice now...

Sent from my Xoom using Tapatalk

nevcairiel
15th January 2012, 15:00
So, is LAV DXVA2 somewhere similar to DXVA Renderless? :)

I've never used PotPlayer, but i guess the overall concept is similar, there isn't all that much that can be done differently.
Either you send the DXVA surface directly to the renderer ("native", or "pure"), or you copy the frame back to system memory and send that ("copy back", or "renderless", or however you call it)

wanezhiling
15th January 2012, 15:19
:thanks:

I think LAV CUVID is my favor.:D

Pat357
15th January 2012, 16:39
I've never used PotPlayer, but i guess the overall concept is similar, there isn't all that much that can be done differently.
Either you send the DXVA surface directly to the renderer ("native", or "pure"), or you copy the frame back to system memory and send that ("copy back", or "renderless", or however you call it)
Just FYI, from the PotP History.txt (10/01/2012) :
"Added Nvidia CUDA Decoder (MPEG1/2, H264/AVC, VC1 are supported)"
:D

pirlouy
16th January 2012, 00:14
Strangely, they stole CUVID source code, but they don't share their code for their DXVA (renderless) decoders. Not fair.
I know, Nevcairiel doesn't need it, but it's just by principle. Let's imagine they can't use Open Source license because of some law. At least, they could just give a bit...

CruNcher
16th January 2012, 01:02
Strangely, they stole CUVID source code, but they don't share their code for their DXVA (renderless) decoders. Not fair.
I know, Nevcairiel doesn't need it, but it's just by principle. Let's imagine they can't use Open Source license because of some law. At least, they could just give a bit...

"VC1 Adapter Filter" looks like Madshis VC1tweak (though the code was never officially released,was it ?) though it's not really hard for someone like Potplayer devs todo a similar one (see AV Splitter dev who also did something similiar though it doesn't work for .wmv) and surely also the CUVID stuff Nev wasn't the first one who implemented it ;)
So far the only prove what they steal (borrow) is ffmpeg and i would say the subtitle engine, some MPC-HC stuff (shader implmentation,EVR Custom), the rest seems speculation, maybe some parts of ffdshow too ;) )
But yeah they could be a little more open, though they integrate everything very well together in their Player it's optimized together for one purpose to work combined the best way possible questionable if it makes sense to port back most of the improvements as they are optimized for their Player framework (hides) @ least though they could have helped by giving back their DXVA improvements to MPC-HC and ffdshow-dxva decoders.

The Mp4 Splitter might be also based on Gabest (MPC-HC) as it doesn't show the same issues like the ffmpeg/libav one implemented in lav splittter and av splitter or they fixed it ;)

blexley
16th January 2012, 01:33
Excuse me? I don't have an Intel device so I don't use egur's work. If I did I'd thank him just like I do all we've who's work I use. I think it's time to follow madshi's advice now...

Sent from my Xoom using Tapatalk

So you are not Thankful for the work of people who you don't use but you expect new users that have never used LAV filters to be thankful.

The word hypocrite comes to mind.

SamuriHL
16th January 2012, 01:34
So you are not Thankful for the work of people who you don't use but you expect new users that have never used LAV filters to be thankful.

The word hypocrite comes to mind.

Yea, that's what I said. Goodbye.

Aleksoid1978
16th January 2012, 04:41
Hi nevcairiel. I will try to test LAVFilters-0.44-dxva2-test-gma.zip and LAVFilters-0.44-dxva2-test3.zip and cannot select DXVA - there is no in list.
Video - integrate Intel HD(i5 - Intel Havendale/Clarkdale Processor - Integrated Graphics Controller)

DXVAChecker - http://s017.radikal.ru/i434/1201/f9/6b751e69ae72t.jpg (http://radikal.ru/F/s017.radikal.ru/i434/1201/f9/6b751e69ae72.png.html)

Andy o
16th January 2012, 04:57
So you are not Thankful for the work of people who you don't use but you expect new users that have never used LAV filters to be thankful.

The word hypocrite comes to mind.

Seems pointless to argue with you, so not gonna do it. Just to mention you're not coming off only as a doom9 newbie, but as a forum newbie as well. You keep this behavior on almost any forum, and either you'll get ignored by most of the people you are asking help from, or outright banned.

blexley
16th January 2012, 07:38
Seems pointless to argue with you, so not gonna do it. Just to mention you're not coming off only as a doom9 newbie, but as a forum newbie as well. You keep this behavior on almost any forum, and either you'll get ignored by most of the people you are asking help from, or outright banned.

If Nev was abit more professional and had an official step by step guide for setting up the popular Mediaplayers or linked to an officially endorsed guide on the net that he supported as accurate and should be used as the standard starting point.Then the non-technical minded that do not understand filters and splitters would not have to signup and wait 5 days to be ignored because you lot are fedup with the same question being asked over and over again from 400 page thread that is impossible to comprehend and is outdated.

You are looking for fault in the wrong person when newbies shouldn't have to come here to start with.

Andy o
16th January 2012, 09:47
It's amazing what some people feel they're owed by others giving them stuff for free. Not gonna comment anymore, you probably shouldn't until you have an actual question about the topic.

blexley
16th January 2012, 09:55
It's amazing what some people feel they're owed by others giving them stuff for free. Not gonna comment anymore, you probably shouldn't until you have an actual question about the topic.

You are being as illogical and nonsensical as everybody else.

What use is something that is free if you can't use it and how can you be greatful or acting if you are owded for something that you haven't even used. ?

Sebastiii
16th January 2012, 10:49
Please blexley, try to be constructive :)

blexley
16th January 2012, 10:57
I thought pointing out the gapping flaw that the non-technical minded are being forced to come here for help with the same problem was being constructive. :)

Keiyakusha
16th January 2012, 11:20
Then the non-technical minded that do not understand filters and splitters would not have to signup and wait 5 days to be ignored because you lot are fedup with the same question being asked over and over again from 400 page thread that is impossible to comprehend and is outdated.

I'm sorry, I probably a bit out of topic but if you want to watch video on PC you should have at least basic understanding about filters stuff and how it works. "non-technical mindedness" is not an excuse. If someone is not non-technical minded to this point, I wonder why he uses PC at all.

blexley
16th January 2012, 11:45
That is like suggesting you need to understand the micro electronics and how HDMI leads work with digital broadcast just to use the remote on your TV.

Why would anybody need to understand anything more than install and play. ?

psymed
16th January 2012, 11:52
delete this post
Why would anybody need to understand anything more than install and play. ?If you want "plug and play", download cccp? It includes lav filters.

blexley
16th January 2012, 12:08
Why are you supposed to install loads of stuff from cccp just to use the included lav filters. ?

MokrySedeS
16th January 2012, 12:13
You are really stubborn, aren't you? nevcairiel already told you that he won't write a guide and even explained why he won't do it, but you just keep on pushing...
Would you like to do something in your free time that doesn't benefit you in any way (fun, experience, money etc)? Of course you wouldn't. So stop expecting (or demanding, actually) it from anyone else.

blexley
16th January 2012, 12:25
Who say he has to write it. ?

Wouldn't the fact that so many ask the same question mean that something is needed.Nothing stopping him from appointing a volunteer to write it for him to look over it once it is done.

Keiyakusha
16th January 2012, 12:25
That is like suggesting you need to understand the micro electronics and how HDMI leads work with digital broadcast just to use the remote on your TV.

Not at all. I'm talking about required level of knowledge. You don't need to know micro electronics to use HDMI. The same you don't need to know programming to use filters. However you should know the process of how to actually plug in HDMI, aren't you? The same way you should have understanding of how to "plug in" filters and related stuff. Or another example, when you making some tea, you don't have to know about chemical processes behind it, but you have to know how to prepare it, add some sugar (if you like it) and stuff.
For me, what you're asking for sounds similar to "how to prepare some tea?"
EDIT: also as was already mentioned, there is some cafe called "CCCP" and it seems they have nice tea in their menu.

blexley
16th January 2012, 12:35
I think we will have to agree to disagree on that one. :)

It's like when security software is installed and all the settings are turned off it is also suggested that the user has a grasp of security to use it when i'm of the frame of mind that everything should be turned on to protect the user until they know what they are doing.

Not sure if that made sense but we just see two different sides.

[ReX]
16th January 2012, 13:11
Here's an updated guide.
1-Download LAV Filters (http://code.google.com/p/lavfilters/downloads/list)
2-Download MPC-HC (http://xhmikosr.1f0.de/index.php?folder=bXBjLWhj)
3-Install LAV Filters (and check every single checkbox on the setup screen)
4-Install MPC-HC
5-Open MPC-HC, go to "View" -> "Options"
6-Go to "Internal Filters" and uncheck AVI, Matroska, MP4/MOV, MPEG PS/TS/PVA and OGG under "Source Filters", and everything under "Transform Filters"
7-Still on the Options screen, go to "External Filters"
8-Add both "LAV Audio Decoder" and "LAV Video Decoder", set both to "Prefer".

(not sure if step 6 is really needed, but never hurts to do it)

Now you have no reason to keep clogging the thread with your nonsense, if you do, we can be 100% sure that you're trolling.

blexley
16th January 2012, 13:20
;1551956']Here's an updated guide.

< snip >

Now you have no reason to keep clogging the thread with your nonsense, if you do, we can be 100% sure that you're trolling.

Well what you posted is what i thought but Nev said it was outdated. ?

LOL :p

I spent a week going around in circles trying to find an easy step by step install guide and i've just found it at the very page i first started at, nevcairiel own HomePage. :D :D :D

http://1f0.de/lav-cuvid/guide/

nevcairiel is there any chance you could make this guide more obvious in the first post of this thread and on the start of your webpage so newbies and people with low computer skills such as myself can avoid the obnoxious comments from users when asking for guidance as if LAV filters is some how for the elite and you need to search and read about codecs and splitters before you can use it.As all we need to know is what to adjust and tick/untick so we can just get on and use it and your guide does that prefectly.

Plus the more accessible it is and the quicker we can get going the quicker newbies can hit the paypal donate. :)

Thanks

No, because the guide is outdated, and LAV CUVID has been deprecated.

nevcairiel
16th January 2012, 13:24
I said the guide is outdated because LAV CUVID is outdated (its now integrated into LAV Video). The install steps are very generic and can probably apply to any filter install, just replace LAV CUVID with <filter of your choice>.

SamuriHL
16th January 2012, 13:44
I said the guide is outdated because LAV CUVID is outdated (its now integrated into LAV Video). The install steps are very generic and can probably apply to any filter install, just replace LAV CUVID with <filter of your choice>.

That ignore function of the forum really rocks. I owe madshi for that recommendation. :D

Sent from my Xoom using Tapatalk

nevcairiel
16th January 2012, 13:54
That ignore function of the forum really rocks.


No it doesn't. It doesn't hide posts of people responding to ignored people, and it doesn't really hide the posts - i can still see that the posts are there, just the text is hidden. :p

SamuriHL
16th January 2012, 13:59
No it doesn't. It doesn't hide posts of people responding to ignored people, and it doesn't really hide the posts - i can still see that the posts are there, just the text is hidden. :p

That's true. I see a lot of responses to posts I can't see. Lol. In tapatalk the op's aren't shown at all. So I do only see replies. Still, better than wading through useless nonsense. :) anyway, sadly I'll be away from my htpc's this week so I won't be able to do any testing. I'm very impressed with the dxva2 support.

Sent from my Xoom using Tapatalk

Pat357
16th January 2012, 14:22
Nev,
I noticed a cosmeric bug : the Lav-splitter out-pin and the video-in and -out pin from Lavvideo report 6fps instead of 60fps
(In DXVA mode on my GTX 570)
I use LAVFilters-0.44-fc517a9
Filter : LAV Video Decoder - CLSID : {EE30215D-164F-4A92-A4EB-9D4C13390F9F}

- Connected to:
CLSID: {B98D13E7-55DB-4385-A33D-09FD1BA26338}
Filter: LAV Splitter Source
Pin: Video

- Connection media type:
Video: MPEG4 Video (H264) 1920x1080 6fps 20897kbps

AM_MEDIA_TYPE:
majortype: MEDIATYPE_Video {73646976-0000-0010-8000-00AA00389B71}
subtype: Unknown GUID Name {31435641-0000-0010-8000-00AA00389B71}
formattype: FORMAT_MPEG2_VIDEO {E06D80E3-DB46-11CF-B4D1-00805F6CBBEA}
bFixedSizeSamples: 0
bTemporalCompression: 1
lSampleSize: 1
cbFormat: 162

VIDEOINFOHEADER:
rcSource: (0,0)-(1920,1080)
rcTarget: (0,0)-(1920,1080)
dwBitRate: 20897898
dwBitErrorRate: 0
AvgTimePerFrame: 166667

VIDEOINFOHEADER2:
dwInterlaceFlags: 0x00000000
dwCopyProtectFlags: 0x00000000
dwPictAspectRatioX: 16
dwPictAspectRatioY: 9
dwControlFlags: 0x00000000
dwReserved2: 0x00000000

MPEG2VIDEOINFO:
dwStartTimeCode: 0
cbSequenceHeader: 30
dwProfile: 0x0000004d
dwLevel: 0x0000002a
dwFlags: 0x00000004

BITMAPINFOHEADER:
biSize: 40
biWidth: 1920
biHeight: 1080
biPlanes: 1
biBitCount: 24
biCompression: AVC1
biSizeImage: 6220800
biXPelsPerMeter: 0
biYPelsPerMeter: 0
biClrUsed: 0
biClrImportant: 0

PS: it seems independent from the choosen vid mode : softw, CUVID en DXVA all report the same (Lav-splitter is sending 6 fps this downstream)
50 fps is reported as 5fps, 25fps is OK.

nevcairiel
16th January 2012, 14:27
There is nothing wrong with the media type.

AvgTimePerFrame defines the FPS, and 166667 (in 100ns units) is 16.6667ms, which is 60fps.
For comparison, 24p is 417084

MPC-HC must be messing it up.

rack04
16th January 2012, 14:36
Where is the best place to download the latest IntelQuickSyncDecoder.dll?

It appears that these (http://xhmikosr.1f0.de/index.php?folder=bGF2ZmlsdGVycw==) builds don't include the dll.

nevcairiel
16th January 2012, 14:38
I would generally recommend to stick with the official versions, they come out pretty regularly.
Most people that run inbetween versions either build them themselves, or use the ones i post for testing specific features. :)

On the specific question, there is no official place to download that dll.

blexley
16th January 2012, 14:49
Where is the best place to download the latest IntelQuickSyncDecoder.dll?

It appears that these (http://xhmikosr.1f0.de/index.php?folder=bGF2ZmlsdGVycw==) builds don't include the dll.

I would generally recommend to stick with the official versions, they come out pretty regularly.
Most people that run inbetween versions either build them themselves, or use the ones i post for testing specific features. :)

On the specific question, there is no official place to download that dll.


What about the poor chap that i pointed out that nobody thanked. ?

http://forum.doom9.org/showpost.php?p=1551632&postcount=8205

http://forum.doom9.org/showthread.php?p=1551632#post1551632

egur
QuickSync Decoder author

Eric Gur, Processor Client Application Engineer
Intel QuickSync Decoder author
Intel Corp.

http://sourceforge.net/projects/qsdecoder/

rack04
16th January 2012, 15:10
What about the poor chap that i pointed out that nobody thanked. ?

http://forum.doom9.org/showpost.php?p=1551632&postcount=8205

http://forum.doom9.org/showthread.php?p=1551632#post1551632

egur
QuickSync Decoder author
__________________
Eric Gur, Processor Client Application Engineer
Intel QuickSync Decoder author
Intel Corp.

http://sourceforge.net/projects/qsdecoder/

I'll take the higher road and not say what I was thinking.

BTW, I don't want to install ffdshow just to get the dll.

Pat357
16th January 2012, 15:11
There is nothing wrong with the media type.

AvgTimePerFrame defines the FPS, and 166667 (in 100ns units) is 16.6667ms, which is 60fps.
For comparison, 24p is 417084
MPC-HC must be messing it up.
I know there is nothing wrong in the way that it works wouthout any problems.
I did not know that the FPS rate reported at the pins is done by the player ; I really thought it was the filter that sets this fps value.
Now I have learned something new ; thanks Nev for your patience !

Any news (good or bad) about the corruption in DXVA-copy-back mode at the beginning at some samples ?
It seems that properly cut samples (starting with I frame - closed GOP) don't have this corruption.
Is this corruption shown only with samples that start with open GOP ?

nevcairiel
16th January 2012, 15:28
What about the poor chap that i pointed out that nobody thanked. ?

He doesn't provide a download of only that dll, and even if he did, i cannot guarantee that versions other then the one i bundle with my versions will work properly, so i would not recommend downloading them.


I did not know that the FPS rate reported at the pins is done by the player ; I really thought it was the filter that sets this fps value.

It is set by the filter, in form of that AvgTimePerFrame value (the FPS is just the inverse of that), the line on top that says 6fps is completely computed by MPC-HC from the values below, though.

Keiyakusha
16th January 2012, 16:34
What about the poor chap that i pointed out that nobody thanked. ?

There is specific threads for quicksync decoder. Wasn't he was thanked there? Why someone should thank him here? This is thread about LAV, here we should say thanks to nev.

ranpha
16th January 2012, 17:05
He doesn't provide a download of only that dll, and even if he did, i cannot guarantee that versions other then the one i bundle with my versions will work properly, so i would not recommend downloading them.


Whoa, don't tell me that the dll you provide in your test binaries isn't just a simple compile of what you can get from egur's repository? Did you modify it the way you do with ffmpeg/libbluray?

nevcairiel
16th January 2012, 17:13
Whoa, don't tell me that the dll you provide in your test binaries isn't just a simple compile of what you can get from egur's repository? Did you modify it the way you do with ffmpeg/libbluray?

No its a vanilla build, but not from the latest version, which has quite some changes (and i'm not sure egur aims at maintaining ABI/API compatiblity)

STaRGaZeR
16th January 2012, 17:26
I'm using the last version with no modifications to LAV's code. Of course the new options are at their default values but that's it. Works just fine. I feel like it seeks a bit faster, but since that's already difficult it may just be placebo. Have you taken a look at the corrupted frame flag it reports now?

nevcairiel
16th January 2012, 17:27
Have you taken a look at the corrupted frame flag it reports now?

That flag is not what one would expect - ie. not helpful.
It would probably help if Eric can manage to make it output the proper frame type, then i could just use that information and wait for an I frame - which is what the software and cuvid decoders also do.

STaRGaZeR
16th January 2012, 17:32
It doesn't now? That's something that should be fixed. Apps using the dll could make wrong asumptions based on wrong information.

jos99
16th January 2012, 18:08
Can you confirm whether LAV video decoder will decode WMV9 content, and whether it can use acceleration as Windows Media Player/Center can do so (possibly VLC also which has similar low CPU usage)?

nevcairiel
16th January 2012, 18:11
Can you confirm whether LAV video decoder will decode WMV9 content, and whether it can use acceleration as Windows Media Player/Center can do so (possibly VLC also which has similar low CPU usage)?

The next version will be able to decode WMV9 using a new DXVA2 decoder on AMD and NVIDIA, the current version only accelerates VC-1 (and only on NVIDIA and Intel)

Speakign of, I'm actually not sure why CUVID doesn't like WMV9, but oh well, not an important format.

jos99
16th January 2012, 18:30
The next version will be able to decode WMV9 using a new DXVA2 decoder on AMD and NVIDIA, the current version only accelerates VC-1 (and only on NVIDIA and Intel)

Speakign of, I'm actually not sure why CUVID doesn't like WMV9, but oh well, not an important format.
That is great news, thx:). Maybe CUVID/CUDA development came after VC1 was already developed? Quite a lot of sites still have HD content in WMV9 so it would be useful.

wanezhiling
16th January 2012, 18:53
nev, I found a inderlaced MPEG-4 video, use LAV CUVID with mpc-hc, it runs at 30fps as fraps shows, but 60 with PotPlayer.

Absolutely PotP is right.

So this's a mpc-hc's problem or CUVID?

With mpc-hc (http://i.imgur.com/CusQR.png)
With PotPlayer (http://i.imgur.com/79dXM.png)

Sample is here: http://www.gokuai.com/f/dP8EPZ6P8I3S8F7E

nevcairiel
16th January 2012, 19:30
nev, I found a inderlaced MPEG-4 video, use LAV CUVID with mpc-hc, it runs at 30fps as fraps shows, but 60 with PotPlayer

I see the problem, its a timestamp issue, i'll try to fix it. Stupid MPEG4 reports wrong framerate.

NikosD
16th January 2012, 19:54
The next version will be able to decode WMV9 using a new DXVA2 decoder on AMD and NVIDIA, the current version only accelerates VC-1 (and only on NVIDIA and Intel)


Just to inform you that your current version is able to DXVA decode WMV9 (both WMV3 and VC-1 advanced profile) on ATI's hardware.

I can see it in front of my eyes!

nevcairiel
16th January 2012, 20:14
I see the problem, its a timestamp issue, i'll try to fix it. Stupid MPEG4 reports wrong framerate.

Fixed

wanezhiling
16th January 2012, 20:39
Fixed

Good job.:thanks:

Thunderbolt8
16th January 2012, 20:39
The next version will be able to decode WMV9 using a new DXVA2 decoder on AMD and NVIDIA, the current version only accelerates VC-1 (and only on NVIDIA and Intel)
will it be possible to play interlaced (BBC) VC-1 stuff via decoding with WMV9 then using ATI? I only remember that you can choose between vc-1 and wmv9 with ffdshow decoder, but I guess it wouldnt work, would it?

nevcairiel
16th January 2012, 20:43
will it be possible to play interlaced (BBC) VC-1 stuff via decoding with WMV9 then using ATI? I only remember that you can choose between vc-1 and wmv9 with ffdshow decoder, but I guess it wouldnt work, would it?

VC-1 is fully supported as well, not only WMV9.
The question was about wmv9.

NikosD
16th January 2012, 21:02
Just FYI there is a common misconception regarding WMV9.
Microsoft's WMV9 is the implementation of the whole VC-1 standard.
The VC-1 Simple and Main profiles are fully compliant with WMV3.
The VC-1 Advanced profile is the "pure" VC-1 format (FourCC=WVC1) not included in the initial WMV9 family before the standardization of VC-1, but it was added after.
So WMV9 includes all three profiles of VC-1.

nevcairiel
16th January 2012, 21:13
the 7-9 and 1-3 names are used all around randomly, WMV8=WMV2, WMV9=WMV3 .. WMV9 Advanced Profile = WVC1 .. its a whole mess of things. :p

NikosD
16th January 2012, 21:34
Yes indeed.
Because VC-1 was based on WMV9, adding essentially only one profile (advanced) the WVC1, but Microsoft had already an advanced profile called WMVA which is not VC-1 compliant and it's now deprecated.

The old WMV9 with WMV3 and WMVA etc was kept with same name WMV9, adding VC-1 advanced profile WVC1 in order to support the whole VC-1 (simple, main, advanced) profile.
The simple and main profiles of VC-1 are essentially WMV3.

I hope it's clear now :D

PeQuE
16th January 2012, 21:35
Is it my fault, or Dolby Digital Plus (EAC3) bitstreaming in Lav Audio is broken in 0.44?

In Graphstudio I can connect Lav Splitter audio pin to Lav Audio, but then it refuses to connect to audio renderer.

EDIT: it only affects if bitstreaming is selected... If not, no problem

Could anyone using similar config confirm or not this bug?

I could provide some sample files if needed...

Thanks a lot...

nevcairiel
16th January 2012, 21:36
If the audio renderer doesnt want to connect to E-AC3, then it usually means it thinks your hardware doesn't support it.

VipZ
16th January 2012, 21:44
Nev, are you planning to add MPEG4 for DXVA? not that its really needed, but cool to have anyway if it doesn't require much work to add :)

I may be missing something here, but does AMD support WMV9 via DXVA2?

http://dl.dropbox.com/u/19338638/7970.JPG

nevcairiel
16th January 2012, 21:47
WMV9/3 is handled by the VC-1 decoder in DXVA2

I'll possibly add MPEG4-ASP to DXVA some day, but ffmpeg doesnt have support for it yet, so it is a bit more effort.

VipZ
16th January 2012, 21:56
Thanks :)

With this sample, CrowdRun_2160p50.x264.CRF24.mkv. Its to be expected not to decode properly on any hardware with DXVA copy-back? When I test its very slow and 1/2 the screen is green (GPU load is always under 20% and CPU under 5%)

nevcairiel
16th January 2012, 22:05
If the hardware claims it supports such a resolution, it'll most certainly try to decode it.
Sadly i don't have any hardware to test this on (yet), mine just blocks the attempt at using DXVA

jos99
16th January 2012, 22:57
Just FYI there is a common misconception regarding WMV9.
Microsoft's WMV9 is the implementation of the whole VC-1 standard.
The VC-1 Simple and Main profiles are fully compliant with WMV3.
The VC-1 Advanced profile is the "pure" VC-1 format (FourCC=WVC1) not included in the initial WMV9 family before the standardization of VC-1, but it was added after.
So WMV9 includes all three profiles of VC-1.
Confusingly, the mediainfo from my HD wmv clip says VC-1 and WMV3- I thought WMV3 was older standard?

Video
ID : 2
Format : VC-1
Format profile : MP
Codec ID : WMV3
Codec ID/Info : Windows Media Video 9
Codec ID/Hint : WMV3
Description of the codec : Windows Media Video V9 - 1920 x 1080, 29.96 fps, Q=90 (VBR)

nevcairiel
16th January 2012, 22:58
VC-1 Main Profile (MP) is 100% compliant with WMV3. Only Advanced Profile is new in VC-1

jos99
16th January 2012, 23:29
VC-1 Main Profile (MP) is 100% compliant with WMV3. Only Advanced Profile is new in VC-1
I see, thanks, so should this be accelerated with NVidia 9800GT on the current version or on the next LAV version? DXVA Checker lists

ModeVC1_IDCT: DXVA1/2, 720x480 / 1280x720 / 1920x1080
ModeWMV9_IDCT: DXVA1/2, 720x480 / 1280x720 / 1920x1080

nevcairiel
16th January 2012, 23:37
The CUVID decoder should be able to use that, yes.
Its only partial acceleration, but better then nothing.

jos99
17th January 2012, 01:25
The CUVID decoder should be able to use that, yes.
Its only partial acceleration, but better then nothing.
So are you saying CUVID works now but in the next version DXVA will be better:o? I just found this document searching for diff between VC1_IDCT and VC1_VLD, dunno if this is any use to you

http://download.microsoft.com/download/5/d/c/5dc91455-0a8c-4d57-8975-0afe1cfaf0dd/DXVA_WMV.pdf

mindbomb
17th January 2012, 04:06
possible problem.
my friend was playing an mkv, and within the folder that the mkv was in was an mka with the same file name.

This caused another instance of lav splitter to open, which caused another lav audio to open, and he had multiple audios playing at the same time.

I believe the audio was AAC in both files, but I'm not certain.

the player used was mpc hc v1.5.2.3456, lav audio .44 and lav splitter .44.

Andy o
17th January 2012, 06:54
That's likely a feature for playing video and audio streams from separate files, same as you can do with subs. Just change the file name or put it in another folder. (mka is matroska's audio container just in case.)

golagoda
17th January 2012, 06:57
Here is a new test version:

http://files.1f0.de/lavf/LAVFilters-0.44-dxva2-test3.zip

New is DXVA2 VC-1 interlaced support (only in DXVA, not software), software fallback should now be functional, as well as some general fixings and improvements.

Here is also a second test version for users of older Intel GPUs, it may or may not improve things when using DXVA2:
http://files.1f0.de/lavf/LAVFilters-0.44-dxva2-test-gma.zip

I'm only going by some documentation i found, i don't actually have such a GPU. :(

Thanks for this test build, I assume you're now using 0.9.x or the latest git master of ffmpeg now because this build at least in my tests has improved seeking in H.264 greatly, before when I would use LAVFilters in combination with madVR there would be a bit of a lag when you tried to seek, but now it's instantaneous and I noticed 'update ffmpeg' a few times on google code.

nevcairiel
17th January 2012, 07:46
So are you saying CUVID works now but in the next version DXVA will be better:o?[/url]

CUVID supports partial acceleration, my DXVA2 does not and will never. Only VLD in DXVA is supported.


my friend was playing an mkv, and within the folder that the mkv was in was an mka with the same file name.

This caused another instance of lav splitter to open, which caused another lav audio to open, and he had multiple audios playing at the same time

Thats a MPC-HC feature. You can either turn that feature off completely (Playback -> Auto-load audio files), or just enable the internal Audio Switcher, which will allow you to choose which file is actually output, instead of outputting both.
LAV has zero influence on this.

mindbomb
17th January 2012, 07:52
i see, thanks for clearing that up.

Aleksoid1978
17th January 2012, 07:56
Nev,
I noticed a cosmeric bug : the Lav-splitter out-pin and the video-in and -out pin from Lavvideo report 6fps instead of 60fps
(In DXVA mode on my GTX 570)
I use LAVFilters-0.44-fc517a9
Filter : LAV Video Decoder - CLSID : {EE30215D-164F-4A92-A4EB-9D4C13390F9F}

- Connected to:
CLSID: {B98D13E7-55DB-4385-A33D-09FD1BA26338}
Filter: LAV Splitter Source
Pin: Video

- Connection media type:
Video: MPEG4 Video (H264) 1920x1080 6fps 20897kbps

AM_MEDIA_TYPE:
majortype: MEDIATYPE_Video {73646976-0000-0010-8000-00AA00389B71}
subtype: Unknown GUID Name {31435641-0000-0010-8000-00AA00389B71}
formattype: FORMAT_MPEG2_VIDEO {E06D80E3-DB46-11CF-B4D1-00805F6CBBEA}
bFixedSizeSamples: 0
bTemporalCompression: 1
lSampleSize: 1
cbFormat: 162

VIDEOINFOHEADER:
rcSource: (0,0)-(1920,1080)
rcTarget: (0,0)-(1920,1080)
dwBitRate: 20897898
dwBitErrorRate: 0
AvgTimePerFrame: 166667

VIDEOINFOHEADER2:
dwInterlaceFlags: 0x00000000
dwCopyProtectFlags: 0x00000000
dwPictAspectRatioX: 16
dwPictAspectRatioY: 9
dwControlFlags: 0x00000000
dwReserved2: 0x00000000

MPEG2VIDEOINFO:
dwStartTimeCode: 0
cbSequenceHeader: 30
dwProfile: 0x0000004d
dwLevel: 0x0000002a
dwFlags: 0x00000004

BITMAPINFOHEADER:
biSize: 40
biWidth: 1920
biHeight: 1080
biPlanes: 1
biBitCount: 24
biCompression: AVC1
biSizeImage: 6220800
biXPelsPerMeter: 0
biYPelsPerMeter: 0
biClrUsed: 0
biClrImportant: 0

PS: it seems independent from the choosen vid mode : softw, CUVID en DXVA all report the same (Lav-splitter is sending 6 fps this downstream)
50 fps is reported as 5fps, 25fps is OK.

Can you upload a sample - for test with MPC-HC, i try to fix.

nevcairiel
17th January 2012, 08:07
Can you upload a sample - for test with MPC-HC, i try to fix.

Its in src/DSUtil/MediaTypeEx.cpp, line 72 and 82, the TrimRight will remove all extra zeros and dots, so 60.000 becomes 6 and 30.000 becomes 3.
To preserve the current functionality, do it with two TrimRights, first trim zeros (60.000 becomes 60.), then trim a dot. Don't trim both at the same time. :)

PeQuE
17th January 2012, 09:57
If the audio renderer doesnt want to connect to E-AC3, then it usually means it thinks your hardware doesn't support it.

That's what I though first... but then I realised that ffdshow is also bitstreaming EAC3 to my hardware, and audio renderer is connecting with no problem... ¿? May be ffdshow is fooling the renderer...

But ok... I'll need to investigate a little bit more. I'll try downgrading to 0.43 or lower, as I don't remember to have seen this problem before. Anyway, not a big problem disabling bitstreaming for EAC3 if that's the case.

Thanks a lot.

nevcairiel
17th January 2012, 11:19
Bitstreaming didn't change at all for quite a while, i doubt the issue goes away.

Aleksoid1978
17th January 2012, 13:20
Nev - bad playback http://www.megaupload.com/?d=2KAAU7RR VC1 1080p60. MPC-HC also bad ))

nevcairiel
17th January 2012, 13:34
Thanks for this test build, I assume you're now using 0.9.x or the latest git master of ffmpeg now because this build at least in my tests has improved seeking in H.264 greatly, before when I would use LAVFilters in combination with madVR there would be a bit of a lag when you tried to seek, but now it's instantaneous and I noticed 'update ffmpeg' a few times on google code.

I usually update quite frequently (always against master), typically at least once shortly after a LAV release (so that any bugs that might come from ffmpeg can be resolved until the next release).

nevcairiel
17th January 2012, 16:37
Nev - bad playback http://www.megaupload.com/?d=2KAAU7RR VC1 1080p60. MPC-HC also bad ))

CUVID decides to not use DXVA for that sample and call the GPU directly, who knows whats wrong with that. Maybe i have some time another day to look.

Remicade
17th January 2012, 18:55
DRC in LAV Audio works only in AC3 ? DTS don't ? It's very annoying to watch movies night.

STaRGaZeR
17th January 2012, 18:58
I usually update quite frequently (always against master), typically at least once shortly after a LAV release (so that any bugs that might come from ffmpeg can be resolved until the next release).

What "ffmpeg" do you use anyway? ffmpeg, libav, or a mix?

nevcairiel
17th January 2012, 19:29
What "ffmpeg" do you use anyway? ffmpeg, libav, or a mix?

ffmpeg and some of my own patches.

STaRGaZeR
17th January 2012, 22:58
ffmpeg and some of my own patches.

Nice. If you do right-click in the ffmpeg folder and use TortoiseGit's "Show log" feature, your custom patches are all those above the origin/upstream tag? Git noob here :p

BTW I was bored and since I couldn't multithread the Fraps decoder (there are repeated P frames in there and stuff), I just multithreaded the DNxHD decoder. Couldn't have been easier. On the 598MB sample here (http://www.limpans.com/DNxHD%20Samples.htm), I get this when benchmarking:

1 thread: ~28 fps and around 25% CPU usage. Real time playback is impossible, too slow.
4 threads: ~144 fps and around 100% CPU usage. More than 4x speedup, smooth playback is possible.

Tested on my 4 core 4 thread 2500K. Here's a build in case anyone wants to try it: http://www.mediafire.com/?lvvq1bl4d63ad2l

What's the best way to get this reviewed and eventually commited to ffmpeg?

nevcairiel
17th January 2012, 23:06
What's the best way to get this reviewed and eventually commited to ffmpeg?

http://ffmpeg.org/developer.html#Submitting-patches

In general, just make sure it looks good, then do git format-patch, and send it to the mailing list with a comment or two.
Running fate is usually a good idea, if the codec even has fate.. but the developers would probably do it before commiting it.

boyumeow
18th January 2012, 06:42
Hi, I'm confuse, should my
GeForce GTX 550 Ti, GF116, VP4, C
be able to play
Video
ID : 1
Format : MPEG-4 Visual
Format profile : Advanced Simple@L5
Format settings, BVOP : Yes
Format settings, QPel : No
Format settings, GMC : No warppoints
Format settings, Matrix : Default (H.263)
Codec ID : XVID
Codec ID/Hint : XviD
using CUVID, cause I can't...
Thanks.

Zachs
18th January 2012, 06:56
I've recently made the switch from AC3Filter to LAV audio but found it to be lacking a few things that the former has.

1) Equalizer

I need to boost the high frequencies a little to compensate for my satelite speakers' lack of tweeter.

and

2) night-mode

where you can dumb down the loudness of explosion and all but keep the dialogue volume the same. In AC3Filter, this is adjustable and applies to all sources (i.e. not just AC3 Dynamic Range Compression).

Is there an equivalent way of achieving the above with LAV Audio?

thomaz909
18th January 2012, 06:57
hi nevcairiel,

im using dvbviewer together with lavvideo decoder and lav audio decoder.

sometimes after "discontinuities" happen when watching mpeg2 sdtv audio and video gets async.

the same problem happens with ffdshow.

with mainconcept mpeg2 decoder, bitcontrol and etymonix mpeg2 decoder the async issue doesnt happen.

ffdshow and lavfilters show picture errors (blocks) when a "discontinuitie" happens.
with mainconcept, etc the picture freezes and skips but no block errors are displayed.

im running lav filters in software mode (ati card) with yadif enabled and set to 50/60fps.

i hope this gets fixed cause i love lavfilters :)

Superb
18th January 2012, 12:47
Hi, I'm confuse, should my

be able to play

using CUVID, cause I can't...
Thanks.The next version of LAV Filters will be able to decode MPEG-4 ASP via CUVID.

dukey
18th January 2012, 12:47
Is there any way to increase the max playback rate under directshow ? For fast forwarding. I am using lav splitter, lav audio/video decoder and EVR. My max playback rate seems to be locked at 2x. Yet for fast forwarding DVDs I can get 16x. Is this the video renderer telling downstream the max rate ? Ie how many actual frames it can draw. It would be nice if we could fast forward, the same as we can with DVDs. Something like muting the audio past 2x, and skipping frames with video should do the trick ..

iSunrise
18th January 2012, 12:47
I've recently made the switch from AC3Filter to LAV audio but found it to be lacking a few things that the former has.

1) Equalizer

I need to boost the high frequencies a little to compensate for my satelite speakers' lack of tweeter.

and

2) night-mode

where you can dumb down the loudness of explosion and all but keep the dialogue volume the same. In AC3Filter, this is adjustable and applies to all sources (i.e. not just AC3 Dynamic Range Compression).

Is there an equivalent way of achieving the above with LAV Audio?
WRT to night-mode, there is an option to DRC (dynamic range compression) some output streams that support it under the Audio settings tab. Never tried it though.

Apart from the options in the Audio settings tab, LAV currently cannot do post-processing like AC3filter or the ffdshow post-processor filter does.

There is a fair share of people who asked nevcairiel before, but I´m not sure if it´s still on his todo list to add another tab in the lav audio filter and put in some post-processing stuff like down-mixing to stereo, for example.

You can use the ffdshow post-processor filter in your chain and achieve what you want, even though I would prefer to use LAV filters for that, because I don´t use ffdshow anymore, nevcairiel already made it 99% obsolete for me.

ryrynz
18th January 2012, 12:51
I've recently made the switch from AC3Filter to LAV audio but found it to be lacking a few things that the former has.

1) Equalizer

I need to boost the high frequencies a little to compensate for my satelite speakers' lack of tweeter.

and

2) night-mode

where you can dumb down the loudness of explosion and all but keep the dialogue volume the same. In AC3Filter, this is adjustable and applies to all sources (i.e. not just AC3 Dynamic Range Compression).

Is there an equivalent way of achieving the above with LAV Audio?

You can do both those things with ffdshow's audio processor under the volume and equalizer settings if you enable output to it via your media player.

Ahh iSunrise beat me to it while I was playing around. :D

Most sound card drivers tend to a configuration with equalizer settings, you could look into that and play with the DRC option in LAV.

nevcairiel
18th January 2012, 13:20
Only post-processing i will offer is a mixer, all the other features i have absolutely zero interest in supporting.

Mercury_22
18th January 2012, 13:38
Only post-processing i will offer is a mixer, all the other features i have absolutely zero interest in supporting.

Yes but will please, please.... :D add independent volume for each channel (like in FFD) too, cause there are a lot of movies/videos with surround channels with too lower volume or "vice versa", or people have different speakers (see above :) ) or....:p

jos99
18th January 2012, 13:43
The CUVID decoder should be able to use that, yes.
Its only partial acceleration, but better then nothing.
Hmm...when I remove all internal codecs and filters in MPC-HC, set LAV Video Decoder and LAV Splitter Source as preferred in external filters, MPC decoder just shows up when I try to play WMV9/VC1 MP content, even if I block it, any ideas?

nevcairiel
18th January 2012, 13:52
Yes but will please, please.... :D add independent volume for each channel (like in FFD) too, cause there are a lot of movies/videos with surround channels with too lower volume or "vice versa", or people have different speakers (see above :) ) or....:p

If there is a full matrix, you can do whatever you please (except increase the overall gain above 100%, because that would have a potential for clipping).

Keiyakusha
18th January 2012, 13:59
Am I doing something wrong or currently there is no way to output YV24 when decoding h264 4:4:4 8 or 10 bit? It outputs only AYUV which is not supported by avisynth or virtual dub.

nevcairiel
18th January 2012, 14:00
Am I doing something wrong or currently there is no way to output YV24 when decoding h264 4:4:4 8 or 10 bit? It outputs only AYUV which is not supported by avisynth or virtual dub.

Did you see a checkbox for YV24 in the formats configuration? No? There you have it! :)
I only implemented the most common formats, which are the recommended formats by Microsoft.

However, supporting YV24 would most likely be rather easy, because its a simple enough format.

Keiyakusha
18th January 2012, 14:04
Aha ok. No problems, I still can open files in virtual dub through hacky 3rd party ffmpeg-based plugin and save them to yv24. But hey, are you saying that AYUV is more common than yv24? Is there anything except LAV that supports it (and benefits from it)? ^__^

However, supporting YV24 would most likely be rather easy, because its a simple enough format.
Well, I'm not going to request this as a feature. I don't really need this on daily basis. Just now I need to process exactly this kind of content.

nevcairiel
18th January 2012, 14:21
AYUV is the recommended format by Microsoft for 8-bit 4:4:4, and madVR accepts it, as does EVR (although EVR depends on hardware support, which is usually not very good).

Adding YV24 support is probably a matter of half an hour or so, i'll put it somewhere on my list of things to do when i need something quick. :)

jos99
18th January 2012, 14:38
Hmm...when I remove all internal codecs and filters in MPC-HC, set LAV Video Decoder and LAV Splitter Source as preferred in external filters, MPC decoder just shows up when I try to play WMV9/VC1 MP content, even if I block it, any ideas?
Doh, solved, WMV3 & VC1 weren't ticked on LAV Formats tab, same for LAV audio, wma and wmapro were unticked (using Shark codec pack).

CruNcher
18th January 2012, 15:32
Lav wouldn't it be a good idear to merge AV Splitter and Lav Splitter resources into 1, both of you seem to reach for the same goal :)
Av Splitter feels a tad more stable has Program Switching (a custom graph system,including mediatype manipulation) for .ts and Lav Splitter has some more neat features combining both would be a good idea i guess :)
Though of course Mr Kytmanov would need to open up his code completely in the first place :)

nevcairiel
18th January 2012, 15:46
Not going to happen.
I only work open source, and with people that respect and endorse open source, both of which don't apply here.

Also, I don't want any graph manipulation things, that's not the splitters job.

CruNcher
18th January 2012, 16:06
Sad i found the idea of this German/Russian Collaboration exciting though i fully understand your points and as said obviously Mr Kytmanov would have to wakeup maybe starting to contributing to lav splitter would be a good start ;)

PS: Though i guess still Lav Splitter leads in overal interoperability recent test show that it's still ahead in playback efficiency also Av Splitters stream switching fails on many Streams in the test database even fails @ detecting programms sometimes @ all :) Also it has a lot of Latm issues that Nev can perfectly fix on both sides having the Audio decoder also under his control ;).
And because of this there are much more unsync situations as well which require to change the whole decoding chain just because of 1 file gets unsync with this decoder using another decoder its fine but another stream gets unsync that worked with the first ;) (this is practically eliminated in Lav Splitter and most video audio decoder combinations especially lav video + lav audio + coreavc + ffdshow + cyberlink).

RBG
18th January 2012, 17:48
CruNcher
Can you, please, upload any video samples, on which you encounter problems with AV. By the way, what version of AV do you use in your testing?

nevcairiel
I don't know, maybe you were asked this question few times before, but do you plan implementing mkv segment linking support to your splitter in the future? That surely is a nice feature supported by both AV and Haali, and it is sad that LAV is some kind of outsider here.:(

nevcairiel
18th January 2012, 17:55
Honestly, its a useless feature, not a nice feature. :p
It will be implemented some day, but don't hold your breath.

CruNcher
18th January 2012, 17:59
CruNcher
Can you, please, upload any video samples, on which you encounter problems with AV. By the way, what version of AV do you use in your testing?

nevcairiel
I don't know, maybe you were asked this question few times before, but do you plan implementing mkv segment linking support to your splitter in the future? That surely is a nice feature supported by both AV and Haali, and it is sad that LAV is some kind of outsider here.:(

1.2.0.6

they're a lot of issues (also some which it shares with lav splitter best example the mp4/mov freeze issue, most probably a core ffmpeg,libavformat issue) though im not putting the same effort into 2 different projects with the same goal (based on the same core) either they become 1 and share the same principals of the core they use or i have to decide for 1 and this decision is fairly clear for me for now (and it wont be the dark side ;) ) :)

MasterNobody
18th January 2012, 18:24
nevcairiel
Why there is no FFV1 support in LAV Video (I didn't found it on formats tab)? I thought it is rather easy to support it when you use ffmpeg. So why it is not there?

adam777
18th January 2012, 18:41
If there is a full matrix, you can do whatever you please (except increase the overall gain above 100%, because that would have a potential for clipping).

Just when I was about to ask if there are some news regarding this issue.
Glad to see it's still somewhere on the to-do list :)
(Oh, I'm totally ignoring the "if" at the beginning :p)

Pat357
18th January 2012, 18:44
Only post-processing i will offer is a mixer, all the other features i have absolutely zero interest in supporting.
By a mixer, I suppose you mean a 7+1 channel matrix or what did you have in mind ?
With a matrix based solution (maybe with presets for 2.0, 5.1 downmix), I think everybody could do it's own thingie :D

RBG
18th January 2012, 19:12
CruNcher
Dark side, lol...:D I have to agree with you, merging both projects would be great especially for end users, one splitter to rule them all:D, but sadly I must admit, that is something of the plane of dreams.:(

nevcairiel
Personally, I don't think it's useless, segment linking is relatively common among anime society. But again, nobody cares about anime lovers...

nevcairiel
18th January 2012, 19:14
common doesnt make useful. :D

Midzuki
18th January 2012, 19:43
Not to mention that presently most animés are N0T worth watching :devil: , to begin with ;) :) :p :D

Pat357
18th January 2012, 19:56
1.2.0.6
they're a lot of issues (also some which it shares with lav splitter best example the mp4/mov freeze issue, most probably a core ffmpeg,libavformat issue)
I thought these issues were fixed in the latest Lav version(s)
Can you upload a mov/mp4 that still freezes with LAV 0.44 ?
Thanks ! :D
PS: just to quickly rule out the core : do these mp4/mov's also crash in FFplay ? Output FFprobe ? Also freeze in Mplayer / Mplayer2 / Avplay (libav only) ?

RBG
18th January 2012, 20:25
Midzuki
Don't be so stodgy, I've got plenty of old good titles in my collection, where segment linking is being used, eg Clannad, Darker than Black, Kanon, and I still want to have perfect seamless playback of them. Also if this year was poor for worthy anime, that doesn't mean in any way that next year will be the same.;)

Andy o
18th January 2012, 20:51
Is that such a big deal? I've watched a couple of cartoons with that feature for openings/endings, and why would anyone want to watch an identical opening/closing every episode? Is there any other way it is widely used?

clsid
18th January 2012, 20:57
You could just continue to use Haali splitter for MKV and use LAV Splitter for everything else.

RBG
18th January 2012, 21:20
Andy o
They are not always identical, that's the point. Also I prefer to have full episode and not cut version and it is better when you yourself can decide to watch or not to watch op/ed, you can always rewind them, gladly there are markers on the time line of potplayer.

clsid
That could be an option, if it wasn't so damn annoying to switch between splitters manually, especially when you're watching something for the first time.

CruNcher
18th January 2012, 22:11
ehh you just register haali disable all the other plugins leave only mkv and then register lav splitter and disable mkv parsing :) the only problem is you need decoder DXVA/Software that work fine in every case with Haalis Splitter (and every supported format haali supports) and bug fixing is problematic most probably the decoder has to adapt but that's also problematic if you think about 3rd party decoder :)
But the Manually switching that really was yesterday :)

Nope the MP4/MOV (H.264 only so far) problem isn't fixed though the streams can be fixed by remuxing them though not every muxer fixes those freezeing streams yet most probably that's also why they have appeared after editing. The Origin for most of these streams seems to be Apple Finalcut :(
Most interesting parser player beahaviour with such a stream is VLC it will reload the stream right in the middle with a small hickup and continue, while ffmpeg based software often just freezing the Video Stream :P

zmaster
18th January 2012, 22:43
A small question about the splitter. It is theoretically possible that the splitter can ignore certain parts with a certain name (the chapter name)?

RBG
18th January 2012, 23:27
CruNcher
Oh my! I need to sleep more, I reread what clsid wrote and understood, that I misread it somehow, shame on me.:o Of course using one splitter for one container format and another for everything else is not a problem at all. I just wanted to say, that there is no universal solution for one format, mkv for example.:( Haali is relatively stable has linked video support, but has its own cons, about them you have already wrote in the post above, also it has no graph builder, which is a noticeable disadvantage for me. AV has a lot of advanced features, including linked video support, graph builder, but it has some bugs, as I use it most of the time, sometimes I am just forced to switch to another splitter to get correct video playback. LAV is relatively stable too, its open sourse, which is great, compatible decoders pack included, but has no linked video support at all.

CruNcher
19th January 2012, 00:04
Yeah i know in what for a situation you are exactly for example something just came up a second ago ;) going through my database finally moved forward from .ts to .mov now i came up to a streams with ima4_adapcm_qt the problem i want to use MPC-HCs Mp4 splitter for .mov because of the (freeze) problems above and now i just found out that tada ima4 is not compatible with MPC-HCs splitter and Lav Audio only MPC-HCs audio splitter connects (so solutions would be have both registered and lower the others merit and hope it wont interfere with anything else, try to make them interoperable or try to fix the mp4/mov freeze) ;)
Linked segment is something i don't use myself also graph building is not really that interesting it's a nice advanced feature same for the mediatype changer the only really interesting thing in AV splitter compared to Lav Splitter for me personally are Programm switching support and proper working raw .(h)264 parsing ;)

Zachs
19th January 2012, 00:28
You can do both those things with ffdshow's audio processor under the volume and equalizer settings if you enable output to it via your media player.

Ahh iSunrise beat me to it while I was playing around. :D

Most sound card drivers tend to a configuration with equalizer settings, you could look into that and play with the DRC option in LAV.

Ah my sound card driver used to have equalizer in XP but it's no longer available in Win7.

How would one go about setting up ffdshow to use with LAV Audio? I'm using MPC-HC.

Edit: nevermind. Found out I could use AC3Filter (select PCM only under "Use AC3Filter for:") as the audio post processor for LAV Audio. I'm using it for EQ and DRC. I could even use it to downmix / upmix! Don't think Nev needs to add post processing capabilities to his filters at all.

CruNcher
19th January 2012, 01:35
@Nev
going through the database for .mov i experienced something odd with Lav Audio ?
@ Decoding PCM sowt Lav Audio showed considerably higher CPU utilization then compared to the MPC-HC Audio Decoder also tried to disable 32 bit floating point and also 32 bit in general and enabled same for MPC-HC Audio no difference (was rendering with the default 16 for both from the start)
i wonder where that difference of Lav Audio = 5%, MPC-HC Audio = 2% (3% Utilization difference) comes from in this decoding case ?

Audio
ID : 2
Format : PCM
Format settings, Endianness : Little
Format settings, Sign : Signed
Codec ID : sowt
Duration : 3mn 30s
Bit rate mode : Constant
Bit rate : 1 536 Kbps
Channel(s) : 2 channels
Channel positions : Front: L R
Sampling rate : 48.0 KHz
Bit depth : 16 bits
Stream size : 38.6 MiB (16%)
Language : English
Encoded date : UTC 2010-01-22 22:09:41
Tagged date : UTC 2010-01-22 22:13:16
Material_Duration : 211064
Material_StreamSize : 40524192

same difference visible for this twos

Audio
ID : 2
Format : PCM
Format settings, Endianness : Big
Format settings, Sign : Signed
Codec ID : twos
Duration : 3mn 55s
Bit rate mode : Constant
Bit rate : 1 411.2 Kbps
Channel(s) : 2 channels
Channel positions : Front: L R
Sampling rate : 44.1 KHz
Bit depth : 16 bits
Stream size : 39.6 MiB (14%)
Language : English
Encoded date : UTC 2011-02-12 13:05:33
Tagged date : UTC 2011-02-12 13:08:10

AAC Decoding doesn't show such a huge difference

nevcairiel
19th January 2012, 07:55
The raw PCM "decoders" are more designed for readability and clarity then for speed.
There are also some features that are only activated with PCM, like the detection of DTS-in-WAV, which might take a bit processing power.

Its no CPU load to worry about.

proper working raw .(h)264 parsing ;)

raw h264 streams play fine for me, they just don't allow seeking, like all raw streams.
Those streams are only intermediates between encodes/muxings, and i don't think its that crucial to be able to seek in them. :p

Its more something that ffmpeg would need to support, not me, anyway. :)


Nope the MP4/MOV (H.264 only so far) problem isn't fixed

Can you dig up a link to such a broken file? I cannot find the one i downloaded. :(

nevcairiel
19th January 2012, 08:17
A small question about the splitter. It is theoretically possible that the splitter can ignore certain parts with a certain name (the chapter name)?

No, thats not possible.
Chapters are just markers on the stream, like bookmarks, they are no logical units.

In theory you could try to jump to the next chapter of course, but it wouldn't be accurate or seamless.

e-t172
19th January 2012, 08:36
Ah my sound card driver used to have equalizer in XP but it's no longer available in Win7.

Sounds like you have a Creative card. In that case you can get back equalization and a bunch of other features by installing the unofficial drivers by Daniel_K.

zmaster
19th January 2012, 15:29
No, thats not possible.
Chapters are just markers on the stream, like bookmarks, they are no logical units.

In theory you could try to jump to the next chapter of course, but it wouldn't be accurate or seamless.Thanks for reply.

Midzuki
19th January 2012, 16:59
Hummm, LAV Splitter apparently does not use "MEDIASUBTYPE_MPEG1AudioPayload". As a consequence, it cannot connect to the MPG123 decoder (mpg123dsf.ax), for example.

CruNcher
19th January 2012, 18:35
The raw PCM "decoders" are more designed for readability and clarity then for speed.
There are also some features that are only activated with PCM, like the detection of DTS-in-WAV, which might take a bit processing power.

Its no CPU load to worry about.



raw h264 streams play fine for me, they just don't allow seeking, like all raw streams.
Those streams are only intermediates between encodes/muxings, and i don't think its that crucial to be able to seek in them. :p

Its more something that ffmpeg would need to support, not me, anyway. :)



Can you dig up a link to such a broken file? I cannot find the one i downloaded. :(

This is strange i can definitely seek with AV Splitter in any raw file (ffmpeg), with Lav Splitter (ffmpeg) the seek bar is just grey (disabled) so i doubt Kytmanov fixed it in ffmpeg though i guess he has the ability todo it and then it would be really sad if he didn't contributed that back :(

Av Splitter

http://img37.imageshack.us/img37/640/avsplitterraw.png

Lav Splitter

http://img267.imageshack.us/img267/9241/lavsplitterraw.png

http://forum.doom9.org/showpost.php?p=1534773&postcount=15
http://forum.doom9.org/showpost.php?p=1534774&postcount=6567

On the yahoo trailer its @ the cut point (introduction,cut,trailer) seems to be always @ a cut (Finalcut) point on the sample in my video its also @ a cut point (first scene,cut,next scene) (Finalcut)

here is the issue propagated to .mp4 http://www.mediafire.com/?yx92afb8mrn7r3r (mp4box) it looks identical in what happens @ least (freeze)

I thought these issues were fixed in the latest Lav version(s)
Can you upload a mov/mp4 that still freezes with LAV 0.44 ?
Thanks ! :D
PS: just to quickly rule out the core : do these mp4/mov's also crash in FFplay ? Output FFprobe ? Also freeze in Mplayer / Mplayer2 / Avplay (libav only) ?

yes they freeze several 3rd party applications based on ffmpeg (many of the all in one converters, video editors ect)
though i have to admit i didn't tried it with the latest trunks of libav or ffmpeg libav alone yet but earlier freezed since i first reported it way back, though what i hoped for is that feedback finds it way back to ffmpeg through lav splitter as well (dynamically) this is why i didn't reported it directly yet i agree that most probably was wishful thinking and in retrospect a mistake (i really hoped for it). Also since the project splitt it got more chaotic as you knew that these software was using ffmpeg now you need to make sure first what it uses ;)

http://win32.libav.org/win32/libav-win32-20110318.7z <- still freezes

http://ffmpeg.zeranoe.com/builds/win32/static/ffmpeg-latest-win32-static.7z <- yep also freezes

just to make sure

ffmpeg:

G:\aero-solveig-all-60fps.264: error while seekingsq= 0B f=0/0 0/0

libav:

also fails with seeking raw

So it seems Kytmanov indeed did some own fix here or uses some experimental fix in Av Splitter that works really well :)

aufkrawall
19th January 2012, 19:14
Will it be possible to use QuickSync to accelerate decoding of 4k video also with Sandy Bridge? Or just with Ivy Bridge?

CruNcher
19th January 2012, 19:26
most probably only with Ivy Bridge even if it would be possible with SB i doubt Intel would disable the Software Restriction because of certain implications
It's rare when Restrictions are disabled last time i remember it was for the Sony PSP when UMD wasn't as commercial successful as Sony imagined it so they finally removed the resolution limit from 640 to the full capable full SD (UMD) for any Stream (though before hackers circumvented this already it was a little adventurous workaround anyways creating a own file format hehe), also after many users where unhappy that they couldn't playback their Apple Quicktime Movs and only being restricted to Sonys .mp4 they reacted after the protest was heavy enough to hurt their sales ;)

So don't hope for it especially for 64bit it's not easy anymore manipulating a driver anymore or you endup in a F8 session every time (though for how long this will still work is also not clear) ;)
With Secureboot on ARM (though also different ARM Firmware protection schemes are already broken) and upcoming devices it will be even much harder and devices fully under control of who developed them those will be hard times. and only the smartest of the smartest low level geniuses that know up to the Bios (Uefi,TPM) level with cryptography experience will survive (and i don't speak about the new .Net or Java zombie Generation growing up or better been grown up ;) ), though this started already with (Pay TV, though companies weren't low level getting sued is still the nice way if you make something public, they even killed genius people in the past to protect their protection schemes and systems) moved to the Xbox,PSP,PS3 and moved on to Smartphones and its continuing to the Desktop PC.

Of course we will benefit in Security from it but also Device Restrictions will get secured with it more and more and Open Source will get threatened by it even if everyone says we will allow it it is a major problem in this new future, it will tough also open up new interesting ways see Intels Chip Research you buy only 1 Chip and later you enable what you need with a code and get more performance for extra money this will also have a major impact on the Economy itself as yields get better it will help to avoid to much production costs developing a line of 30 different chips instead their will be only 1 chip where you decide what you want to pay for extra you can virtualy build your own chip deciding if you need VC-1,H.264,Mpeg-2 whatever any instruction (SSE, AVX) could be restricted Cores Mhz you name it it could be done in the future, and evaluation wise it is already being done by Intel) :)
So everything will be on the Chip allready only 1 production line 1 time the cost but due to the restrictions it will allow to create a tremendous efficient Ecosystem, of course there need to be evaluated how do these products will look like and how do you market certain features to costumers (will there be base configs,do i need certain instructions and for what, where does it help me in my current usage ect) but all this is currently researched on as we speak ;)

aufkrawall
19th January 2012, 19:31
most probably only with Ivy Bridge even if it would be possible with SB i doubt Intel would disable the Software Restriction because of certain implications

Maybe just a patched version of the ffdshow QuickSync decoder would be needed. :)
Since it's open source I don't see any legal conflicts.

CruNcher
19th January 2012, 20:30
Nope also Egur can only do what others allow him todo ;)

Zachs
20th January 2012, 00:07
Sounds like you have a Creative card. In that case you can get back equalization and a bunch of other features by installing the unofficial drivers by Daniel_K.

Actually it's Realtek crap that comes with the mobo.
I've looked around but Realtek said EQ is no longer supported under Win7.

ranpha
20th January 2012, 00:57
Actually it's Realtek crap that comes with the mobo.
I've looked around but Realtek said EQ is no longer supported under Win7.

There is an EQ in Realtek drivers in Windows 7. Your hardware probably doesn't support it?

CruNcher
20th January 2012, 01:51
There is an EQ in Realtek drivers in Windows 7. Your hardware probably doesn't support it?

yep though its made for the avg joe so you are first greeted with the user friendly GUI which is basically the choice of some presets (there are more in the drop down menu) on the side of that dropdown is a little Equalizer button pushing that makes the EQ visible.

http://img854.imageshack.us/img854/6421/realtekeq.png

Pat357
20th January 2012, 02:39
raw h264 streams play fine for me, they just don't allow seeking, like all raw streams.
Those streams are only intermediates between encodes/muxings, and i don't think its that crucial to be able to seek in them. :p:(
I could use some advise concerning playback from raw .264/h264/x264 streams. Please give me good tip guys.
This is wat I have : playback from raw h264 streams using Lav-filter + lav video gives me a black screen in MPC-HC.
(File async, lav-spiltter, lav-audio and lav-video as "prefered", all internal stuff disabled)

I also created a graph in Graphstudio like :
asyn file -> lav-splitter-> lav-video -> EVR. Very simple and should work, but it doesn't, again the same black screen.
The info about the source (fps, higth, width, ratio,..) is properly passed from the source -> lav-splitter -> lav video -> evr.
Oh, I have enabled "raw video" in the lavsplitter tab after the lavfilter installation, not during installation.

I'm starting pulling out my hair because every time I create a raw .x264, I want to have a look at it before I start muxing the rest, and I bump against this issue.
If I put a .x264 that gives only a black screen in a container (MKV, MP4, whatever) it plays normal ! So the stream itself is OK, I guess.
If I replace lav-splitter with for ex. AVsplitter, Sonic HD demuxer or any other capable splitter, everything plays fine (still using Lav-video !).
The problem I have seems somehow related to Lav-spliiter, maybe in combination with something else.
All Lav-filters have their original merit (80000, 80004) and no other filter has a higher merrit in Dshow. (others or all below lav-video/audio).

How can it be that a raw x264 file does not play, but as soon as I put it in a container (which doesn't matter, mkv, mp4,..are all ok) it plays perfect !
It seems that something (filter with a lower merit) is highjacking all extensions for raw x264 files....
Your advise please.
System : Lav-filters 0.44, MPC-HC v3970, Nvidia gtx-570, Intel 970 @ 3.6 , Win 7 prof x64 (all updates).

mindbomb
20th January 2012, 03:05
i have a question.

I'm playing around with dxva copy back and rgb 32 output in lav video.

I'm noticing that interlaced video is being deinterlaced. How is this happening? I don't have yadif enabled. Is lav video doing it or can evr cp deinterlace rgb32 video?

goldie
20th January 2012, 06:26
@mindbomb
You can check whether it's deinterlaced in renderer or not by temporarily switch renderer from EVR to VMR or something else (overlay?).
Then check if there is a another deinterlaced-feature-enabled filter in your directshow filter chain (mayde is ffdshow).

ryrynz
20th January 2012, 06:56
ffdshow Quick Sync

http://img31.imageshack.us/img31/2531/ffdshowj.jpg

LAV Quick Sync

http://img860.imageshack.us/img860/485/lavki.jpg

Oh yeahh, grats Nev.

goldie
20th January 2012, 08:07
@ryrynz
Where can get the Decoder Performance Test tool?

wanezhiling
20th January 2012, 08:36
http://code.google.com/p/graph-studio-next/downloads/list
view -> Decoder Performance Test

goldie
20th January 2012, 09:02
http://code.google.com/p/graph-studio-next/downloads/list
view -> Decoder Performance Test
:thanks:

NikosD
20th January 2012, 09:10
@Nevcariel

For Intel's hardware only (QuickSync):

1) Have you managed to implement a DXVA VC1_VLD mode ?

2) Have you managed to implement DXVA VC-1 interlaced mode in either VC1_VLD or VC1_IDCT ?

3) Have you managed to accelerate H.264 in L5.1 and ReF>8 ?

nevcairiel
20th January 2012, 09:21
@Nevcariel

For Intel's hardware only (QuickSync):

1) Have you managed to implement a DXVA VC1_VLD mode ?

2) Have you managed to implement DXVA VC-1 interlaced mode in either VC1_VLD or VC1_IDCT ?

3) Have you managed to accelerate H.264 in L5.1 and ReF>8 ?

With the QuickSync decoder, yes to all (only VLD is supported for everything, no partial accelerations)

With pure DXVA, VC-1 does not work at all, and i didn't check such H264 files.
I generally recommend to use the QuickSync decoder, anyhow. :)

NikosD
20th January 2012, 09:31
I was referring to pure DXVA, I know that QS decoder with the help of Intel's MSDK can do all.

If you try H.264 with Ref>7 maybe lower, you will see that they are falling back to software mode.

UPDATE:
Are you ready for a pure DXVA mode for all hardware (ATI, Nvidia, Intel) soon ?

nevcairiel
20th January 2012, 09:34
Eric is planning to add a direct mode to the QuickSync decoder as well, which means it could run without the copying overhead and still be able to play everything. Because of that, i do not plan to spend a whole lot of time trying to get Intels special DXVA mode working.

I'm not sure yet when i'll get to finish pure DXVA2 otherwise, or if i should release the current state before that. Time is running a bit thin right now.

Thunderbolt8
20th January 2012, 09:48
when I use lav video decoder with dvxa2 (test3 build), then my picture stays black, theres only sound and subtitles to be seen. as soon as I untick dxva2, its fine again. how come?

nevcairiel
20th January 2012, 11:30
when I use lav video decoder with dvxa2 (test3 build), then my picture stays black, theres only sound and subtitles to be seen. as soon as I untick dxva2, its fine again. how come?

Its possible that your card doesn't support that specific file.
Do other DXVA decoders work on that?

A MediaInfo of that file may be helpful.

My code doesn't do any card-specific compat checks right now, it only checks for general compatibility (so only high profile, etc).

With recent drivers, do ATI/AMD still need checks on older cards regarding level and ref-frames? Do i need to implement this?

goldie
20th January 2012, 11:38
I have do some test with LAV Filter test3 build on my laptop.

LAV Filters - None (software) mode
301 .. 313 fps
http://i.imgur.com/R9u3C.png

LAV Filters - Intel QuickSync mode
231 .. 233 fps
http://i.imgur.com/WGJPI.png

LAV Filters - DXVA (copy-back) mode
123 .. 124 fps
http://i.imgur.com/QtN7l.png

The software mode on my laptop is fastest but it made the fan too loud, so I prefer the hardware acceleration mode.
And I'm very interesting that DXVA is slower than Intel QuickSync, thanks for nev and egurs' great work. :)
But the Intel QuickSync on my laptop has problem that decoded frames are out of order, result in unusable.
It may can be fixed by update graphic driver (I saw some dicussions about this problem in previous threads here), but I have to wait for laptop manufacture release since latest Intel official driver said it could not be installed please intall the one from manufacture instead; consequently, I now have to stick at DXVA mode for usually video playback. :(

System info:
Chipsset: Intel HM67
CPU spec: Intel Core i7-2640M 4MB-L3cache 2.8GHz (Turbo 3.5GHz)
GPU spec: Intel HD 3000
GPU driver version: 8.15.10.2418
Memory info: Dual-channel DDR3-1333 8GB

nevcairiel
20th January 2012, 11:57
120 fps is still fast enough to play any current content, so don't sweat it. :p

NikosD
20th January 2012, 12:05
With recent drivers, do ATI/AMD still need checks on older cards regarding level and ref-frames? Do i need to implement this?

HD 5000 series and above can play everything, in general, you don't need to implement any checks for ReF or Level.

HD 3000 series and below, can't play anything beyond BluRay spec, last time I tested.

HD 4000 series, most of them use UVD2.2 like HD 5000 and a few use UVD2.0, but the "hardware/drivers" should be able to play everything or almost everything.
ATI says that UVD2.2 inside HD 4000 series is a little different than UVD2.2 inside HD 5000 series, but who really knows.

Don't waste your time with those kind of checks.

when I use lav video decoder with dvxa2 (test3 build), then my picture stays black, theres only sound and subtitles to be seen. as soon as I untick dxva2, its fine again. how come?

Can you upload somewhere the sample in order to test it with my 5750 ?

chewi
20th January 2012, 13:27
Hello!
Great thanks to the authors for LAV Filters. If you allow, I have a suggestion/wish: the possibility of simultaneous playback of two subtitle' streams (sewn in mkv) by LAV-splitter. My family and me usually watch foreign films without the voice translation - using the original track and translated subtitles (it really helps in learning foreign language). But often, when the voice translation is not very good or not very clear pronunciation — really need to be able to display simultaneously 2 subs (translated and original).

Is it possible to add this functionality to LAV splitter?
Thanks in advance.

nevcairiel
20th January 2012, 13:50
That wouldn't be all that useful, they would show in the same spot and overlap each other.

CruNcher
20th January 2012, 14:36
I could use some advise concerning playback from raw .264/h264/x264 streams. Please give me good tip guys.
This is wat I have : playback from raw h264 streams using Lav-filter + lav video gives me a black screen in MPC-HC.
(File async, lav-spiltter, lav-audio and lav-video as "prefered", all internal stuff disabled)

I also created a graph in Graphstudio like :
asyn file -> lav-splitter-> lav-video -> EVR. Very simple and should work, but it doesn't, again the same black screen.
The info about the source (fps, higth, width, ratio,..) is properly passed from the source -> lav-splitter -> lav video -> evr.
Oh, I have enabled "raw video" in the lavsplitter tab after the lavfilter installation, not during installation.

I'm starting pulling out my hair because every time I create a raw .x264, I want to have a look at it before I start muxing the rest, and I bump against this issue.
If I put a .x264 that gives only a black screen in a container (MKV, MP4, whatever) it plays normal ! So the stream itself is OK, I guess.
If I replace lav-splitter with for ex. AVsplitter, Sonic HD demuxer or any other capable splitter, everything plays fine (still using Lav-video !).
The problem I have seems somehow related to Lav-spliiter, maybe in combination with something else.
All Lav-filters have their original merit (80000, 80004) and no other filter has a higher merrit in Dshow. (others or all below lav-video/audio).

How can it be that a raw x264 file does not play, but as soon as I put it in a container (which doesn't matter, mkv, mp4,..are all ok) it plays perfect !
It seems that something (filter with a lower merit) is highjacking all extensions for raw x264 files....
Your advise please.
System : Lav-filters 0.44, MPC-HC v3970, Nvidia gtx-570, Intel 970 @ 3.6 , Win 7 prof x64 (all updates).

yep same here lot of bitstreams are just black those work with Av Splitter, so the only solution so far would be combining Lav and AV splitter by managing it the way that Lav Splitter doesn't handle raw and Av Splitter only handles raw, so you go in Av Splitter in the advanced options disable everything except rawvideo and for Lav Splitter you disable rawvideo (which by default should be disabled anyways) voila ;)

chewi
20th January 2012, 15:42
That wouldn't be all that useful, they would show in the same spot and overlap each other.

Why only "in the same spot"? One sub - at the top, another - at the bottom of the screen; this way I mean.

ajp_anton
20th January 2012, 16:19
HD 3000 series and below, can't play anything beyond BluRay spec, last time I tested.I've played 16-ref 1080p with DXVA (MPC-HC) on my integrated HD3300.

NikosD
20th January 2012, 16:49
Drivers ?
OS ?

I have a lot of 16 ReF H.264 1080p files and a 3650 I could try.

wanezhiling
20th January 2012, 16:50
Drivers ?
OS ?

I have a lot of 16 ReF H.264 1080p files and a 3650 I could try.

http://forum.doom9.org/showthread.php?t=163652
Try to set DXVA to "skip all checks" in MPC

NikosD
20th January 2012, 17:06
No, we aren't talking about HD 4000 series.

We are talking about HD 3000 series.

It's UVD2.x vs UVD+.

wanezhiling
20th January 2012, 17:23
I mean maybe you can try this way on your HD3650.:)

HeadlessCow
20th January 2012, 17:27
Why only "in the same spot"? One sub - at the top, another - at the bottom of the screen; this way I mean.

Because subtitles contain positioning.

nevcairiel
20th January 2012, 17:28
Because subtitles contain positioning.

Exactly, i cannot control the positioning of the subtitles.

Paladin77
20th January 2012, 17:30
I mean maybe you can try this way on your HD3650.:)
I am interested to know as well as DXVA2 copyback shows available but is unusable and painfully slow on the 3430 system. slow copy back I guess

NikosD
20th January 2012, 17:53
I've played 16-ref 1080p with DXVA (MPC-HC) on my integrated HD3300.

I am interested to know as well as DXVA2 copyback shows available but is unusable and painfully slow on the 3430 system. slow copy back I guess

I mean maybe you can try this way on your HD3650.:)

I have done a lot of tests with Windows Xp - Catalyst 11.2 - Radeon 3650, you can find here. (http://forum.doom9.org/showthread.php?t=159486) Look at the second system results.

That's why I asked ajp_anton about drivers version and OS.

Maybe after almost a year of my last try with 3650, I could try again under Win 7 system with latest Catalysts 11.12

Thunderbolt8
20th January 2012, 18:36
Its possible that your card doesn't support that specific file.
Do other DXVA decoders work on that?

A MediaInfo of that file may be helpful.

My code doesn't do any card-specific compat checks right now, it only checks for general compatibility (so only high profile, etc).

With recent drivers, do ATI/AMD still need checks on older cards regarding level and ref-frames? Do i need to implement this?I tested one AVC and a VC-1 file, it was the same in both cases.
still using ATI drivers 11.5, dunno whether its recommend to update with my rather old card or which version is supposed to be best for the 4770

STaRGaZeR
20th January 2012, 18:40
I am interested to know as well as DXVA2 copyback shows available but is unusable and painfully slow on the 3430 system. slow copy back I guess

Expected behavior, its Merom core doesn't support SSE 4.1. Maybe it's a good idea to change copy back as not available on CPUs older than Penryn, since it'll be painfully slow otherwise.

I wonder about the speed on AMD CPUs. But since SSE 4.1 is only available on Bulldozer, I think I already know the answer :rolleyes:

nevcairiel
20th January 2012, 18:51
On my Sandy Bridge, i didn't really see any big differences between a default memcpy and the sse4 copy function, but i guess it all depends on the CPU/GPU combination.

I'll do the pure dxva at some point in the future. I will however mmost likely release one new version before then

mindbomb
20th January 2012, 19:34
on my amd e350, i couldn't use motion adaptive deinterlacing with 1080i30 video with dxva copyback.

with pure dxva, i could, but all the dxva decoders i have used thus far are terrible, so it is a small price to pay to use just regular adaptive deinterlacing.

Pillars of Creation
20th January 2012, 20:32
I've got a question regarding LAV splitter:

General
Format : Matroska
Format version : Version 2
File size : 2.29 GiB
Duration : 54mn 50s
Overall bit rate : 5 980 Kbps
Encoded date : UTC 2012-01-19 18:23:05
Writing application : mkvmerge v5.1.0 ('And so it goes') built on Nov 28 2011 23:58:28
Writing library : libebml v1.2.3 + libmatroska v1.3.0

Video
ID : 1
Format : AVC
Format/Info : Advanced Video Codec
Format profile : High@L4.1
Format settings, CABAC : Yes
Format settings, ReFrames : 5 frames
Codec ID : V_MPEG4/ISO/AVC
Duration : 54mn 50s
Bit rate : 5 120 Kbps
Width : 1 280 pixels
Height : 720 pixels
Display aspect ratio : 16:9
Frame rate : 25.000 fps
Original frame rate : 23.976 fps
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Progressive
Bits/(Pixel*Frame) : 0.222
Stream size : 2.00 GiB (87%)
Writing library : x264 core 68 r1195 5d75a9b
Encoding settings : cabac=1 / ref=5 / deblock=1:0:0 / analyse=0x3:0x133 / me=umh / subme=7 / psy_rd=1.0:0.0 / mixed_ref=1 / me_range=16 / chroma_me=1 / trellis=1 / 8x8dct=1 / cqm=0 / deadzone=21,11 / chroma_qp_offset=-2 / threads=12 / nr=0 / decimate=1 / mbaff=0 / bframes=3 / b_pyramid=1 / b_adapt=1 / b_bias=0 / direct=1 / wpredb=1 / keyint=250 / keyint_min=25 / scenecut=40 / rc=2pass / bitrate=5120 / ratetol=1.0 / qcomp=0.60 / qpmin=10 / qpmax=51 / qpstep=4 / cplxblur=20.0 / qblur=0.5 / ip_ratio=1.40 / pb_ratio=1.30 / aq=1:1.00
Default : Yes
Forced : No

Audio #1
ID : 2
Format : AC-3
Format/Info : Audio Coding 3
Mode extension : CM (complete main)
Codec ID : A_AC3
Duration : 54mn 50s
Bit rate mode : Constant
Bit rate : 192 Kbps
Channel(s) : 2 channels
Channel positions : Front: L R
Sampling rate : 48.0 KHz
Bit depth : 16 bits
Compression mode : Lossy
Stream size : 75.3 MiB (3%)
Language : German
Default : Yes
Forced : No

Audio #2
ID : 3
Format : AC-3
Format/Info : Audio Coding 3
Mode extension : CM (complete main)
Codec ID : A_AC3
Duration : 54mn 50s
Bit rate mode : Constant
Bit rate : 448 Kbps
Channel(s) : 6 channels
Channel positions : Front: L C R, Side: L R, LFE
Sampling rate : 48.0 KHz
Bit depth : 16 bits
Compression mode : Lossy
Stream size : 176 MiB (7%)
Language : English
Default : No
Forced : No

Which frame rate gets transported to the decoder?

Frame rate : 25.000 fps

or

Original frame rate : 23.976 fps

Splitter displays 23.976 but 25 fps is correct...

nevcairiel
20th January 2012, 20:42
The frame rate is never guaranteed to be accurate.
Usually it would take the container frame rate, however Matroska does not have a container frame rate, so it trys other sources, probably the H264 timing info in its headers.

The frame rate value reported is meaningless, how fast the video is played is defined by its timestamps, and nothing else.

Andy o
20th January 2012, 20:57
So I got around to testing DXVA2 copy back (with your "test3" build posted a few pages back), and it's working fine apparently for VC-1 and H.264, but is it supposed to be working yet for MPEG-2, specifically 1080i in wtv container? It's not getting enabled here.

Pillars of Creation
20th January 2012, 21:03
The frame rate is never guaranteed to be accurate.
Usually it would take the container frame rate, however Matroska does not have a container frame rate, so it trys other sources, probably the H264 timing info in its headers.

The frame rate value reported is meaningless, how fast the video is played is defined by its timestamps, and nothing else.

Thanks for your fast respond.

Nevertheless with lav splitter the video file is played with 23.97x fps and droppes frames and jiggles

Without lav splitter (using standard mpc-hc splitter) the same video file is played with 25 fps and without jiggles. vlc (internal splitter) also plays the video file smooth (displays also 25 fps).

What am I doing wrong? :(

nevcairiel
20th January 2012, 21:11
So I got around to testing DXVA2 copy back (with your "test3" build posted a few pages back), and it's working fine apparently for VC-1 and H.264, but is it supposed to be working yet for MPEG-2, specifically 1080i in wtv container? It's not getting enabled here.

If you're running on your 5770 from your signature, then no. Only 6xxx and above have an enabled MPEG2 decoder, afaik.
The 5000 cards were advertised to have it, but apparently it never worked.


What am I doing wrong? :(
If you're running a refresh rate switcher, it might get it wrong if it only relys on that information. Its not reliable, it never was and never will be (specifically if the container has no fps field)

Andy o
20th January 2012, 21:16
Ah, thanks never knew that about mpeg 2.

Pillars of Creation
20th January 2012, 21:20
If you're running a refresh rate switcher, it might get it wrong if it only relys on that information. Its not reliable, it never was and never will be.

Yes you are right. I usually use madVR to switch my refresh rate. I tried mpc-hc (latest beta) refresh rate switcher but got the same problem.

Thanks for your help.

nevcairiel
20th January 2012, 21:24
With madVR, you can override the choosen refresh rate by putting a refresh rate token into the filename, like "filename.25p.mkv"

wanezhiling
20th January 2012, 21:28
5xxx is still MPEG2_IDCT...

Only UVD3.0 and above support MPEG2_VLD just like Nvidia VP3/4/5.:rolleyes:

Pillars of Creation
20th January 2012, 21:38
With madVR, you can override the choosen refresh rate by putting a refresh rate token into the filename, like "filename.25p.mkv"

Thank you again, it works.

mindbomb
20th January 2012, 21:52
The 5000 cards were advertised to have it, but apparently it never worked.


the radeon 4000-5000 series radeons have idct, not vld for mpeg 2.

Actually, for cuda and quicksync, idct is working for mpeg 2, right?

Also, i think most equipment that has vld mpeg 2 can also do vld mpeg 4 asp, since they are apparently very similiar.

But yea, they did advertise as full mpeg 2 decoding, and they don't technically do that.

nevcairiel
20th January 2012, 23:03
Actually, for cuda and quicksync, idct is working for mpeg 2, right?

No, only CUDA can do that, QuickSync not (also, QuickSync is only for Sandy Bridge, it has full VLD)


Also, i think most equipment that has vld mpeg 2 can also do vld mpeg 4 asp, since they are apparently very similiar.

Thats not true at all.
MPEG-2 was supported much earlier - just not in radeons. ;)

RBG
21st January 2012, 00:07
Bug report.

I tried Sony demo ts file and encountered a problem with LAV splitter. It shows incorrect duration time and rewind is some kind of broken.

LAV splitter
http://img4.imageshack.us/img4/6686/20120121023824.png

Built-in MPEG2 TS Source
http://img685.imageshack.us/img685/2653/20120121023853.png

To check rewind function just do jump forward after 1 minute 20 seconds point of time line and you'll be send to the real-time 8th minute of the video. Again on Built-in MPEG2 TS Source everything works fine including rewind. Also it seems that field order auto detection in LAV video doesn't work as it should here, at least I had to force "Top-Field First" option to get read of picture doubling artifacts.

Here is the sample (http://www.filefactory.com/file/c2a1371/n/test1.zip) file. It's not original Sony demo, I have just cut a piece of it in a hex editor.

Mediainfo.
Video
ID : 33 (0x21)
Menu ID : 1 (0x1)
Format : MPEG Video
Format version : Version 2
Format profile : High@High
Format settings, BVOP : Yes
Format settings, Matrix : Custom
Codec ID : 2
Duration : 1mn 34s
Bit rate mode : Variable
Bit rate : 130 Mbps
Maximum bit rate : 24.0 Mbps
Width : 1 920 pixels
Height : 1 080 pixels
Display aspect ratio : 16:9
Frame rate : 29.970 fps
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Interlaced
Scan order : Bottom Field First
Compression mode : Lossy
Bits/(Pixel*Frame) : 2.093
Stream size : 1.43 GiB (95%)

Audio
ID : 36 (0x24)
Menu ID : 1 (0x1)
Format : MPEG Audio
Format version : Version 1
Format profile : Layer 2
Codec ID : 3
Duration : 1mn 34s
Bit rate mode : Constant
Bit rate : 224 Kbps
Channel(s) : 2 channels
Sampling rate : 48.0 KHz
Compression mode : Lossy
Stream size : 2.51 MiB (0%)

Mixer73
21st January 2012, 12:08
Hey guys

I've been chasing "ordinary" video quality for a while in my system and I think I might have found the cause.

My configuration is PotPlayer, LAV Filters & MadVR. Typical content is DVR-MS files containing DVB MPEG2 files. The complaint I have is with upscaling 720x576 files.

When I am playing standard def files in PotPlayer with LAV Filters I see resolution in as 720x576 and resolution out as 1024x576 and they look very blocky.

If I switch to MS-DTV decoder I see 720x576 in and 2048x1080 out, and apart from the fact that I don't have the better deinterlacing of LAV Video the picture quality is much better, ie the scaling looks correct.

Can anybody think of a reason the above might happen?

DragonQ
21st January 2012, 12:20
1024x576 is the correct "display resolution" for 16:9 SD content (in countries that previously used PAL). No idea where 2048x1080 is coming from or why it would look better. What's your monitor/TV's native resolution and what's your desktop resolution set to?

wanezhiling
21st January 2012, 15:59
http://www.gokuai.com/f/5KE1J0M0f0F76F68
Frames drop all the time.. file problem?

NikosD
21st January 2012, 16:26
I've played 16-ref 1080p with DXVA (MPC-HC) on my integrated HD3300.

http://forum.doom9.org/showthread.php?t=163652
Try to set DXVA to "skip all checks" in MPC

Nope.

Unfortunately I did all the tests, there is a ReF limitation in UVD+ (HD 3000 series).

System: Win 7 SP1 x64 - Radeon 3650 - Catalyst 11.12

All three players I tried MPC-HC 1.5.3.3983 (latest), PotPlayer v1.5.31590 (latest) and WMP12 have artifacts playing even ReF 8.

Also:

MS DS, MS MFT, FFDshow DXVA have same problems.

Case closed. (For me)

UPDATE:

I found out a 1280 x 532 (High@L5.1) (CABAC/ 13 ReF Frames) which is OK.

Only one file and not 1080p file.

From the known 1080p L5.1 ReF>7 files, nothing is playable.

UPDATE2:

LAV 0.44 test3 is completely inappropriate for HD 3000 hardware (at least for 3650)

1) It can't accelerate WMV3_VLD at all (corrupted green images)

2) H.264 & VC-1 "acceleration" achieves about 2-3 fps regardless the clip

PotPlayer's DXVA is fine for all of the above.

In my opinion don't enable DXVA2 copy frame mode for HD 3000 series and below.
I know nothing about HD 4000, but i think they should be OK.

UPDATE3:

My card tested - 3650 - is an AGP card working at x8 mode and I enabled AGP Fast Writes.

If someone has a PCI-E HD 3000 series card, he could give one last chance for the hardware.

wanezhiling
21st January 2012, 19:06
LAV DXVA2(copy-back) works fine with mhd4650(UVD2.0) :)

nevcairiel
21st January 2012, 19:32
http://www.gokuai.com/f/5KE1J0M0f0F76F68
Frames drop all the time.. file problem?

The file seems to just be broken. I couldn't find anything that plays it properly


LAV 0.44 test3 is completely inappropriate for HD 3000 hardware (at least for 3650)


I doubt i'll ever spent much time to make it work on that old hardware, either it works, or it doesn't. The decoding bugs with WMV3 are unfortunate, and can probably be fixed somehow, however the copyback performance in such old cards is unlikely to ever change.
I already fixed some WMV3/VC1 bugs since the last test build, but who knows if that was the cause.

Mixer73
22nd January 2012, 00:12
1024x576 is the correct "display resolution" for 16:9 SD content (in countries that previously used PAL). No idea where 2048x1080 is coming from or why it would look better. What's your monitor/TV's native resolution and what's your desktop resolution set to?

Monitor is 1920x1080 and resolution is set to that in PotP.

It just appears that upscaling out of LAV Filters is somehow not working right, while other decoders with the same renderer look OK.

nevcairiel
22nd January 2012, 00:37
It just appears that upscaling out of LAV Filters is somehow not working right, while other decoders with the same renderer look OK.

The decoder should never upscale, or its doing it wrong.

zmaster
22nd January 2012, 04:26
When I am playing standard def files in PotPlayer with LAV Filters I see resolution in as 720x576 and resolution out as 1024x576...

If I switch to MS-DTV decoder I see 720x576 in and 2048x1080 out...It's look like a crazy OSD, which shows PotPlayer, if built-in video transform filter is disabled. But this does not explain the difference in quality. May be the screenshots will clarify something?

Mixer73
22nd January 2012, 04:29
The decoder should never upscale, or its doing it wrong.

I'm just wondering where I go to troubleshoot this, it never happened before I reformatted this system and it is a way cleaner install now than it was before.

MadVR shows from LAV Video that its getting 1024x576 in and 1920x1080 out and I've tried various scaling options (mostly using Lanczos 4 taps for chroma & luma), and it just looks blocky.

Is there a tickbox in LAV or a colour space or something I could inadvertently tick that can affect the scaling upstream?

Its doing my head in ;)

psymed
22nd January 2012, 05:00
What's file source async? Does that come with ffdshow and can it be removed or no?

nand chan
22nd January 2012, 05:33
Forgive me for not reading the whole thread, I haven't checked up on it for ages; is there any update on the status of OC + Linked files support? I'm still waiting for that one feature so I can finally say goodbye to Haali's Splitter.

mindbomb
22nd January 2012, 06:27
ever since advanced subtitle selection was added, i can't imagine ever using another splitter.

I'd actually say it is more important than linked file support.

nevcairiel
22nd January 2012, 09:59
Forgive me for not reading the whole thread, I haven't checked up on it for ages; is there any update on the status of OC + Linked files support? I'm still waiting for that one feature so I can finally say goodbye to Haali's Splitter.

Its still on the list, but its not high priority for me.

Thunderbolt8
22nd January 2012, 12:58
Its possible that your card doesn't support that specific file.
Do other DXVA decoders work on that?

A MediaInfo of that file may be helpful.

My code doesn't do any card-specific compat checks right now, it only checks for general compatibility (so only high profile, etc).

With recent drivers, do ATI/AMD still need checks on older cards regarding level and ref-frames? Do i need to implement this?got it working now with CCC 11.12

but it seems to be rather slow in case of a HD concert (AVC, 29,97fps). madVRs present queue is at 0/3 or 0/4 all the time, even with madvr deinterlacing off. using coreavc, the queue is at 3/4 all the time, with or without madvr deinterlacing.

RBG
22nd January 2012, 14:09
nevcairiel
Haven't you checked my sample yet? Is it possible to fix it?

iSunrise
22nd January 2012, 17:29
Monitor is 1920x1080 and resolution is set to that in PotP.

It just appears that upscaling out of LAV Filters is somehow not working right, while other decoders with the same renderer look OK.
Why would you set your monitor resolution in PotPlayer? madVR automatically does very high-quality scaling to your native display resolution when in fullscreen.

You don´t need to set PotPlayer to any fixed resolution, else PotPlayer always scales to that resolution and PotPlayer should not do that, that is the renderer´s job.

RBG
22nd January 2012, 19:20
iSunrise

I doubt that he actually turned on potplayer upscaler, most likely it is an osd bug.

http://img28.imageshack.us/img28/9863/20120122221620.th.jpg (http://img28.imageshack.us/i/20120122221620.jpg/)

http://img6.imageshack.us/img6/7194/20120122221720.th.jpg (http://img6.imageshack.us/i/20120122221720.jpg/)

http://img831.imageshack.us/img831/7169/20120122221825.th.jpg (http://img831.imageshack.us/i/20120122221825.jpg/)

fairchild
22nd January 2012, 20:13
iSunrise

I doubt that he actually turned on potplayer upscaler, most likely it is an osd bug.

It's definitely an OSD/Potplayer bug. If you use MadVR and you do CTRL+J to look at stats, you will see that MadVR is receiving the correct movie resolution and then scales to the correct target size (whatever your display device's resolution is).

BTW Nev, I just found out that you can get Lav Filters builds compiled by XhmikosR @ http://xhmikosr.1f0.de/index.php?folder=bGF2ZmlsdGVycw==

Are these basically builds you update with small fixes and whatnot when someone reports a bug on this forum? I figure they should be safe to use albeit maybe not as stable as your release builds. :)

nevcairiel
22nd January 2012, 21:02
Are these basically builds you update with small fixes and whatnot when someone reports a bug on this forum? I figure they should be safe to use albeit maybe not as stable as your release builds. :)

There have been a few issues with his builds in the past, i would recommend to just stick with the official builds, its not like it takes a year for a new one to appear.

Thunderbolt8
22nd January 2012, 23:17
but apparently moar than 2 weeks this time :devil:

Asmodian
22nd January 2012, 23:24
LOL at the :devil:

Two whole weeks! Forever ;)

Great work Nev, don't rush anything.

sexus
22nd January 2012, 23:52
hows that dvd support comin along nevcairiel?

Thunderbolt8
23rd January 2012, 00:05
LOL at the :devil:

Two whole weeks! Forever ;)

Great work Nev, don't rush anything.I like Rush, great songs. :sly:

SamuriHL
23rd January 2012, 00:12
hows that dvd support comin along nevcairiel?

Do you see it in the change list? Has a test build been posted with it? No? Hmmm. I wonder what that can mean. Asking constantly isn't going to make it get done faster. ;)

ryrynz
23rd January 2012, 00:26
There have been a few issues with his builds in the past, i would recommend to just stick with the official builds, its not like it takes a year for a new one to appear.

Yeah, I had an rendering issue a few days ago before c63c9b3 came out, that build of his has been running fine.
Sometimes you just see something like 'updated Quick Sync to latest revision and tweaked it to perfection" and you think, yeah I want this NOW.
Nev is pretty good, every three weeks or so there's a new version which is awesome, but even that can be hard to stomach when your following development.
I swear following software updates can be just a little addictive :D

Mixer73
23rd January 2012, 00:49
You don´t need to set PotPlayer to any fixed resolution, else PotPlayer always scales to that resolution and PotPlayer should not do that, that is the renderer´s job.

I only set it when I was having the issues with poor quality - its like a renderer setting, MPC-HC has the same.

It's definitely an OSD/Potplayer bug. If you use MadVR and you do CTRL+J to look at stats, you will see that MadVR is receiving the correct movie resolution and then scales to the correct target size (whatever your display device's resolution is).

OK so just an OSD bug for the res then, anybody got any bright ideas on something I might have touched which will affect scaling in MadVR and EVR? I guess its gotta be a Nvidia setting? Anything that would make scaling go blocky? Seeing as I tried GPU & custom shader scaling in MadVR this maybe the most likely.

It is strange though that its so much more pronounced using LAV Video than MS-DTV.

robpdotcom
23rd January 2012, 01:22
It is strange though that its so much more pronounced using LAV Video than MS-DTV.

Are you playing dvr-ms or wtv files? If so, you need to make sure you are using LAV Splitter - LAV Video doesn't like what the Microsoft splitter sends to it.

In mpc-hc, for wtv files, you just need to block "StramBufferSource" and make sure wtv is checked in LAV Splitter's configuration. For dvr-ms files, I also needed to block "WM AF Reader".

sexus
23rd January 2012, 04:52
Do you see it in the change list? Has a test build been posted with it? No? Hmmm. I wonder what that can mean. Asking constantly isn't going to make it get done faster. ;)

well atleast it was worth an try :D

SamuriHL
23rd January 2012, 04:56
well atleast it was worth an try :D

Lol. Just keep an eye on the change lists and test builds. If it's not announced it's not done.

Sent from my Xoom using Tapatalk

ryrynz
23rd January 2012, 06:51
@sexus You wanna be looking here :D

http://code.google.com/p/lavfilters/issues/detail?id=47&colspec=ID%20Type%20Status%20Priority%20Filter%20Summary

Mixer73
23rd January 2012, 07:14
Are you playing dvr-ms or wtv files? If so, you need to make sure you are using LAV Splitter - LAV Video doesn't like what the Microsoft splitter sends to it.

In mpc-hc, for wtv files, you just need to block "StramBufferSource" and make sure wtv is checked in LAV Splitter's configuration. For dvr-ms files, I also needed to block "WM AF Reader".

They are DVR-MS files.

Unfortunately LAV splitter produces out of sync playback for my PAL DVR-MS files. Also the MS splitter seems to handle the lossy broadcast files substantially better, as otherwise reported.

Also what is strange is that in PotP, no Splitter is appearing in the Filter list..

alegend
23rd January 2012, 11:50
I get curious picture problems when trying to use Intel QUicksync support with lsav 0.44

You can see it here:

http://www.sendspace.com/file/58tequ

Any idea whats the problem??

I have a brand new intel i5-2390t with QS Support

vivan
23rd January 2012, 12:13
alegend,
Do you have .2559 intel HD Graphics driver installed? If so - try downgrading to 15.22.50.64.2509 (8.15.10.2509)...

alegend
23rd January 2012, 12:21
yes newest intel driver is installed.

did u also have this problem?

sexus
23rd January 2012, 18:01
@sexus You wanna be looking here :D

http://code.google.com/p/lavfilters/issues/detail?id=47&colspec=ID%20Type%20Status%20Priority%20Filter%20Summary

really who wouldve thought :cool:

ratbert
23rd January 2012, 18:52
First after waiting five days to get speech, i wanted to thank nevcairiel for his great work. I never liked the haali media splitter because it kept crashing my mashines. The lav splitters are more than just a good replacement. I only wished i would have found this project sooner, that could have saved me some frustration and cash. I like that simple solution that solves all my problems maybe some day i will start again building a mediacenter pc. At moment the lav splitters can use cuda or even intel quicksync. I own a notebook with nvida optimus that has an intel and a nvida gpu so theoretical i could use both gpu´s at the same time. But to do that i would have to use another codec for one gpu. My question ist if it would be possible to have a second instance of the lav video decoder that can be configured separately? For me that could save a lot of time but maybe that is to much effort for one person. Im nearly sure there are not much wanting that feature.

CrAsH cro
23rd January 2012, 19:14
Is there any updates on bitstreaming DTS-WAV files? MPC-HC or MediaPortal?

hoborg
23rd January 2012, 19:36
For me, DTS in WAV working just fine, but i dont have reciever, so i cannot test bitstreaming :/
Anyone here can test it please?

mylle
23rd January 2012, 20:24
I can test if you provide me with a sample!

hoborg
23rd January 2012, 21:06
I can test if you provide me with a sample!

Here (http://hobring.esero.net/saf/samples/dtswav_test.zip).

mylle
23rd January 2012, 21:35
ok. i put in in the same folder where i have my other music on mediaportal. Tried to play it but it only outputs noise. Do i need to change settings anywhere?

Andy o
23rd January 2012, 21:40
DTS-WAV files are meant to be played bit-perfectly, so for Win 7/Vista you'll have to use WASAPI exclusive mode, and not use something like ReClock, or otherwise apply any processing to the audio. There are some decoders for it though, but they're not your normal DTS decoder.

mylle
23rd January 2012, 21:45
OK. I guess im in over my head here then. I bitstream HD audio on my movies and the codec section in MediaPortal is set to LAV audio and it works perfect. I have no clue how to set how the dts-wav file should be played.

nevcairiel
23rd January 2012, 22:42
LAV Audio can decode DTS-WAV, if you trick it into being used for raw PCM. With LAV Splitter, it should be used, with other splitters, probably not.

hoborg
23rd January 2012, 22:56
And yes, DTS in WAV working just fine, but i am not sure if DTS in WAV bitstream is working correctly.

Thunderbolt8
23rd January 2012, 23:52
got it working now with CCC 11.12

but it seems to be rather slow in case of a HD concert (AVC, 29,97fps). madVRs present queue is at 0/3 or 0/4 all the time, even with madvr deinterlacing off. using coreavc, the queue is at 3/4 all the time, with or without madvr deinterlacing.
with my laptop (radeon 4850 mobility) and a normal 24fps movie, present queue hovers between 3/4 and 0/4 all the time. its fine for a few secs, but then drops down to zero for a few seconds and so forth.

is there any chance of improving the speed? (coreavc is faster)
or are my gpus simply too slow?

ajp_anton
24th January 2012, 01:08
I have done a lot of tests with Windows Xp - Catalyst 11.2 - Radeon 3650, you can find here. (http://forum.doom9.org/showthread.php?t=159486) Look at the second system results.

That's why I asked ajp_anton about drivers version and OS.

Maybe after almost a year of my last try with 3650, I could try again under Win 7 system with latest Catalysts 11.12
Sorry for the late answer, been busy.

It was a very long time ago.
I found this post (http://forum.doom9.org/showthread.php?p=1397440#post1397440). I think I used Windows Server 2003 64-bit back then, no idea about drivers.
I have a newer OS now, but still the same hardware. Can't get DXVA working at all now though, and I don't really care since it's my server computer.

Mr.White
24th January 2012, 01:29
LAV Audio can decode DTS-WAV, if you trick it into being used for raw PCM. With LAV Splitter, it should be used, with other splitters, probably not.

MPC-HC, LAV Splitter, ReClock (wasapi) = Bitstreaming
But if LAV Audio is set to Prefer MPC-HC crash!

dts_wav (http://www.haringetliv.nu/pub/temp/dts_wav.jpg)

NikosD
24th January 2012, 08:39
Sorry for the late answer, been busy.

It was a very long time ago.
I found this post (http://forum.doom9.org/showthread.php?p=1397440#post1397440). I think I used Windows Server 2003 64-bit back then, no idea about drivers.
I have a newer OS now, but still the same hardware. Can't get DXVA working at all now though, and I don't really care since it's my server computer.

Well, things are getting a little worse for HD 3300.
It seems that it doesn't even have a UVD+. It has UVD, which was the first generation inside HD 2000 series.

From the OS version used (Windows Server 2003) combined with UVD, I think there is no chance at all to play HW accelerated L5.1-Ref 16-1080p clips in that system.

The only way to actually see such clips played in your system, is with a software decoder.
In my opinion, your media player/decoder made a fall back to software mode and you didn't notice.

If your newer version of Windows Server is 2008 SP1, you could try the modified Win 7 drivers, which install the whole Catalyst suite on Server 2008.

Anyway, I did my little research for UVD+ with newer system and drivers here (http://forum.doom9.org/showpost.php?p=1553048&postcount=8409)

PeQuE
24th January 2012, 09:14
OK. I guess im in over my head here then. I bitstream HD audio on my movies and the codec section in MediaPortal is set to LAV audio and it works perfect. I have no clue how to set how the dts-wav file should be played.
I asume you're trying to play dts-wav file using MyMusic section of MediaPortal... If that's the case, then you're probably using internal BASS Player engine. I can play dts-wav files and they bitstream perfectly. Please check that:

- You've got volume at 100%
- You are not using crossfading.
- You don't need to, but I recomend you to install Asio4all (http://www.asio4all.com/), and configure MediaPortal to use it for MyMusic.

Sorry for the offtopic, nev...

alegend
24th January 2012, 11:04
alegend,
Do you have .2559 intel HD Graphics driver installed? If so - try downgrading to 15.22.50.64.2509 (8.15.10.2509)...

HEY!!! This one really worked!!!
I downgraded to that .2509 driver and it works like a charm!
Big thx!

Pat357
24th January 2012, 12:01
with my laptop (radeon 4850 mobility) and a normal 24fps movie, present queue hovers between 3/4 and 0/4 all the time. its fine for a few secs, but then drops down to zero for a few seconds and so forth.

is there any chance of improving the speed? (coreavc is faster)
or are my gpus simply too slow?
In which mode do you set Coreavc when you compare it to Lav DXVA-copy-back ?
Is coreavc DXVA mode ? Then it would not work which MadVR, and probably switch to SW mode.

ajp_anton
24th January 2012, 12:21
Well, things are getting a little worse for HD 3300.
It seems that it doesn't even have a UVD+. It has UVD, which was the first generation inside HD 2000 series.

From the OS version used (Windows Server 2003) combined with UVD, I think there is no chance at all to play HW accelerated L5.1-Ref 16-1080p clips in that system.

The only way to actually see such clips played in your system, is with a software decoder.
In my opinion, your media player/decoder made a fall back to software mode and you didn't notice.

If your newer version of Windows Server is 2008 SP1, you could try the modified Win 7 drivers, which install the whole Catalyst suite on Server 2008.

Anyway, I did my little research for UVD+ with newer system and drivers here (http://forum.doom9.org/showpost.php?p=1553048&postcount=8409)I'm 100% sure DXVA was used. The MPC-HC DXVA decoder showed heavy artifacting with the default amount of video RAM (only using the onboard 128MB) while software always worked, but with 256MB video RAM it worked.
Why would you think it only has UVD? The UVD series is pretty much what defined the integrated series since they didn't really improve the performance. HD2000 was UVD, HD3000 was UVD+ and HD4000 was UVD2.

nevcairiel
24th January 2012, 12:23
UVD+ is indeed present in the HD3000 series, however AMD never made a big distinction between UVD and UVD+, they marketed it as pure UVD. Only with UVD2, they started naming it differently.

NikosD
24th January 2012, 12:52
Why would you think it only has UVD? The UVD series is pretty much what defined the integrated series since they didn't really improve the performance. HD2000 was UVD, HD3000 was UVD+ and HD4000 was UVD2.

UVD+ is indeed present in the HD3000 series, however AMD never made a big distinction between UVD and UVD+, they marketed it as pure UVD. Only with UVD2, they started naming it differently.

As I wrote to my first post, HD 3000 series should all have UVD+.

But Radeon HD 3300 and HD 3200 are IGPs paired with AMD 790GX Chipset and AMD 780G Chipset.

Well, Wikipedia says net UVD and not UVD+ for both IGPs.

So it seems it's not the same like the other HD 3000 series cards, probably because HD 3200 and HD 3300 aren't "real" cards.

According to Tom's Hardware, 780G and 790GX chipsets have both UVD 1.0 and only 785G has UVD 2.0.
But of course because UVD+ was also being marketed as simply UVD, that means that could have and should have UVD+. Or UVD 1.0.

As I have written before, even UVD 2.2 inside HD 4000 series is different than UVD 2.2 inside HD 5000 series, which is different than UVD 2.2 inside HD 6700 series.

In my opinion the REAL DIFFERENCE between UVD, UVD+, UVD2, UVD2.2(HD 4000), UVD2.2(HD 5000), UVD2.2(HD 6700), UVD 3(HD 6000), UVD 3? (HD 7000) is the way that ATI/AMD supports the hardware by exposing certain features through drivers/BIOS of a hardware which more or less the same.

If you know what I mean ;)

And that last point ajp_anton means, that maybe you could actually play 1080p L5.1 Ref 16 files, with suitable drivers/OS/BIOS (Graphics card), even in hardware (UVD/UVD+) that it's not supposed to play such files "officially".

Check this (http://www.agile-news.com/news-323468-Photo:-no-fear-of-4K-x-2K!-Fire-whirlwind-2-HD6570-2G-Daniel-Edition-hardware-solution-2160p!.html) out and try to explain it.

How is it possible for UVD3 to accelerate 4K x 2K video files with 238Mbps bitrate and 14% CPU usage and the same card cannot play 1080p60 ???

ajp_anton
24th January 2012, 13:16
Oh well, I don't care what it has, when the fact is that 1080p 16ref worked on my 790GX =).

NikosD
24th January 2012, 13:24
Sure, but you should care two times why it's not working now, if it worked sometime ago.

It looks like MPEG-2 VLD support.

It was a time that it worked in HD 4000 series with a registry hack, but then stopped and never reappeared.

UVD features are like ghosts, afterall :scared:

CruNcher
24th January 2012, 16:37
In my opinion the REAL DIFFERENCE between UVD, UVD+, UVD2, UVD2.2(HD 4000), UVD2.2(HD 5000), UVD2.2(HD 6700), UVD 3(HD 6000), UVD 3? (HD 7000) is the way that ATI/AMD supports the hardware by exposing certain features through drivers/BIOS of a hardware which more or less the same.

It is most likely partly a programmable architecture same as Nvidia and Intel have and microcodes get updated via the Drivers (firmware), in terms of GPU Hardware Asics people shouldn't think anymore in the old terms of full fixed things (it would be much to unflexible) ;)
Though obviously somewhere are limits but even those limits can be workedarround see Nvidias Mpeg-2 fix via the Shader ;)

NikosD
24th January 2012, 17:07
Agree.

ATI put a fast ASIC inside from the beginning and started to play with features and names with the help of drivers and firmware/BIOS.

I know about ATI HD 3000 that supports MPEG2_IDCT in GPU shaders, not in UVD+.

What do you mean about Nvidia fix for MPEG2?

nevcairiel
24th January 2012, 17:30
NVIDIA fixed MPEG-2 with CUDA, saying "with shaders" is incorrect.
Their hardware MPEG-2 decoder had a bug in earlier generations (a hardware bug), so with some driver they just implemented a new MPEG-2 decoder in CUDA which didn't have the bug, but still runs plenty fast.

CruNcher
24th January 2012, 17:30
There is a decoding problem for certain bitstreams in everything pre VP4 and Nvidia fixes it by using the Shaders for Decoding instead the VPx it's a design bug most probably that couldn't get fixed :). Donald was the first implementing this workarround via the Shaders later Nvidia implemented it in the Drivers directly.

@Nev
just for testing could you implement Arcsofts and Cyberlinks VC-1 GUIDs in the .wmv parser part, they both work fine with .ts but with .wmv Arcsoft don't get the DXVA connection on the used GUID though Cyberlinks decoder doesn't even connect ;)

PS: It seems you need some special bit somewhere GUID alone doesn't seem enough :(

clsid
24th January 2012, 17:40
Here (http://hobring.esero.net/saf/samples/dtswav_test.zip).A crash happens in avcodec-lav-53.dll when DCA is disabled in LAV Audio. Interestingly enough, LAV Audio is part of the graph. Crash happens when pressing play in GraphStudio. Also, it does not connect to ffdshow. Could you make it output a normal mediatype as well?

nevcairiel
24th January 2012, 17:45
A crash happens in avcodec-lav-53.dll when DCA is disabled in LAV Audio. Interestingly enough, LAV Audio is part of the graph. Crash happens when pressing play in GraphStudio.

I changed it so that DTS-in-WAV detection is now disabled when the DTS codec is disabled, this should avoid the crash, but will of course only output garbage.

The problem with DTS-in-WAV is that the splitter does not know that its DTS, it'll claim its normal 16-bit PCM. Only LAV Audio will try to figure it out...

ajp_anton
24th January 2012, 17:51
Sure, but you should care two times why it's not working now, if it worked sometime ago.Like I said, it's a server computer, never used for video playback. I don't have any media players or decoders installed, and video memory is back to only using the onboard 128MB. I can't even remember why I did test it back then.

CruNcher
24th January 2012, 18:21
Nah this seems not easy todo seems only PowerDVD 11 (default) and Potplayer (using either Arcsoft or Cyberlinks Decoder and their VC-1 Adapter filter with Microsofts parser) are capable of this ( .wmv VC-1 AP playback via DXVA2 no copy back on Intel Quicksync) currently even with Av Splitters connection tricks it endsup in a black screen on MPC-HC with most streams else its switching to YUY2 input and consumes much more CPU then both solutions above, so for MPC-HC only Quicksync works (reliable ffdshow-quicksync or lav video quicksync) in this .wmv VC-1 AP case though with a lot higher CPU consumption :P

hoborg
24th January 2012, 18:40
I changed it so that DTS-in-WAV detection is now disabled when the DTS codec is disabled, this should avoid the crash, but will of course only output garbage.

The problem with DTS-in-WAV is that the splitter does not know that its DTS, it'll claim its normal 16-bit PCM. Only LAV Audio will try to figure it out...
But did you tried to bitstream it?

nevcairiel
24th January 2012, 19:37
Here is a "release candidate" of the next version.
Its basically feature complete for what i wanted to do, just hoping no severe bugs are still in there.

x86: http://files.1f0.de/lavf/LAVFilters-0.45-rc.zip
x64: http://files.1f0.de/lavf/LAVFilters-0.45-rc-x64.zip

New since 0.44 is the DXVA2 decoder.
Its known to work pretty well, however its not really all that super fast on ATI hardware. On older models, like HD3000 and older, it'll most likely be unusable.
DXVA2 is currently only available in "copy back" mode, which means it needs to copy the video frame back from the GPU to the CPU, which is the reason for the low performance. "Native" DXVA2 is planned for a later version.

I'm mostly interested in bugs with the DXVA2 decoder, or regressions since 0.44.

RBG
24th January 2012, 21:05
nevcairiel
Mpeg-4 ASP acceleration is back... Man, you've made my day. :thanks:

nevcairiel
24th January 2012, 21:06
I would however not recommend using it, it bugs out on some files, seems to be related to bitstream packing or something in that area. Its off by default for that reason.

NikosD
24th January 2012, 21:27
Perfect decoding for every clip I tried H.264/VC-1/WMV3 in DXVA2 copy frame mode.

Same "slow" speed for Radeon (6)750 as with Lav 0.44 DXVA2 copy frame.

We have to wait for direct DXVA2 implementation.

I have also Deinterlacing for DXVA VC-1 even with VMR renderer in DXVA checker.
Is VMR renderer capable of Deinterlacing with decent quality ?

Because I think that HW Deinterlacing is not supported in LAV for ATI (I didn't see an option).

Maybe the drivers are involved and do Vector adaptive deinterlacing regardless of decoder/renderer.

One "bug" only.
You have kept version 0.44 for all your filters (splitter, audio, video)

I don't have VP4 right now to check, but did you implement MPEG4 ASP for VP4/VP5 and UVD3 in DXVA2 copy frame or just NVCUDVID for Nvidia HW only?

nevcairiel
24th January 2012, 21:30
I have also Deinterlacing for DXVA VC-1 even with VMR renderer in DXVA checker.
Is VMR renderer capable of Deinterlacing with decent quality ?

Because I think that HW Deinterlacing is not supported in LAV for ATI (I didn't see an option).

If you use VMR or EVR, it'll do hardware deinterlacing through the driver.


One "bug" only.
You have kept version 0.44 for all your filters (splitter, audio, video)

Its not 0.45 yet. :p


I don't have VP4 right now to check, but did you implement MPEG4 ASP for VP4/VP5 and UVD3 in DXVA2 copy frame or just NVCUDVID for Nvidia HW only?
Only in CUVID mode. I might do DXVA later, if i get bored enough.

Gser
24th January 2012, 21:34
DXVA2 is currently only available in "copy back" mode, which means it needs to copy the video frame back from the GPU to the CPU, which is the reason for the low performance. Does this mean I could use LAVF to deinterlace my videos threw DXVA2 in avisynth? MadVR for instance deinterlaces threw DXVA2, it just doesn't return the frames.

nevcairiel
24th January 2012, 21:37
Does this mean I could use LAVF to deinterlace my videos threw DXVA2 in avisynth? MadVR for instance deinterlaces threw DXVA2, it just doesn't return the frames.

Its only decoding with DXVA2, not deinterlacing.
Deinterlacing uses another surface type as render target, downloading from that is incredibly slow (~5fps) for ATI, so its not worth it.

clsid
24th January 2012, 21:38
Perhaps the DTS-in-WAV detection could be moved from decoder to splitter? Those files can successfully be parsed by for example AC3File, allowing any DTS decoder to be used.

nevcairiel
24th January 2012, 21:42
Perhaps the DTS-in-WAV detection could be moved from decoder to splitter?

You're welcome to write a patch.
There is a reason its like this. If it would be trivial to change, it would've been done already.

Thunderbolt8
24th January 2012, 21:54
"Native" DXVA2 is planned for a later version.looking forward ;)

Mangix
25th January 2012, 02:08
LAV Audio's FLAC decoder fails when i use reclock's accept bitstream formats. reclock falsely believes that lav audio is sending a bitstreamed format... mpc-hc's built-in FLAC decoder works fine.

CruNcher
25th January 2012, 03:25
@Nev
http://git.1f0.de/gitweb?p=ffmpeg.git;a=commit;h=fe15af222b118ac350c7e0b2178b85ff364f4b7e <-finally, good that Reimar pushed it through :) 1 line that can change so much ;)

hmm this should be in 0.45 RC1 shouldn't it, if it should it seems to fail for dshow :( ?

hmm that lav splitter also shows 0.44 in its version number ?

File Version = 0.44.0.0

boyumeow
25th January 2012, 04:56
...

hmm that lav splitter also shows 0.44 in its version number ?

File Version = 0.44.0.0

I believe this should be nev answer.

...
Its not 0.45 yet. :p
...

Aleksoid1978
25th January 2012, 05:23
Here is a "release candidate" of the next version.
Its basically feature complete for what i wanted to do, just hoping no severe bugs are still in there.

x86: http://files.1f0.de/lavf/LAVFilters-0.45-rc.zip
x64: http://files.1f0.de/lavf/LAVFilters-0.45-rc-x64.zip

New since 0.44 is the DXVA2 decoder.
Its known to work pretty well, however its not really all that super fast on ATI hardware. On older models, like HD3000 and older, it'll most likely be unusable.
DXVA2 is currently only available in "copy back" mode, which means it needs to copy the video frame back from the GPU to the CPU, which is the reason for the low performance. "Native" DXVA2 is planned for a later version.

I'm mostly interested in bugs with the DXVA2 decoder, or regressions since 0.44.

Very bad work in DXVA mode with H.264 interlaced and MBAFF on Intel HD. If need - i can upload a problem sample. Progressive H.264 - work good.

Aleksoid1978
25th January 2012, 05:55
Also - i have sample what play in DXVA with artefacts in MPC-HC and in LAV with DXVA mode. Play fine with CUDA and other DXVA decoder like Microsoft or Cyberlink.

dead_screem
25th January 2012, 06:59
ok, minor bug, DXVA2 shows as "Available" on XP.

Also in CUVID certain VC-1 interlaced files drop frames alot, used to work fine. And with CUVID adaptive deinterlacing in a certain vc-1 sample you might have the sample already "Slicies VC-1 NTSC 1080i 29.97 Bilder.ts". in the bottom half of the screen several of the horizontal white lines disappear completely.
(also CUVID+YADIF=very slow framerate, like 7fps on my 9800GTX, but this may be expected?)

Aleksoid1978
25th January 2012, 07:21
Another bug - Indeo3 http://aleksoid.tosei.ru/Test/Sample/indeo30.avi playback is choppy with LAV, ffdshow - playback is ok.

nevcairiel
25th January 2012, 07:48
LAV Audio's FLAC decoder fails when i use reclock's accept bitstream formats. reclock falsely believes that lav audio is sending a bitstreamed format... mpc-hc's built-in FLAC decoder works fine.

Thats a ReClock bug, nothing that can be done.
Enable MPC-HCs Audio Switcher, it'll prevent this from happening.


And with CUVID adaptive deinterlacing in a certain vc-1 sample you might have the sample already "Slicies VC-1 NTSC 1080i 29.97 Bilder.ts". in the bottom half of the screen several of the horizontal white lines disappear completely.

The highest quality deinterlacer is not available on Windows XP (which i guess from your previous comment you're using).
Enabling it caused severe image degredation for alot of people, so i opted to just completely disable it for an overall better experience.


hmm this should be in 0.45 RC1 shouldn't it, if it should it seems to fail for dshow :( ?

Its not in the build i posted, will be in release.

nevcairiel
25th January 2012, 08:30
Another bug - Indeo3 http://aleksoid.tosei.ru/Test/Sample/indeo30.avi playback is choppy with LAV, ffdshow - playback is ok.

Fixed, was a bug in libav.

nixties
25th January 2012, 08:49
With DXVA2 MPEG2 still artefacts - Sample (http://www.mediafire.com/download.php?79f42499cc12964), Full file (http://www.mediafire.com/download.php?e2v8eva8780a2lk)

NikosD
25th January 2012, 09:05
@nev

Is it possible to implement a Multi-threaded CPU decoding of VC1-/WMV3 ?

Because if for some reason the DXVA is not working or doesn't exist at all for VC-1/WMV3, the next option is a single threaded CPU decoding which is very slow, especially for VC-1 which is an advanced video format and the most difficult to decode after H.264 of course.

I have tried a lot of decoding implementations - none of them is Multi-threaded.

Is there a particular reason or difficulty ?

madshi
25th January 2012, 09:33
The MS software VC-1 decoder is multithreaded, I think?

nevcairiel
25th January 2012, 09:45
Is it possible to implement a Multi-threaded CPU decoding of VC1-/WMV3 ?

Possible, sure, but for a complex format like VC-1, its a bit out of my skill set. I'm not nearly advanced enough in video decoding itself to be able to do that. Maybe some day someone will do it in ffmpeg/libav.

The MS software VC-1 decoder is multithreaded, I think?
I'm not sure, but at least it always was fast enough to decode 1080p even on my old Core2Duo

madshi
25th January 2012, 10:04
The only thing I have problems with on my development PC is interlaced VC-1 (when using software decoding). But I think it's not the decoding (MS decoder) itself which is too slow, I think the MS decoder tries to deinterlace internally and I've found no way to turn that off.

NikosD
25th January 2012, 10:56
The MS software VC-1 decoder is multithreaded, I think?

The so called "multi-threaded" MS VC-1 decoder, although it seems that is filling both 2 cores at my Core2Duo, it actually has the decoding performance of the single-threaded FFmpeg/ LAV/ PotPlayer and every other single-threaded decoder out there.

So no, we can forget MS "multi-threaded" VC-1 decoder.


I'm not sure, but at least it always was fast enough to decode 1080p even on my old Core2Duo

No, unfortunately it's not.
Try to play 1080/60p clip like Devil_may_cry_gameplay.wmv (ftp://helpedia.com/pub/multimedia/x264/testvideos/2010%20-%2009%20-%20DXVA%20benchmarks%20-%20Avivo%20vs%20PureVideo%20vs%20Clear%20Video/) on every Core2Duo you want.
You will never see 60fps, unless you use DXVA.

Possible, sure, but for a complex format like VC-1, its a bit out of my skill set. I'm not nearly advanced enough in video decoding itself to be able to do that. Maybe some day someone will do it in ffmpeg/libav.

I see.

italospain
25th January 2012, 11:09
The only thing I have problems with on my development PC is interlaced VC-1 (when using software decoding). But I think it's not the decoding (MS decoder) itself which is too slow, I think the MS decoder tries to deinterlace internally and I've found no way to turn that off.

You can try wmv PowerToy.
You can disable there the internal deinterlacer.

Mosley
25th January 2012, 13:44
I recently reported HERE (http://sourceforge.net/apps/trac/mpc-hc/ticket/1970)a bug involving (in part) LAV splitter. Have you ever experienced a similar problem with blu-rays using MPC-HC or another player?

madshi
25th January 2012, 13:47
The so called "multi-threaded" MS VC-1 decoder, although it seems that is filling both 2 cores at my Core2Duo, it actually has the decoding performance of the single-threaded FFmpeg/ LAV/ PotPlayer and every other single-threaded decoder out there.
The MS VC-1 decoder is noticeably faster on my PC than the ffmpeg VC-1 decoder.

You can try wmv PowerToy.
You can disable there the internal deinterlacer.
Cool, thanks!

nevcairiel
25th January 2012, 13:49
I recently reported HERE (http://sourceforge.net/apps/trac/mpc-hc/ticket/1970)a bug involving (in part) LAV splitter. Have you ever experienced a similar problem with blu-rays using MPC-HC or another player?

Bit-streaming of E-AC3 (DDPlus) from Blu-rays is not supported at this time, and possibly never will be.
That one demo disc you mention in the ticket is the *only* known disc to actually use that type of audio, no real world movie ever used it to my knowledge.

NikosD
25th January 2012, 14:48
I would however not recommend using it, it bugs out on some files, seems to be related to bitstream packing or something in that area. Its off by default for that reason.

You mean you have to deal with new bugs that appeared during the integration of your LAV CUVID, or the bugs you mention were always there from the beginning even in LAV CUVID ?

Maybe a hardware's bug like MPEG2 ?

nevcairiel
25th January 2012, 14:50
You mean you have to deal with new bugs that appeared during the integration of your LAV CUVID, or the bugs you mention were always there from the beginning even in LAV CUVID ?

Maybe a hardware's bug like MPEG2 ?

Its a bug in the CUVID implementation, most likely in the parser, which is part of the NVIDIA driver, not the hardware.

mindbomb
25th January 2012, 16:25
The MS software VC-1 decoder is multithreaded, I think?

multithreaded for progressive, single threaded for interlaced iirc.
ofc, better than not being able to decode interlaced at all.

there was talk before about integrating the microsoft decoder into lav video, was that scrapped?
with all the hardware acceleration available now, it is not terribly important, but it would be cool.
and then we could have vc-1 turned on in lav video by default.

Mangix
25th January 2012, 18:49
Thats a ReClock bug, nothing that can be done.
Enable MPC-HCs Audio Switcher, it'll prevent this from happening.

like i said, mpc-hc's built-in audio decoder works perfectly. i forgot to mention that this only happens to 2 channel FLAC files. 6ch FLAC files work fine(encoded as ac3 and sent to my receiver by ReClock).

also your suggestion does not work. lav audio does not even make a connection. here are two screenshots.

http://imageshack.us/photo/my-images/220/clipboard01mo.jpg/
http://imageshack.us/photo/my-images/856/clipboard02pb.jpg/

mindbomb
25th January 2012, 19:09
wait, lav splitter is a flac source filter?

since when?

nevcairiel
25th January 2012, 19:44
With DXVA2 MPEG2 still artefacts - Sample (http://www.mediafire.com/download.php?79f42499cc12964), Full file (http://www.mediafire.com/download.php?e2v8eva8780a2lk)

That file is odd.
Go decode it with software. :p

nevcairiel
25th January 2012, 20:32
like i said, mpc-hc's built-in audio decoder works perfectly. i forgot to mention that this only happens to 2 channel FLAC files. 6ch FLAC files work fine(encoded as ac3 and sent to my receiver by ReClock).

also your suggestion does not work. lav audio does not even make a connection. here are two screenshots.

http://imageshack.us/photo/my-images/220/clipboard01mo.jpg/
http://imageshack.us/photo/my-images/856/clipboard02pb.jpg/

First of all, its not a audio decoder bug. MPC-HCs internal decoder works because MPC-HC forces it into the graph, but the decoder doesn't control if its used or not. :)
I fiddled around with the flac mediatype as exposed by the splitter, and it seems to be fixed. Its still a ReClock bug though, it shouldn't accept FLAC if it cannot bitstream it.

wait, lav splitter is a flac source filter?
Sure, its a source filter for a gazillion of formats, most i just don't expose in the installer.

CruNcher
25th January 2012, 21:33
ok confirmed the H.264 concatenate fix its working as expected no freezes for all samples anymore in mplayer as well as lav splitter, adios MPC-HC splitter :)

Thunderbolt8
25th January 2012, 21:45
The MS software VC-1 decoder is multithreaded, I think?
from what Ive once heard, it only looks like being multithreaded, because task manager is juggling with the load or something like that. not sure if thats still the latest news, though.
I guess if it was really multithreaded, I wouldnt need to complain about speed issues with (interlaced) VC-1 stuff. (normal=progressive VC-1 HD videos also tend to create more speed problems than in case of multithreaded AVC decoding for example (in cases of comparable video bitrate of course))

nixties
25th January 2012, 21:46
That file is odd.
Go decode it with software. :p
In this case will be very useful to add option for preventing H/W decoding of SD-video ;)
P.S. That file plays fine via CUVID, Cyberlink DXVA/HAM, Microsoft DTV/DVD; PotPlayers DXVA VLD plays with light garbling.


And another related bug - on many SAT-captured video garbling appears at playback beginning/when seeking (for example this file (http://www.mediafire.com/download.php?kamofff5u9fc916)), fast seeking of that files cause VGA-driver hanging (NV GTS450).

nevcairiel
25th January 2012, 21:56
If you have an NVIDIA, i would always prefer CUVID, its just better. :p
Anyhow, i fixed it a tiny bit that it doesn't crap out completely at the beginning anymore, but its still not really good. :(
Maybe i can still find something, we'll see.

kolak
25th January 2012, 22:57
Bit-streaming of E-AC3 (DDPlus) from Blu-rays is not supported at this time, and possibly never will be.
That one demo disc you mention in the ticket is the *only* known disc to actually use that type of audio, no real world movie ever used it to my knowledge.

Yes- no point to bother with DD Plus.

SamuriHL
25th January 2012, 23:20
I think there's some confusion here about DD Plus. LAV Audio supports it just fine when it comes from sources other than Blu-ray. HD DVD's, for example, bitstream just fine and decode just fine. I haven't yet seen a broadcast in the US with DD Plus but I know other countries do. And those probably work ok. (The US will be using it soon....I keep telling myself that. :D)

roytam1
26th January 2012, 01:33
new nightly:
http://roy.orz.hm/lavf-w32-nightlies/lavf-my120126-rv34seekkeyframe-f2b8cb3.7z

diff:
http://roy.orz.hm/lavf-w32-nightlies/lavf-my120126-rv34seekkeyframe-f2b8cb3.diff

Andy o
26th January 2012, 02:30
I think there's some confusion here about DD Plus. LAV Audio supports it just fine when it comes from sources other than Blu-ray. HD DVD's, for example, bitstream just fine and decode just fine. I haven't yet seen a broadcast in the US with DD Plus but I know other countries do. And those probably work ok. (The US will be using it soon....I keep telling myself that. :D)

Do you know if there any news of it working with AMD 6000 or 7000? Googling only takes me to a post of my own asking the same question at AVS. :confused:

SamuriHL
26th January 2012, 02:39
Do you know if there any news of it working with AMD 6000 or 7000? Googling only takes me to a post of my own asking the same question at AVS. :confused:

It works on my 5870 so I can't see why it wouldn't work on the 6xxx or 7xxx cards.

mrcorbo
26th January 2012, 05:29
Do you know if there any news of it working with AMD 6000 or 7000? Googling only takes me to a post of my own asking the same question at AVS. :confused:

Just tested a DD+ sample I had muxed in MKV and it bitstreamed fine on my 6970.

Andy o
26th January 2012, 06:16
It works on my 5870 so I can't see why it wouldn't work on the 6xxx or 7xxx cards.

That's news to me, maybe AMD fixed it in drivers?

EDIT: yeah it does work here too, IIRC works with LAV but it didn't use to work with ffdshow.

nevcairiel
26th January 2012, 07:44
It indeed always worked in LAV, it was just a ffdshow bug that it didn't work before.

Mosley
26th January 2012, 08:21
That one demo disc you mention in the ticket is the *only* known disc to actually use that type of audio, no real world movie ever used it to my knowledge.
I didn't know. I thought DDPlus Audio were fairly widespread.

nevcairiel
26th January 2012, 08:22
I didn't know. I thought DDPlus Audio were fairly widespread.

Sure, its used on HDDVD and in some broadcasts, BUT, the Blu-ray version of DDPlus is "special", its not comparable to normal DDPlus, and its not used on any actual Blu-rays. :)

Blu-rays use something rather odd, there is a normal DD/AC3 5.1 stream, and then there is a 4 channel DDPlus stream. If the player is capable of playing that DDPlus stream, its used to replace the surround channels from the 5.1 stream, as well as adding two new channels, giving 7.1 ... its a odd setup, just for the sake of supporting players that don't understand DDPlus.

Andy o
26th January 2012, 08:33
It indeed always worked in LAV, it was just a ffdshow bug that it didn't work before.

Didn't work with the commercial players either, so we all thought it was the driver or hardware at the time. So kudos!

SamuriHL
26th January 2012, 14:33
That's news to me, maybe AMD fixed it in drivers?

EDIT: yeah it does work here too, IIRC works with LAV but it didn't use to work with ffdshow.

Ffdshow has a bug that prevents it from working. The hardware has always supported it, though. When bitstreaming was added to lav audio it was one of the first things I tried and it worked perfectly.

Sent from my Xoom using Tapatalk

SamuriHL
26th January 2012, 14:34
Didn't work with the commercial players either, so we all thought it was the driver or hardware at the time. So kudos!

None of them ever actually added support for bitstreaming ddp. :)

Sent from my Xoom using Tapatalk

Andy o
26th January 2012, 16:50
I remember it working with NV cards... I think WinDVD did at least.

SamuriHL
26th January 2012, 16:55
I don't think I ever tried in WinDVD cause the only content I have with DDP currently is HD DVD sources.

Pat357
26th January 2012, 20:11
like i said, mpc-hc's built-in audio decoder works perfectly. i forgot to mention that this only happens to 2 channel FLAC files. 6ch FLAC files work fine(encoded as ac3 and sent to my receiver by ReClock).

also your suggestion does not work. lav audio does not even make a connection. here are two screenshots.

What happens if you use MadFlac -> Reclock ?

RBG
26th January 2012, 20:32
nevcairiel

I found a bug in 0.45-rc, video playback is sluggish and choppy with AV splitter on this sample (http://www.mediafire.com/?i5a343iotmczic3), on previous version of LAV video(0.44) everything is fine. This problem occurs only when LAV video is in the CUVID mode. BTW I packed AV splitter .ax with video sample into archive file, in case you don't have one.

MediaInfo.
Video
IGeneral
Unique ID : 216221930323604009574219194167342582500 (0xA2AAD43AF07CD585807AE6544B22CAE4)
Complete name : D:\test\New folder (7)\The_thing .mkv
Format : Matroska
Format version : Version 2
File size : 81.8 MiB
Duration : 19s 545ms
Overall bit rate mode : Variable
Overall bit rate : 35.1 Mbps
Encoded date : UTC 2012-01-26 18:51:24
Writing application : mkvmerge v5.2.1 ('A Far Off Place') built on Jan 2 2012 23:21:10
Writing library : libebml v1.2.3 + libmatroska v1.3.0

Video
ID : 1
Format : VC-1
Format profile : AP@L3
Codec ID : WVC1
Codec ID/Hint : Microsoft
Duration : 19s 561ms
Bit rate : 31.0 Mbps
Width : 1 920 pixels
Height : 1 080 pixels
Display aspect ratio : 16:9
Frame rate : 23.976 fps
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Progressive
Compression mode : Lossy
Bits/(Pixel*Frame) : 0.624
Stream size : 72.3 MiB (88%)
Title : 30357 kbps
Language : English
Default : No
Forced : No

Mangix
26th January 2012, 20:52
What happens if you use MadFlac -> Reclock ?

no idea. however that nightly build posted by roytam1 fixed the bug. even though it is a reclock bug, the workaround is still appreciated :)

nevcairiel
26th January 2012, 20:56
I found a bug in 0.45-rc, video playback is sluggish and choppy with AV splitter on this sample (http://www.mediafire.com/?i5a343iotmczic3), on previous version of LAV video(0.44) everything is fine. This problem occurs only when LAV video is in the CUVID mode.

Seems fine here, try the nightly build posted by roytam on the previous page.

NikosD
26th January 2012, 21:23
I tried LAV CUVID MPEG4 on a GT440 VP4 card and I can say for sure that there is no ASP HW acceleration.

I tried every video file I have in ASP format (Advanced Simple Profile@L5), I mean all of my 720p and 1080p video files and I got a black screen.

They couldn't play at all.

The only clips they got HW acceleration were Simple@L3 clips and Mpeg-4 Visual (DivX 5) clips.

The decoding was bug-less and fast enough.
Hardest clip 1080/30p gave me 70fps in benchmark mode.

CruNcher
26th January 2012, 21:31
@Nev
congrats on beating Intel Quicksync on the Mpeg-2 CPU Consumption with the current DXVA2 copy back not sure yet what are the differences visually between Quicksyncs VLD and Libavcodecs DXVA2 but it in consumption (decoding efficiency) it wins so far :)

lav video quicksync = 9%/37%
lav video dxva2 = 7%/41%

that looks like a good balance CPU/GPU :)

nevcairiel
26th January 2012, 21:34
I tried LAV CUVID MPEG4 on a GT440 VP4 card and I can say for sure that there is no ASP HW acceleration.

Must be a driver issue, alot of the clips i tested with LAV CUVID back in the day don't work anymore.
But i cba to install older drivers, maybe someone else feels compelled to do it. :)

Thats why the option is off by default. :p
I also added a warning message in the hover tip

Afterall, its MPEG4, it rarely comes in HD, and even if it does, its decoded very easily in software.

NikosD
26th January 2012, 21:58
A lot of problems with MPEG4 ASP and Nvidia hardware/ drivers.

It makes sense now that nobody supports MPEG4 ASP on Nvidia HW in DXVA mode and you are the only one who tried to support in NVCUVID, with limited success though.

I'm wondering what is the situation regarding UVD3 HW and MPEG4 ASP.

Does anyone have Advanced Simple Profile@L5 clips and UVD3 to test ?

RBG
26th January 2012, 22:01
Seems fine here, try the nightly build posted by roytam on the previous page.


Indeed works fine on nightly build. Thank you.:)

nevcairiel
26th January 2012, 22:03
A lot of problems with MPEG4 ASP and Nvidia hardware/ drivers.

It used to work much better, must've been maybe around 270 driver or so.
Its probably just a issue in the cuvid library, maybe a DXVA solution would work better.

We'll see, maybe a new driver fixes it again, if it doesn't, i may remove it again some day.

CruNcher
26th January 2012, 22:32
If that isnt efficient i dunno :)

Playback (1080i 60 fps)

lav video quicksync = 9%/37%
lav video dxva2 = 7%/41%

Benchmark

lav video = 63% = 438 fps
lav video dxva2 = 26% = 350 fps
lav video quicksync = 40% = 312 fps
ffdshow-quicksync = 56% = 335 fps

nevcairiel
26th January 2012, 22:33
Too bad DXVA2 on Intel still has some corruption issues with H264. :p

Aleksoid1978
27th January 2012, 08:16
I have freeze my PC (help only Restart button :)) when try to playback this sample http://aleksoid.tosei.ru/Test/Sample/1080i%20VC-1%5b1%5d.split.2.m2ts at DXVA mode. Play fine at CUVID mode.

NikosD
27th January 2012, 09:27
Too bad DXVA2 on Intel still has some corruption issues with H264. :p

On SNB HW I don't have any issues with DXVA H.264 and PotPlayer.
It even plays fine 16 Ref and any H.264 clip I've thrown on it.

Do you have any odd sample to test ?

nevcairiel
27th January 2012, 09:28
I assume you're not talking about PotPlayers DXVA decoder, but mine? Because thats all i care about. :)
The issues seem to happen mostly with MBAFF coded files.

I just have to spend a few hours trying to figure out where the problem is, but at least on SNB there is QuickSync to fix that..

NikosD
27th January 2012, 09:38
I thought you were talking about general problems on Intel's HW.

I try to keep equal distances from all the developers that I use their products.

Different view from you obviously ;)

roytam1
27th January 2012, 13:35
new nightly:
http://roy.orz.hm/lavf-w32-nightlies/lavf-my120127-9adb451.7z

No diff provided as the diff was checked-in, thanks nev.

vivan
27th January 2012, 16:20
On SNB HW I don't have any issues with DXVA H.264 and PotPlayer.
It even plays fine 16 Ref and any H.264 clip I've thrown on it.

Do you have any odd sample to test ?http://www.mediafire.com/?id7cd8aujnmwslf

CruNcher
28th January 2012, 04:02
@ Nev
another .ts issue

http://www.mediafire.com/?1pm5y68sbf35fst

lav splitter = fails
mpc-hc splitter = works
av splitter = works
vlc = works
mplayer = works
libav ffplay = works
ffmpeg ffplay = works

nevcairiel
28th January 2012, 08:53
@ Nev
another .ts issue

http://www.mediafire.com/?1pm5y68sbf35fst


Fixed

nevcairiel
28th January 2012, 12:33
I have freeze my PC (help only Restart button :)) when try to playback this sample http://aleksoid.tosei.ru/Test/Sample/1080i%20VC-1%5b1%5d.split.2.m2ts at DXVA mode. Play fine at CUVID mode.

Fixed

clsid
28th January 2012, 13:58
Nev, could you have another look at MsVideo1/MSRLE and other paletted formats? With LAV Splitter the palette data is handled automatically, but for the MS AVI splitter you need to do that manually. Relevant code can be found in libavformat\avidec.c line 624 and 1133.

nevcairiel
28th January 2012, 15:00
The funny thing is, i don't even know why it works with LAV. I don't supply the palette, but for some reason it figures it out.

I did however add some code that should try to send the palette to the decoder.

Edit:
I also figured out why/how it works with LAV, hooray. :)

vinnytx
28th January 2012, 18:37
Hello,

Please someone can post the link of lav filters dxva?
Thanks

wanezhiling
28th January 2012, 18:37
Hello,

Please someone can post the link of lav filters dxva?
Thanks


http://xhmikosr.1f0.de/index.php?folder=bGF2ZmlsdGVycw==

nevcairiel
28th January 2012, 19:27
LAV Filters 0.45

LAV Splitter
- Improved playback of concatenated H.264 MOV/MP4 files
- Adjusted FLAC MediaType to avoid ReClock stealing the connection
- Improved playback of MPEG-TS with timestamp discontinuities

LAV Audio
- Fixed playback of very short PCM samples
- Improved COOK decoding without LAV Splitter

LAV Video
- New DXVA2 "copy back" decoder (see release notes)
- Updated Intel QuickSync decoder (r29)
- Multi-threaded decoding for RV30/40, UtVideo and DNxHD
- Support for the "FCC" color matrix
- Improved Dirac support
- CUVID support for MPEG-4 ASP (experimental)
- Support for Intel Indeo 4, MS Video 1, 8BPS, LOCO, ZMBV, VCR1, AASC


Download: Installer (both x86/x64) (http://files.1f0.de/lavf/LAVFilters-0.45.exe) -- Zips: 32-bit (http://files.1f0.de/lavf/LAVFilters-0.45.zip) & 64-bit (http://files.1f0.de/lavf/LAVFilters-0.45-x64.zip)

New and note-worthy? Well....

DXVA2 "copy back"
In our series of hardware decoders, the new DXVA2 decoder now features hardware decoding also for AMD/ATI.
Its a "copy back" (or "renderless") decoder - that means all frames are copied back from the GPU to normal memory.

Sadly, however, AMD/ATI cards suffer from a serious performance penalty, which prevents them from showing their full potential. The decoder is fully Blu-ray compliant, and even goes beyond that, however playing 60fps movies will most likely not work fluidly.
H.264, VC-1 and MPEG-2 are fully supported (progressive and interlaced), as long as your hardware can keep up. Very old AMD/ATI cards, like 3000 series and below, might not be able to use this decoder. 4000 and up seem to be working ok.

I do have plans to work on a "native" DXVA2 decoder, which does not need to copy the frames back from the GPU, avoiding this performance bottleneck - however that goes with the loss of the flexibility of course. No post-processor, only EVR, etc.

For everyone else, i recommend to stick to your native decoders (CUVID/QuickSync), as they generally work better and faster then DXVA2.

Known Issues:
- Interlaced H.264 on Intel can show image corruption
- Some MPEG-2 streams show image corruption (all GPUs)

There is also the potential for crashing the GPU driver. It happened mostly on VC-1 interlaced, and all cases i encountered i also fixed, but if it happens, just let me know, and i'll look into it.

MPEG-4 ASP in the CUVID decoder
There seems to be an issue in the latest NVIDIA drivers that cause the decoder to not work on most MPEG-4 ASP files!
Older drivers, like 270 series, seem to be working OK.

Its experimental, use it at your own risk, its off by default.

And the rest...
Everything else is pretty much standard changes, a few new video formats supported, a number of improvements inherited from ffmpeg, the usual number of bugs fixed.

I still didn't get around to re-writing the installer, but i hope i can get the motivation to do it for the next big release.
What else is up next? I haven't decided. There is some features on the table, audio mixing, dvd support, some subtitle work .. we'll see.

Take care.

wanezhiling
28th January 2012, 19:31
:pFirst

nevcairiel
28th January 2012, 19:33
:pFirst

Actually, you are #8554

fastplayer
28th January 2012, 19:46
4000 and up seem to be working ok.
Confirmed, at least on my HD4670. This is the first time I've used DXVA without having anything to complain about. Dunno if it's because the drivers have matured by now or nev just knows how it's done. I suspect the latter. :D
:thanks:

dukey
28th January 2012, 19:47
excellent work nevcairiel,
Copy back is good, but native DXVA would be even better :) What do you plan to do with subtitles ?

Sebastiii
28th January 2012, 20:00
Great :) Thanks :)

vinnytx
28th January 2012, 20:49
http://xhmikosr.1f0.de/index.php?folder=bGF2ZmlsdGVycw==

Thank you very much

clsid
28th January 2012, 20:57
The funny thing is, i don't even know why it works with LAV. I don't supply the palette, but for some reason it figures it out.

I did however add some code that should try to send the palette to the decoder.

Edit:
I also figured out why/how it works with LAV, hooray. :)
I can confirm it works great now. Just curious, why did it work with LAV?

clsid
28th January 2012, 21:03
What else is up next? I haven't decided. There is some features on the table, audio mixing, dvd support, some subtitle work .. we'll see.Issue 32 would be cool as well. Loading DirectVobSub after LAV Video. Should be needed only for WMP/WMC on Vista/7. Having that feature would mean one less reason to use Haali splitter and ffdshow when wanting subs.

nevcairiel
28th January 2012, 21:09
Just curious, why did it work with LAV?

Apparently the demuxers insert the palette data in a special format at the end of the frame data, which the ffmpeg decoders can then read, and every other normal decoder would just ignore it, because it already has all the data it needs. It seems to be working just fine. :)

ffdshow should also be able to understand that data, unless you broke that part in your libav somehow. :p

Issue 32 would be cool as well.
Since other people also have commented that they want some feature done more then the others

I work on what is fun to work on, not what would be useful! :)
Also, opinions on what feature is important are quite different amongst different people.

Anyhow, i don't take requests, is all. :)

aufkrawall
28th January 2012, 22:53
Hello nevcairiel,
I think it'd be a good idea to force RGB output for FRAPS video, if possible. This is required, the official FRAPS decoder does it too. Without it the video doesn't look right.
FRAPS RGB source LAV default output:
http://www.ld-host.de/uploads/thumbnails/48f239fd34598dc0782f3f2dd78ea060.png (http://www.ld-host.de/show/48f239fd34598dc0782f3f2dd78ea060.png)

LAV forced RGB output:
http://www.ld-host.de/uploads/thumbnails/de55107ab40558b7cd4183903313d0ab.png (http://www.ld-host.de/show/de55107ab40558b7cd4183903313d0ab.png)


FRAPS 4:2:0 source LAV default output:
http://www.ld-host.de/uploads/thumbnails/b33e31341e86df8b95671028b204ef84.png (http://www.ld-host.de/show/b33e31341e86df8b95671028b204ef84.png)

FRAPS 4:2:0 source LAV forced RGB output (it's really how it should look like, it's not too bright ;) ):
http://www.ld-host.de/uploads/thumbnails/7e70bdc94309f2d3da429f0af6c1d2bd.png (http://www.ld-host.de/show/7e70bdc94309f2d3da429f0af6c1d2bd.png)

FRAPS 4:2:0 source FRAPS decoder output:
http://www.ld-host.de/uploads/thumbnails/0e9943732a44bd813e1a50cd369a0396.png (http://www.ld-host.de/show/0e9943732a44bd813e1a50cd369a0396.png)

I'm going to suggest this also for ffdshow.

CruNcher
28th January 2012, 22:59
@Nev

stall issue (like known from the beyonce stream)

http://www.mediafire.com/?k3d0sum8cepydcj

jmone
28th January 2012, 23:38
FYI - Testing with J.River MC (cross post from the Beta Forum)
=========== PC: Tested on GTS450 with DXVA2 Option and madVR =========
- VC-1 and x264 at high framerate (eg 50/60i, 50p) works just fine and is fluid with no dropped frames
- MPEG SD (720x576i) works just fine and is fluid with no dropped frames
- MPEG HD / Non Square Pixel ?? (1440x1080) just presents a static Green Video output
- DTV plays fine (not Green Video on HD Content as well)

Works pretty well on this rig

=========== HTPC =========
Here are the more intersting / exciting results. The above results was on my PC, the following is from my HTPC where I have been using the Intel IGP X3000 (i7-2600K chip). It has been working well with SW decoding (no surprise given the 2600K's processing power) over Dec but I did had to change madVR's Luma Scaling algorithms to something like Mitchell-Netravali over anything with "Taps" as it would drop frames as render times would blow out.

So here is how the IGP X3000 goes with the two HW Accel Options:

Testing on X3000 with DXVA2 Option and madVR:
- VC-1 and x264 at high framerate (eg 50/60i, 50p) works just fine and is fluid with no dropped frames
- MPEG SD (720x576i) works just fine and is fluid with no dropped frames
- MPEG HD / Non Square Pixel (1440x1080) works just fine and is fluid with no dropped frames
- DTV crashed MC

Testing on X3000 with Intel Quicksync Option and madVR:
- VC-1 and x264 at high framerate (eg 50/60i, 50p) works just fine and is fluid with no dropped frames
- MPEG SD (720x576i) works just fine and is fluid with no dropped frames
- MPEG HD / Non Square Pixel (1440x1080) works just fine and is fluid with no dropped frames
- DTV works just fine and is fluid with no dropped frames

This is only some quick testing and:
1) I've not really tried testing the madVR scaling algorithms with the various HW Decoders but it seems to give more overhead to leave as it is. I did see dropped frames in 1440x1080/50i MPEG 2 with Lanczos (3-Taps) and DXVA2
2) I only played short 5-sec clips on my test files. I was looking at madVR's OSD to see any reported issues with dropped frames and long queue times and also half an eye on the playback to see an very obvious issues.
3) I used madVR to do the deinterlacing on both Quicksync and DXVA2. I saw some odd deinterlacing judder using YADIF in some combinations.

Couple of Q:
1) If there is a speed issue with decoding in LAVVideo, apart from seeing visual issues - will madVR report issues in the queues / dropped frames
2) Let me know if there is any specific combos that need testing.

Thanks for the great work and the IGP option for HTPC looks very promising --> one less item, lower heat, power and noise!

CruNcher
29th January 2012, 00:23
Your playback is complex you have to keep an eye on your GPU load as well CPU load alone isnt gonna work these days anymore, so you could transfer some load Deinterlacing to the CPU and keep more free for the GPU that way in Interlaced Scenarios and upscaling, though SB has nice scaling (Adaptive Lanczos4) that should top MadVR in Efficiency (being fixed function) ;)

The most efficient way on SB is trying to get every possible load on the fixed function and only absolute necessary load on the EUs GPU/CPU, though this is currently only doable with EVR or a custom Renderer that can make use of these functions (Cyberlink,Arcsoft)


@Nev
Auto AV sync correction (Lav Audio) fails in this special case (emergency weather warning system)

http://www.mediafire.com/?kzpt3pictktu6bh

ryrynz
29th January 2012, 01:12
Caught this during development builds and it hasn't been fixed in 0.45, there appears to be two issues.

If IntelQuickSyncDecoder.dll is not located or cannot be loaded then LAV falls back to standard decoding but doesn't display any message.
It would be nice if there was some indication about the state of the currently running video decoder which brings me to the second issue.

IntelQuickSyncDecoder 0.24 looks to be either broken or LAV isn't loading it properly and falling back to software mode.
If I install 0.45 set QuickSync and run GraphStudios decoder benchmark I get the same results as with software decoding ~1550FPS.
If I copy IntelQuickSyncDecoder 0.22 over and then benchmark I get ~2418 FPS.

Aspeh
29th January 2012, 01:44
Me and two people i house share with one who is studying Art and Design and the other Photography and Media just don't understand why people insist on running MadVR with LAV filters. ?

Tested on various screens and computers and even one laptop LAV filters looks really nice and natural but MadVR to us looks over processed to the point of looking fake like it's been airbrushed with photoshop with bad colour saturation and is to dark to watch.

Unless we are doing something wrong we just don't get it. ?

DragonQ
29th January 2012, 02:03
LAV Filters contains a splitter and various decoders. It reads the video file, finds the video and audio streams you want, decodes them and passes them to the renderer, which then displays the video and plays the audio. MadVR is generally considered the best quality software renderer. It has advanced algorithms for resizing/upscaling that can make low resolution videos look noticeably sharper than the standard alternative (EVR) can.

For example, the best way to watch SDTV is probably to use LAV Filters (with CUVID decoding and vector adaptive deinterlacing) and MadVR (with high quality settings). This should be able to best even some hardware deinterlacers and upscalers (e.g. what you'd find in a good quality HDTV).

(I think I have all that right. :p)

Fadeout
29th January 2012, 02:05
MadVR on its own doesn't touch colors or anything, it only slightly changes scaling, and it's barely noticeable.

CruNcher
29th January 2012, 02:13
@DragonQ
EVRs quality depends on the Hardware Scaling implementation underneath it you cant generalize it in any way to "EVR is bad" thats a pure myth that MadVR Avg joe users come up with, because the Hardware implementations in their card sucks. Also EVR Custom Renderer support shader based upscaling with different algorithms as well just recently MPC-HC experimental got Lanczos support. The biggest advantage of MadVR is its Playback stability :)

ajp_anton
29th January 2012, 02:30
Does LAV Audio work with stereo speakers (headphones) on its own, or is that what the future "mixer" is for? I'd like to get rid of ffdshow for good, which is still there only for the audio.

iSunrise
29th January 2012, 02:50
Does LAV Audio work with stereo speakers (headphones) on its own, or is that what the future "mixer" is for? I'd like to get rid of ffdshow for good, which is still there only for the audio.
Yeah, I would be interested in that, too. It would be great if nev could do a mixing matrix along with some "standard" presets, so it would make ffdshow´s mixing solution pretty much obsolete for me, too.

Andy o
29th January 2012, 03:34
@DragonQ
EVRs quality depends on the Hardware Scaling implementation underneath it you cant generalize it in any way to "EVR is bad" thats a pure myth that MadVR Avg joe users come up with, because the Hardware implementations in their card sucks. Also EVR Custom Renderer support shader based upscaling with different algorithms as well just recently MPC-HC experimental got Lanczos support. The biggest advantage of MadVR is its Playback stability :)

madVR's advantage with the level expansion and RGB conversion are very visible in test gradients (no banding), but on movies I can't tell the difference, though I haven't bothered to look for it very much since madVR just works for me. It's possible that on CG animation you could find a difference.

aufkrawall
29th January 2012, 04:29
madVR's advantage with the level expansion and RGB conversion are very visible in test gradients (no banding)

What about EVR custom + LAV with forced RGB output?

DrNein
29th January 2012, 06:08
Does LAV Audio work with stereo speakers (headphones) on its own, or is that what the future "mixer" is for? I'd like to get rid of ffdshow for good, which is still there only for the audio.

ffdshow headphone mode is "meh" and I would not expect any non-commercial filters to be any better. I use CyberLink Audio for Dolby Headphone but still must use ffdshow as an intermediate for HD audio. Would like to ditch ffdshow as well.

CruNcher
29th January 2012, 06:26
What about EVR custom + LAV with forced RGB output?

Indeed especially on SB it shows its more efficient to use the RGB32 conversion for chroma upsampling than most shaders (which higher GPU usage heavily, of course if you don't mind 15% more GPU usage and want to save the last drop of CPU in my tests it showed to be roughly 1% saving on SD HD 4% MPEG-2, so 3%/25% vs 2%/40% not a good deal imho)

madVR's advantage with the level expansion and RGB conversion are very visible in test gradients (no banding), but on movies I can't tell the difference, though I haven't bothered to look for it very much since madVR just works for me. It's possible that on CG animation you could find a difference.

Not only Chroma upsampling improvements can be seen also on a lot of TV Sets that use colorful lightning (I don't mean the TV @ your home but the Set in the TV Studio so Broadcast Content any Kind of Shows for example that have complex Stages and Lightning or Pop Concerts with a lot of effects)

mindbomb
29th January 2012, 06:35
i noticed this:
playing the ifo files in dvd video results in the total calculated duration of the video being incredibly off.

Using MPC VTS Reader with Lav Splitter.


What about EVR custom + LAV with forced RGB output?
Good, but outputting rgb32 will result in no deinterlacing from the renderer iirc.

So, you should only use it for cuvid atm, or with yadif.

hissatsu
29th January 2012, 06:40
I posted this several weeks ago, but apparently this isn't a common problem, as I didn't get any responses.

I'm having an issue with seamless blu-ray playback, using every version of LAV Filters after 0.39 (though 0.45). When it switches from one m2ts in the mpls playlist to the next (like in a TV series where each episode is a separate m2ts on the BD), the screen goes black, but audio continues playing normally. Hitting Ctrl-J to get statistics shows nothing, but right clicking brings up the menu. To get the video back, I have to do something that interacts with the splitter. If I seek or skip chapters, the video comes back. Switching audio tracks also brings back the video. Skipping chapters forward or back to the point where the next m2ts starts, the screen does not go black; it only happens during normal playback at the point of the switch.

I'm using MPC-HC with MadVR on Windows 7 x64 with a GTX 550Ti, and I've experienced this on two separate Windows installs with almost completely different hardware (Intel/AMD, different video cards). I've tried every version of LAV Filters through 0.45, with hardware acceleration (CUVID, DXVA2) and without; I've tried every version of MadVR for the past few months and many builds of MPC-HC, along with other renderers (EVR-CP, EVR Sync, Haali), and no setting makes any difference other than going back to LAV Filters 0.39 or earlier. I'd really like to use a newer version, but something appears to have changed with the splitter after that point.

If anyone has any ideas/suggestions to fix this or if anyone is having similar issues, I'd really like to know.

Andy o
29th January 2012, 06:47
I know with ffdshow's HQ RGB conversion + dithering I could achieve similar, but not the same, results with madVR's tests. I haven't tested those with LAV's RGB conversion. In any case, unless you have a reason not to use madVR (like underpowered hardware), why not? It's also the easiest way to do high-quality deinterlacing and IVTC. Pretty much the only way to get it done with AMD cards. With NV cards and CUVID, you still have to select 25/30p output manually (and change to 50/60p for regular deinterlacing) and apply an avisynth script for decimation, for it to work.

CruNcher
29th January 2012, 07:00
i noticed this:
playing the ifo files in dvd video results in the total calculated duration of the video being incredibly off.

Using MPC VTS Reader with Lav Splitter.



Good, but outputting rgb32 will result in no deinterlacing from the renderer iirc.

So, you should only use it for cuvid atm, or with yadif.

Not correct depends on the Hardware (Driver), on SB its not the case (EVR Custom)

nevcairiel
29th January 2012, 07:17
If IntelQuickSyncDecoder.dll is not located or cannot be loaded then LAV falls back to standard decoding but doesn't display any message.
It would be nice if there was some indication about the state of the currently running video decoder which brings me to the second issue.

The property page has a text field next to the dropdown, which will indicate if the DLL could be loaded properly. Feedback during playback would just be terribly annoying.
It'll say "Available" if you just open it and it can load the DLL (and some basic checks succeed), and it'll say "Active" if its actually being used during playback.


IntelQuickSyncDecoder 0.24 looks to be either broken or LAV isn't loading it properly and falling back to software mode.
If I install 0.45 set QuickSync and run GraphStudios decoder benchmark I get the same results as with software decoding ~1550FPS.
If I copy IntelQuickSyncDecoder 0.22 over and then benchmark I get ~2418 FPS.

Works just fine here.
Its known to be broken for non-SandyBridge GPUs, but thats not a bug in my code, but in the DLL, so not a bug report for this thread. :)

nevcairiel
29th January 2012, 07:33
@Nev
Auto AV sync correction (Lav Audio) fails in this special case (emergency weather warning system)

http://www.mediafire.com/?kzpt3pictktu6bh

That file has issues no matter what. When i seek after that announcement, its completely desync, which means that weather thingy completely screws up the timestamps of the audio.

I'll put that into the broken file category. :)

nevcairiel
29th January 2012, 07:59
@Nev

stall issue (like known from the beyonce stream)

http://www.mediafire.com/?k3d0sum8cepydcj

Fixed

ryrynz
29th January 2012, 08:12
The property page has a text field next to the dropdown, which will indicate if the DLL could be loaded properly. Feedback during playback would just be terribly annoying.
It'll say "Available" if you just open it and it can load the DLL (and some basic checks succeed), and it'll say "Active" if its actually being used during playback.


Thanks, I was thinking of a status bar icon or something but that 'active' I guess does the trick. It shows as active so it is loading 0.24 okay.
GraphStudio still benches it slow however, does yours bench the same with 0.22 vs 0.24? I'd like to confirm before posting the bug in the ffdshow forum.
For the record I have a Sandy 2600K, and the latest ffdshow runs QuickSync with the proper performance showing in GraphStudio like in 0.22 with LAV.
It would be odd if we're running the same CPU and your 0.24 benches the same or higher than using 0.22.

nevcairiel
29th January 2012, 08:23
To get your kind of numbers, it would have to be SD or something, wouldn't it?

I benchmarked a pretty default 1080p H264 MKV with LAV 0.45:

0.24:
~200 fps

0.22 (from LAV 0.44):
~170 fps

And a SD file

0.24
~800 fps

0.22:
~750 fps

Something to note:
When decoding very short files, so that the decoding process only takes a fraction of a second for the complete file, the numbers will most likely be very inaccurate.
In fact, the only way i ever see numbers like 1500 or 2400 is when the file is extremely short.

PS:
My version of the QS decoder is slightly newer then the one that ships in ffdshow at the moment, but not sure it would explain anything. :)
Technically, its all 0.23, the latest commit just increased the version to 0.24 without implementing any of the 0.24 features yet. :p

Aleksoid1978
29th January 2012, 09:06
Have issue with LAV Audio http://aleksoid.tosei.ru/Test/Sample/Matrix%201.avi
issue only with internal MPC-HC Avi splitter and system Avi splitter. Play fine with LAV Splitter.

nevcairiel
29th January 2012, 09:43
Have issue with LAV Audio http://aleksoid.tosei.ru/Test/Sample/Matrix%201.avi
issue only with internal MPC-HC Avi splitter and system Avi splitter. Play fine with LAV Splitter.

Fixed

ryrynz
29th January 2012, 09:45
@Nev It's was anime intro 704x480 at 1 min 12 seconds.

I took to a couple of ~1920x1080 Youtube trailers and benched.
I set to the run 5 times and average the result.

Snow White and the Huntsman 1.53 minutes
Normal LAV 650 FPS
QuickSync 0.22 476 FPS
QuickSync 0.24 361 FPS

Wrath of the Titans 2.05 minutes
Normal LAV 684 FPS
QuickSync 0.22 737 FPS
QuickSync 0.24 609 FPS

Same results, 0.22 beats 0.24.

The Snow White trailer leaves me a little confused I thought QuickSync was ALWAYS going to be faster..

Qaq
29th January 2012, 09:46
What about EVR custom + LAV with forced RGB output?
This way LAV adds dither noise after YUV>RGB conversion and EVR upscales that noise. madVR adds dither after upscaling at very final stage.

CruNcher
29th January 2012, 09:51
@Nev

Stream crashes the DXVA copy back Mpeg-2 decoder (currently the only one)

http://www.mediafire.com/?y2dbekgemoeo28m

nevcairiel
29th January 2012, 10:00
Same results, 0.22 beats 0.24.


It seems to depend mostly on the stream in question, i actually found one where 0.22 is faster as well now, but the ones i tested earlier still have 0.23 being faster.

I'll report my findings to Eric.

ryrynz
29th January 2012, 10:02
Awesome, thanks for confirming, I can upload all files I've tested with if Eric would like them.

Sebastiii
29th January 2012, 10:12
I posted this several weeks ago, but apparently this isn't a common problem, as I didn't get any responses.

I'm having an issue with seamless blu-ray playback, using every version of LAV Filters after 0.39 (though 0.45). When it switches from one m2ts in the mpls playlist to the next (like in a TV series where each episode is a separate m2ts on the BD), the screen goes black, but audio continues playing normally. Hitting Ctrl-J to get statistics shows nothing, but right clicking brings up the menu. To get the video back, I have to do something that interacts with the splitter. If I seek or skip chapters, the video comes back. Switching audio tracks also brings back the video. Skipping chapters forward or back to the point where the next m2ts starts, the screen does not go black; it only happens during normal playback at the point of the switch.

I'm using MPC-HC with MadVR on Windows 7 x64 with a GTX 550Ti, and I've experienced this on two separate Windows installs with almost completely different hardware (Intel/AMD, different video cards). I've tried every version of LAV Filters through 0.45, with hardware acceleration (CUVID, DXVA2) and without; I've tried every version of MadVR for the past few months and many builds of MPC-HC, along with other renderers (EVR-CP, EVR Sync, Haali), and no setting makes any difference other than going back to LAV Filters 0.39 or earlier. I'd really like to use a newer version, but something appears to have changed with the splitter after that point.

If anyone has any ideas/suggestions to fix this or if anyone is having similar issues, I'd really like to know.

Hi,

I have tested on Terminator Salvation (on MediaPortal with LAV Splitter ofc).
I didn't have your issue, i have keep running 2 switching on m2ts (quite simple on that BD, one is @ 53 seconds and the second @ 1 minute 56 secs.)
Running on NVidia with CUDA.
I will try to do the same test with my Avatar seamless BD and on MPC-HC.
Thanks

*EDIT* : On Avatar, it's ok too (on MePo)

nevcairiel
29th January 2012, 10:15
Awesome, thanks for confirming, I can upload all files I've tested with if Eric needs them.

I actually found the reason why ffdshow seems faster, and "fixed" that, still doesn't explain why 0.22 is faster then 0.24, i'll investigate a bit before i post.

ryrynz
29th January 2012, 10:31
Nice, you're on a bit of a fixing spree! I wasn't expecting a performance drop with 0.24 after reading about the improvements Eric made over 0.22 and certainly with it depending on the stream as much as it does, I just assumed it was broken. :D

Aspeh
29th January 2012, 10:36
MadVR on its own doesn't touch colors or anything, it only slightly changes scaling, and it's barely noticeable.

If you run LAV output side by side with MadVR it certainly looks alot different but that might be because we can't seem to work out between the 3 of us how to change Colour,Contrast and Brightness of MadVR to closer match that of the LAV output.

nevcairiel
29th January 2012, 10:41
@Nev

Stream crashes the DXVA copy back Mpeg-2 decoder (currently the only one)

http://www.mediafire.com/?y2dbekgemoeo28m

That one is fine with NVIDIA, only fails on Intel for me. I would suspect a driver bug or something like that...
I'll try to figure it out, maybe. :)

DragonQ
29th January 2012, 12:13
@DragonQ
EVRs quality depends on the Hardware Scaling implementation underneath it you cant generalize it in any way to "EVR is bad" thats a pure myth that MadVR Avg joe users come up with, because the Hardware implementations in their card sucks. Also EVR Custom Renderer support shader based upscaling with different algorithms as well just recently MPC-HC experimental got Lanczos support. The biggest advantage of MadVR is its Playback stability :)
Hmmm. EVR gives me pretty bad banding on a GT 430, even when supplied with RGB32. MadVR doesn't give me this effect, but I can't use it for live TV in MediaPortal unfortunately.

I also find it strange that EVR works instantly in MPC-HC whereas MadVR has some weird issues like the aspect ratio being wrong for a split second when switching to/from full screen, and a half second delay between raising/lowering volume and the OSD coming up.

Midzuki
29th January 2012, 12:55
Fixed

Would you mind telling us what was "fixed" exactly? :)

I don't intend to download a +600MB file, just to (TRY TO) find that out by myself. ;)

nevcairiel
29th January 2012, 13:08
Would you mind telling us what was "fixed" exactly? :)

The file used to not play properly, now it does - hence, fixed. :p
It was screwing up timestamps when the demuxer didn't bother to split the data into frames, and just delivered them in one big blob

You can always read the commit log if you want to know more. ;)

Midzuki
29th January 2012, 13:56
The file used to not play properly, now it does - hence, fixed. :p
It was screwing up timestamps when the demuxer didn't bother to split the data into frames, and just delivered them in one big blob

:thanks:

You can always read the commit log if you want to know more. ;)

But of course :o :stupid:

thebigmunch
29th January 2012, 13:56
If you run LAV output side by side with MadVR it certainly looks alot different but that might be because we can't seem to work out between the 3 of us how to change Colour,Contrast and Brightness of MadVR to closer match that of the LAV output.Like someone somewhat explained before, LAV Filters and MadVR are two different things. One is a splitter and decoder package, while the other is a video renderer. You can't really compare them side-by-side. What video renderer are you using where you're saying LAV Filters?

Reino
29th January 2012, 14:21
Hello nevcairiel,
I think it'd be a good idea to force RGB output for FRAPS video, if possible. This is required, the official FRAPS decoder does it too. Without it the video doesn't look right...

...I'm going to suggest this also for ffdshow.You're not the first one: FFDShow feature request - Output colorspace as-is (http://sourceforge.net/tracker/?func=detail&aid=3458249&group_id=173941&atid=867363).

nevcairiel
29th January 2012, 14:40
LAV will always try to output the color space of the video untouched, however, that is not always easy.

For some formats, like Fraps, its impossible to know what the "right" format is before we decoded the first frame.
However, before decoding it, we have to tell the renderer what we're sending, which if unknown is the default NV12 (because it covers like 90% of all important video).

Now, when the first frame was decoded, we know that its actually RGB, and we try to tell the renderer that we actually would like to send RGB.
Here is where the problems start. EVR does not allow changing its input format after the initial connection has been made, so its just not possible to switch from NV12 to RGB on-the-fly.
This problem does *not* exist with any other renderer. VMR is fine, Overlay Renderer is fine, Haali is fine, madVR is fine, just EVR (any variant) fails.

Thats why Fraps has this issue, and thats another reason why forcing RGB output for everything when using EVR is a good idea.
If people tell me that RGB Fraps is far more common then YUV420 Fraps, then i might add a special rule for Fraps to default to RGB instead of NV12, but this has to be done on a per-codec basis.

rica
29th January 2012, 15:19
@nevcairiel,

LavSplitter gives false information; it gives 48k/16 info while the original audio is 96k/24.

Here is a sample:

http://www.mediafire.com/?2jygu42sokq5c52

(Tom Petty and the Heartbreakers, Blu-Ray Audio.)

STaRGaZeR
29th January 2012, 15:19
If people tell me that RGB Fraps is far more common then YUV420 Fraps, then i might add a special rule for Fraps to default to RGB instead of NV12, but this has to be done on a per-codec basis.

It's not IMO, but defaulting to RGB is a good idea with this codec regardless. Levels issues and other stuff.

BTW did you decide what to do when playing yuv444 content with AYUV output disabled, instead of converting to 420?

Reino
29th January 2012, 15:22
Afaik is Fraps always RGB. At least I've never heard of YV12 Fraps files.

Excuse me if this sounds ignorant (I have no experience with coding splitters or decoders), but wouldn't it be much easier then for the splitter to detect the file's colourspace, so the decoder doesn't have to decode the first frame?

Of course this still doesn't explain why FFDShow's own FFV1(RGB) format and HFYU(RGB) are always converted to YV12, but this is also the wrong topic to discuss that.

aufkrawall
29th January 2012, 15:30
The official FRAPS Decoder always converts to RGB, both for YV12/RGB.
Otherwise everything is messed up due to PC range.

If possible there should be hardcoded that when the source's codec is FPS1, always RGB should be outputted.

nevcairiel
29th January 2012, 15:33
Afaik is Fraps always RGB. At least I've never heard of YV12 Fraps files

Fraps has 6 versions

0: YUV 4:2:0
1: RGB
2: YUV 4:2:0
3: RGB
4: YUV 4:2:0
5: RGB

There are differences how the data is encoded, but thats the color spaces used.
The bad thing is, FPS1 is the FourCC for all variants.


Excuse me if this sounds ignorant (I have no experience with coding splitters or decoders), but wouldn't it be much easier then for the splitter to detect the file's colourspace, so the decoder doesn't have to decode the first frame?
Sure, the splitter knows the correct format (at least LAV Splitter does), but there is no "standard" way to communicate that to the decoder. I've been thinking about adding a LAV specific way to communicate this, so the problem can at least be controlled when using LAV Splitter + LAV Video.


BTW did you decide what to do when playing yuv444 content with AYUV output disabled, instead of converting to 420?
It should go to RGB right now, and not to 4:2:0, which is imho the most sensible solution.

hissatsu
29th January 2012, 15:34
Hi,

I have tested on Terminator Salvation (on MediaPortal with LAV Splitter ofc).
I didn't have your issue, i have keep running 2 switching on m2ts (quite simple on that BD, one is @ 53 seconds and the second @ 1 minute 56 secs.)
Running on NVidia with CUDA.
I will try to do the same test with my Avatar seamless BD and on MPC-HC.
Thanks

*EDIT* : On Avatar, it's ok too (on MePo)

Well, thanks for checking. I'd really like being able to use a version of LAV Filters newer than .39 without have a black screen pop up every while. I guess no one else is experiencing this issue?

nevcairiel
29th January 2012, 15:41
@nevcairiel,

LavSplitter gives false information; it gives 48k/16 info while the original audio is 96k/24.

Here is a sample:

http://www.mediafire.com/?2jygu42sokq5c52

(Tom Petty and the Heartbreakers, Blu-Ray Audio.)

Thats DTS-HD, its normal for it to only show the DTS core information. Its not important, a DTS decoder does not need that information.
Its only a cosmetic issue, which however won't be fixed until ffmpeg gets DTS-HD support one day.

rica
29th January 2012, 16:01
OK, thanks!

msoltyspl
29th January 2012, 16:25
Two issues I noticed while remuxing things:

1) I noticed that LAV tends to "flatten" chapter editions in matroska - when there's more than 1 available, for example:

edition1
+chapt1
+chapt2
+chapt3
+chapt4

edition2
+chapt2x
+chapt3x

Now if say - chapt2x comes between edition1's chapters 2:3 and chapt3x between 3:4 - then when LAV splitter is active, it will present:

+chapt1
+chapt2
+chapt2x
+chapt3
+chapt3x
+chapt4

2) Another thing noticed - it looks like EditionFlagOrdered is generally ignored (the behavior is as if EditionFlagOrdered=0 - so no "playlist"-like behavior, though only haali does it, and not without certain quirks either).

I can imagine both things would be far away on a todo list, though at some point perhaps.

Pat357
29th January 2012, 17:24
Not correct depends on the Hardware (Driver), on SB its not the case (EVR Custom)
Not sure whar you mean...Will you still get HW deinterlacing when outputting RGB to EVR-CP on SB than ?

STaRGaZeR
29th January 2012, 17:27
It should go to RGB right now, and not to 4:2:0, which is imho the most sensible solution.

Ugh, I fail. I wanted to say yuv422 if you disable all 422 formats. You said you were going to consider it, and since it's still prefering 420 over 444 or RGB, I'm asking again :p

mindbomb
29th January 2012, 17:50
Not sure whar you mean...Will you still get HW deinterlacing when outputting RGB to EVR-CP on SB than ?

actually, i think the same thing may be happening on my amd e350.

I'm not sure what is happening tbh, im noticing the fps is doubled, so there must be at least a bob going on.

renq
29th January 2012, 19:15
http://forum.doom9.org/showthread.php?p=1554567#post1554567 ;)

Pat357
29th January 2012, 19:18
Can anyone tell me the difference between P210/v210 ?
These are both 10bit and both 4:2:2....
How is the conversion done from one to the other ?
Is this a 100 lossless process ? (think it is, but want to be sure)

Similar for P410 / v410 ?

nevcairiel
29th January 2012, 19:21
Of course its a lossless process. v210 is just for some post-processing filters that don't understand P210. If there is a choice, always use P210, output of that is most likely faster. :)
v210 is a highly packed format, while P210 wastes a few bits to make it easier to work with.

nevcairiel
29th January 2012, 19:31
Ugh, I fail. I wanted to say yuv422 if you disable all 422 formats. You said you were going to consider it, and since it's still prefering 420 over 444 or RGB, I'm asking again :p

Oh i forgot about that.
I still am not sure why anyone would turn that off, but okay.

I'll make it prefer 4:4:4 and then RGB, i suppose? Even though 4:4:4 YUV is so badly supported by renderers ... i wish EVR would just deny connection.

STaRGaZeR
29th January 2012, 19:39
444 then RGB is what I'd ideally do, but with the AYUV situation... vote for RGB over AYUV. Or RGB only, mind you.

nevcairiel
29th January 2012, 20:12
I today implemented YV24 output, which i could use as the primary 4:4:4 output and disable AYUV by default - avoiding the problem entirely (EVR doesnt connect to YV24)
Sadly, madVR seems to have a bug with YV24, causing it to show wrong colors, so i didn't enable it yet for anything. But if its fixed, i may do the switch.

On the other hand, if someone turns off 4:2:2 but keeps AYUV active, its their own damn fault imho. :)

CruNcher
29th January 2012, 21:28
@Nev
im pretty sure you shouldn't fix this but could you if it is needed make a workaround for this based on CoreAVC connecting im not sure tough what would happen if they fix it in the next version and lav splitters workaround tries it to for the current (detecting version before connection seems impossible, or really isn't something a splitter should do) ;)

https://forum.doom9.org/showpost.php?p=1554111&postcount=6977

oddball
29th January 2012, 22:12
Hi again. I've lost my advanced sub settings (I should have saved it someplace). I can't get subs to work how I want. I want only forced subs for English audio. Full subs for Non-English audio and Full subs for anything undertermined like where someone has muxed with both audio and subs set to Und.

CruNcher
29th January 2012, 22:34
@Nev
another stall

http://www.mediafire.com/?5shk3hmbudh8r0j

nevcairiel
29th January 2012, 22:47
@Nev
another stall

http://www.mediafire.com/?5shk3hmbudh8r0j

That file has a broken audio track which causes it (a file with video but only broken audio is problematic)

I changed it now to completely ignore any broken streams, hope this doesn't cause any valid streams to stop playing, but that is a risk i'll just have to take.

CruNcher
29th January 2012, 22:58
That file has a broken audio track which causes it (a file with video but only broken audio is problematic)

I changed it now to completely ignore any broken streams, hope this doesn't cause any valid streams to stop playing, but that is a risk i'll just have to take.

We'll see :)

and now the other way ;) (no audio lav audio until you seek, if you don't seek it stalls @ 1:05)

http://www.mediafire.com/?e2djjxeylkmrsbf

ryrynz
29th January 2012, 23:08
I actually found the reason why ffdshow seems faster, and "fixed" that, still doesn't explain why 0.22 is faster then 0.24, i'll investigate a bit before i post.

Using your build that fixed the default MT settings and it's boosted the performance of QuickSync 0.24 by quite a lot! It's now within 15% of my 0.22 benches.

CruNcher
29th January 2012, 23:27
Though still @ playback Mpeg-2 DXVA2 copy Back is more efficient then the Intel Quicksync Decoder (still no quality compare) :)
and except that 1 stream crash (works with Quicksync) http://www.mediafire.com/?y2dbekgemoeo28m i couldn't find anything problematic all the streams work i wonder what for Mpeg-2 streams don't work for you Nev or show failures @ decoding (would you share those please) :)

Aspeh
30th January 2012, 00:19
Like someone somewhat explained before, LAV Filters and MadVR are two different things. One is a splitter and decoder package, while the other is a video renderer. You can't really compare them side-by-side. What video renderer are you using where you're saying LAV Filters?

just the bare minimum

MPC-HC , Lav Filters and MadVR

Just tried Crystalplayer that someone mentioned loads of pages back and that looks awsome with LAV filters and super sampling enabled as you have full control of brightness and contrast without using video overlays but it doesn't work with MadVR to test as there are no options to set external render.

CruNcher
30th January 2012, 00:33
lav filters -mpc-hc -evr custom bicubic -chroma upsampling -sharpen complex 2 (or experimental sharpen shader in the mpc-hc experimental thread, lower gpu usage)

lav filters (rgb32) dxva2 copy back, quicksync,cuvid -mpc-hc -evr custom bicubic -sharpen complex 2 (or experimental sharpen shader in the mpc-hc thread, lower gpu usage)

internal filters - potplayer evr custom - sharpen complex 2 -chroma upsampling- (or experimental sharpen shader in the mpc-hc thread,lower gpu usage)

On SB
lav filters, dxva 2 copy back (mpeg-2), intel quicksync rest or dxva native capable decoders (everything) -> Mpc-HC - EVR (default) internal Lanczos4 and Chroma upsampling (its really efficient and 0 cost, fixed function) (deinterlacing is not bad needs further comparisons, quality/resource its no catch and clearly intel the winner (fixed function), compared to the massive cpu usage of yadif on 30i->60p) - internal driver sharpen (automatic though most of the times to harsh, fixed function it compensates the quantization blur efficiently)
internal filters (full dxva) -> Potplayer EVR default internal Lanczos4 and Chroma upsampling - internal driver sharpening - (automatic though most of the times to harsh)

sneaker_ger
30th January 2012, 01:12
Hi again. I've lost my advanced sub settings (I should have saved it someplace). I can't get subs to work how I want. I want only forced subs for English audio. Full subs for Non-English audio and Full subs for anything undertermined like where someone has muxed with both audio and subs set to Und.

eng:eng|f;eng:off;*:eng;und:und
or maybe
eng:eng|f;eng:off;*:eng;*:*

Pat357
30th January 2012, 01:31
I today implemented YV24 output, which i could use as the primary 4:4:4 output and disable AYUV by default - avoiding the problem entirely (EVR doesnt connect to YV24)
What is the problem with EVR/AYUV?
I just looked at a 10s intro converted to AYUV, but I did not notice a problem. I could judge the colors though, as I don't know how this intro it should look.
Sadly, madVR seems to have a bug with YV24, causing it to show wrong colors, so i didn't enable it yet for anything. But if its fixed, i may do the switch.
Not sure if it has to do with something I noticed a while ago : with Y24 output, when the colors were OK in FFplay/Mplayer2, I had to swap the U en V to get the right colors in MadVR.
Because I didn't gave anything else that could understand YV24, I was not sure which was wrong and which was right.

On the other hand, if someone turns off 4:2:2 but keeps AYUV active, its their own damn fault imho. :)[/QUOTE]
Why would it be so bad ??

Anyway, I prefer YV24 over AYUV because Avisynth supports only YV24.
If outputting in RGB32 does the HW adaptive deinterlacing still work in EVR on NVidia/AMD/Intel ? What about Madvr, does it still receive the necessary flags to deinterlace and do IVTC ?

mindbomb
30th January 2012, 02:32
I have noticed something very strange regarding dvd playback.
For dvd playback, I use microsoft dtv-dvd video decoder, since the dvd navigator doesnt connect to lav video. I am using lav splitter, lav audio, and madvr as well.

But here is what I don't understand. I will get a macrovision error normally. However, when I add ffdshow audio processor, DVDs play fine.

What is happening?

CruNcher
30th January 2012, 02:54
@Nev

last frame gets lost (display) with yadiff enabled ?

http://www.mediafire.com/?mbuq1zevy5lv59u

[ReX]
30th January 2012, 05:29
I started watching a very old XviD file and I noticed choppy playback. I tested with ffdshow, Koepi's XviD and MPC-HC intenal XviD decoder and they played it with no problems.

Here's a sample (http://www.mediafire.com/?cujlmub3fh443v0) I cut from it.

nevcairiel
30th January 2012, 08:04
@Nev

last frame gets lost (display) with yadiff enabled ?

http://www.mediafire.com/?mbuq1zevy5lv59u

I was wondering when someone would notice. :)
Its a known issue, the yadif filter lacks a function to flush remaining frames, or i just don't know how. :p

what for Mpeg-2 streams don't work for you Nev or show failures @ decoding (would you share those please) :)

http://www.mediafire.com/download.php?79f42499cc12964

Shows random corruption during decoding, seems to happen with all GPUs.
CUVID is fine, as is the MS DXVA decoder.

nevcairiel
30th January 2012, 08:21
What is the problem with EVR/AYUV?
I just looked at a 10s intro converted to AYUV, but I did not notice a problem. I could judge the colors though, as I don't know how this intro it should look.

All formats EVR supports depend on hardware support, and hardware support for AYUV is spotty at best.
On many GPUs it'll either be inredibly slow because some sort of software emulation gets activated, or it'll just show a black screen.

In MPC-HC in particular, AYUV just breaks completely if you have some of the more "advanced" EVR-CP settings activated.

madshi
30th January 2012, 08:38
just the bare minimum

MPC-HC , Lav Filters and MadVR

Just tried Crystalplayer that someone mentioned loads of pages back and that looks awsome with LAV filters and super sampling enabled as you have full control of brightness and contrast without using video overlays but it doesn't work with MadVR to test as there are no options to set external render.
First of all comparing LAV Filters <-> madVR is like comparing a harddisk with a graphics card. LAV Filters is a splitter/decoder. madVR is a renderer. That are 2 totally different things. If you use LAV Filters, only, you get no visible image on screen at all. For that you will need a renderer. If you don't use madVR as the renderer, some other renderer will be used instead. So the comparison you're doing is not LAV Filters <-> madVR. Instead it's madVR <-> some other renderer. LAV Filters got nothing to do with that.

With any renderer, the first thing you should do is to setup the black/white level correctly. Which GPU are you using? When using AMD, set the pixel format to "RGB Full" (or whatever it is named) in the AMD control panel. When using NVidia, make sure you use a custom resolution. Doing this will switch the GPU to full range RGB output. Then in madVR you can switch between video and PC levels. Make sure you use the correct one.

Setting up contrast (or rather the grayscale) should be done at your display, not the video renderer. Also switch your display to the correct colorspace (BT.709), then there's no need to twiddle around with color controls in the renderer. Finally, if your display is bad and shows grossly wrong colors, you can use madVR + yCMS to calibrate your display. But for that you'd need to have your own meter.

This is more a madVR topic than an LAV Filters topic, though, so this discussion should be moved to the madVR thread.

If you want to compare LAV Filters to something else, you need to compare it to another splitter/decoder. If you want to do that, make sure you use the same video renderer. If you switch all splitter, decoder and renderer at the same time, your testing is mostly useless. You gotta compare one thing at a time.

Aspeh
30th January 2012, 09:19
@madshi

Thanks for taking the time to type that out and explain.

It pretty much confirms what i was thinking that i'm out of my depth with this subject and will just leave things on default until i have learnt abit more on the subject as it all seems like a foreign language at the moment. :)

Pat357
30th January 2012, 12:16
Though still @ playback Mpeg-2 DXVA2 copy Back is more efficient then the Intel Quicksync Decoder (still no quality compare) :)
and except that 1 stream crash (works with Quicksync) http://www.mediafire.com/?y2dbekgemoeo28m i couldn't find anything problematic all the streams work i wonder what for Mpeg-2 streams don't work for you Nev or show failures @ decoding (would you share those please) :)
That might be true for Intel SB, but I still have too much "green screens" on my Nvidia GTX 570 (VP4 decoder) with dvxa cb.

nevcairiel
30th January 2012, 12:27
I think its somewhat amusing how people don't care to report issues but then seem to complain in parts of other posts that things don't work. :)

So, if you think you have a bug in the DXVA2 decoder, please report it properly, including a sample, which OS, which GPU and which driver version.

CruNcher
30th January 2012, 13:55
@Nev
that Mediafire link is dead ?

nevcairiel
30th January 2012, 13:59
@Nev
that Mediafire link is dead ?

No file for you, then. :)
I didn't upload it, someone else did.

AmshTemp
30th January 2012, 14:49
@nevcairiel:
Can you make the installer check for the proccessor brand and arch beforce installing IntelQuickSyncDecoder.dll file?
Having an AMD proccessor and Windows 32 bit, I assume IntelQuickSyncDecoder.dll shouldn't get installed at all

nevcairiel
30th January 2012, 14:56
@nevcairiel:
Can you make the installer check for the proccessor brand and arch beforce installing IntelQuickSyncDecoder.dll file?
Having an AMD proccessor and Windows 32 bit, I assume IntelQuickSyncDecoder.dll shouldn't get installed at all

Maybe i could, but a 100-200kb DLL won't hurt you, so i won't.
Its not worth the effort, there is nothing to gain.

CruNcher
30th January 2012, 15:10
@Nev
coming back to this

https://forum.doom9.org/showpost.php?p=1554600&postcount=8624

im not sure anymore if this is actually CoreCodecs fault :(

i have more of these cases and i tested them with AV Splitter and it doesn't show this specific behavior for every of them i really wonder what he does different compared to all the other splitter so that CoreAVC + Lav Audio handles it correct ?

Though therefore his way of doing it seems to get unsync with Lav Video + Lav Audio a little

nevcairiel
30th January 2012, 15:11
im not sure anymore if this is actually CoreCodecs fault :(


One decoder fails, all others work, its their fault.
Taking CoreCodecs communication policys into account, i really don't care to support them either way.

nevcairiel
30th January 2012, 15:16
So, here i was thinking about an audio mixer, mostly how to make it properly configurable, and i hit somewhat of a wall.

I'm thinking about 2 modes, a simple and a advanced mode.

The simple mode would just offer downmixing to 5.1 or stereo, with the appropriate controls (front/back balance, center level)
This really isn't a big problem, and should work out just fine.

The advanced mode should offer a full matrix to toy with.
Here is where the problems come in. To make such a full matrix truly useful, don't you need one matrix per input channel configuration?
I'm thinking, i would want a separate matrix for 7.1, which i want to downmix to 5.1, and i want a separate matrix for stereo, which i want to upmix to 5.1, and of course i want a 5.1 matrix, which shouldn't touch the audio at all. Not to mention all the configurations in between (although i could ponder ignoring them, and converting everything to a "standard" layout before putting it through the matrix)

Looking at ffdshow for example, it doesn't allow this except if you start using profiles. So is this something people really would use? Or would one matrix thats somehow applied to everything just be fine?
Maybe i'm missing something, and ffdshow somehow stores different matrices, but doesn't really explain that on the configuration dialog?

AmshTemp
30th January 2012, 15:31
Fair enough, but what about *64-bit* IntelQuickSyncDecoder.dll installed in my *32-bit* Windows ?

nevcairiel
30th January 2012, 15:32
Fair enough, but what about *64-bit* IntelQuickSyncDecoder.dll installed in my *32-bit* Windows ?

Thats a bug i forgot to fix. :)

madshi
30th January 2012, 15:42
So, here i was thinking about an audio mixer, mostly how to make it properly configurable, and i hit somewhat of a wall.
Having one matrix available for every input channel configuration would be a very powerful but also complicated feature, I guess. I'm not really sure I can say anything about whether it should be done or not, and if how. E.g. what happens if the user changes one matrix (e.g. boost LFE volume)? Should the change also be transported somehow to the other matrixes? Probably not, but that means the user would have to do such changes in all matrixes manually to make it valid globally. Next problem: Should changes always be stored? Maybe the user sometimes wants to change something just for one audio track?

E.g. one thing I sometimes wish for is an easy way to just boost the center channel volume (and only temporarily for one playback session), because some audio tracks have the dialog too low volume. That's the only thing I'm missing right now in LAV Audio.

Midzuki
30th January 2012, 15:46
...

The advanced mode should offer a full matrix to toy with.
Here is where the problems come in. To make such a full matrix truly useful, don't you need one matrix per input channel configuration?

Not sure if this suggestion is "easily-doable" :o , but...

Only one "full-featured matrix"
(à la Channel Downmixer :devil:) ,

plus five "output modes" (2.0, 4.0, 5.1, 7.1, and "as-is").

CruNcher
30th January 2012, 15:51
One decoder fails, all others work, its their fault.
Taking CoreCodecs communication policys into account, i really don't care to support them either way.

More decoder fail its a very different situation (but yeah it seems to be a another special fix from Kytmanov this time for CoreAVC)

recover-sample.ts

Lav Splitter + Lav Audio (No AV sync correction)

CoreAVC = Stutter (after the glitch)
Arcsoft = unsync
Cyberlink = OK
Lav Video = OK
DivX H.264 = Crash (@ the glitch) deterministic
Microsoft DTV = OK
ffdshow quicksync (libavcodec) = OK
ffdshow quicksync (quicksync) = unsync
Mainconcept SDK9 = OK

AV Splitter + Lav Audio (No AV sync correction)

CoreAVC = OK
Arcsoft = unsync
Cyberlink = unsync
Lav Video = a little unsync
DivX H.264 = flys (accelerates) away
Microsoft DTV = Heavy unsync
ffdshow quicksync (libavcodec) = OK
ffdshow quicksync (quicksync) = unsync
Mainconcept SDK9 = flys (accelerates) away

Haali Splitter + Lav Audio (No AV sync correction)

CoreAVC = Stutter (after the glitch)
Arcsoft = little unsync
Cyberlink = recovers slower the sync but does
Lav Video = OK
DivX H.264 = Crash (@ the glitch) deterministic
Microsoft DTV = unsync video way to fast
ffdshow quicksync (libavcodec) = OK
ffdshow quicksync (quicksync) = unsync
Mainconcept SDK9 = OK

MPC-HC Splitter + Lav Audio (No AV sync correction)

CoreAVC = Stutter (after the glitch)
Arcsoft = unsync
Cyberlink = OK
Lav Video = OK
DivX H.264 = Crash (@ the glitch) non deterministic first execution always crashes 2nd doesn't have to
Microsoft DTV = OK
ffdshow quicksync (libavcodec) = OK
ffdshow quicksync (quicksync) = unsync
Mainconcept SDK9 = OK

Potplayer Chain DXVA = a little unsync
Potplayer Chain Software = a little unsync

nevcairiel
30th January 2012, 15:52
one thing I sometimes wish for is an easy way to just boost the center channel volume (and only temporarily for one playback session), because some audio tracks have the dialog too low volume. That's the only thing I'm missing right now in LAV Audio.

From what i'm told, proper downmixing is what most of the people really need/want, because quite a lot of people do only have stereo speakers. That feature is relatively easy, as i can just programmatically build up the matrix based on the input channels.
So maybe i'll postpone the full matrix a bit to figure out how to do it.

Per-channel volume is another matter which would probably be easy to implement, but especially if you want "temporary", it again offers a configuration challenge. It should be easily accessible, preferably over a shortcut even, and of course be temporary. I don't even have a notion of temporary settings yet. :)

I'm just assembling a list of things that all go into the audio mixing category, so we'll see where it goes.

madshi
30th January 2012, 16:02
FWIW, I wouldn't even need per-channel volume settings for all channels. Just a way to boost the center channel would be nice. You could even name it "dialog boost", or something like that, and probably it would be good enough to have only 1-2 boost steps available, so a simple keyboard shortcut could just toggle between the 2-3 possible states (e.g. "off", "small boost", "big boost") . I'm not sure if this really fits in with all the other mixer features, because the other stuff would probably not be temporary settings and also would not need shortcuts. But probably it makes sense to implement it all in one go.

Thanks for considering this... :)

Midzuki
30th January 2012, 16:21
Anyway, just be careful enough to NOT repeat Yamagata's lame workaround :D

http://forum.doom9.org/showthread.php?p=1057543#post1057543

P.S.: also, the volume indicators should (finally) be as fast as the ones in AC3Filter ;)

wanezhiling
30th January 2012, 16:39
audio mixer? Great!
If possible, add normalizer together~~:p


PS: nev, would you like to add this (http://i.imgur.com/I7b8V.png) like mpc-hc for the installer

Mercury_22
30th January 2012, 16:56
...
Per-channel volume is another matter which would probably be easy to implement, but especially if you want "temporary", it again offers a configuration challenge. It should be easily accessible, preferably over a shortcut even, and of course be temporary. I don't even have a notion of temporary settings yet. :)

I'm just assembling a list of things that all go into the audio mixing category, so we'll see where it goes.

So maybe you can start with "Per-channel volume" with ON / OFF and without any other "fancy" options and then go to more challenging options :p

adam777
30th January 2012, 17:07
Regarding FFDShow, as far as I can tell, without the use of profiles, you can choose between 2 options:
- Specifying an output format, in which case it's FFDShow's job to create the matrix in regards to the input.
That is, if you specified stereo as the desired output, a stereo input will create a "no change" matrix, a 5.1 input will send front and back channels to relevant stereo channel and center will be sent to both (all in regards to normalization etc.).
This mode is probably the most useful for most.

- Another option would be specifying a custom matrix, and considering that output format in a specific machine usually does not change, the only difference it has is in regards to volume levels.
If you chose to normalize your custom matrix, you take into account channels that may not be present in the input.

Bottom line, I would be perfectly fine with the first option.

STaRGaZeR
30th January 2012, 18:30
The advanced mode should offer a full matrix to toy with.
Here is where the problems come in. To make such a full matrix truly useful, don't you need one matrix per input channel configuration?
I'm thinking, i would want a separate matrix for 7.1, which i want to downmix to 5.1, and i want a separate matrix for stereo, which i want to upmix to 5.1, and of course i want a 5.1 matrix, which shouldn't touch the audio at all. Not to mention all the configurations in between (although i could ponder ignoring them, and converting everything to a "standard" layout before putting it through the matrix)

If I understand this correctly, I think ffdshow can do this without profiles. I convert everything to 4.0, my matrix upmixes stereo to 4.0 plus downmixes 5.1 or 7.1 to 4.0 . Of course it doesn't allow for different coeffs for each conversion, but is that really needed?

nevcairiel
30th January 2012, 18:31
Of course it doesn't allow for different coeffs for each conversion, but is that really needed?

Thats my question, right there.
What ffdshow does is give you one matrix to rule them all.

STaRGaZeR
30th January 2012, 18:38
Yep. Then I don't need it, but having it won't do any harm I guess. It all depends on how complicated you want to make it then :p

DrNein
30th January 2012, 18:40
Center-channel boost should effectively be handled by Dynamic Range Compression (existing LAV Audio option). Otherwise, Windows Sound Device (Levels Balance). Or for bitstreaming, the external processor/receiver.

nevcairiel
30th January 2012, 18:49
The DRC option in LAV Audio only activates the DRC in the AC3 decoder, which is only present if the file was actually encoded to have DRC - and even then, is not very strong.

clsid
30th January 2012, 18:53
A fully custom matrix is not something you should worry about right now imo. There is probably few people actually using that right now in ffdshow.

Just dynamically construct a matrix based on the output a user has selected, plus a few additional options such as:
- boost center when downmixing to stereo (or maybe everything below 5.1 like quadro as well)
- expand stereo to center/surround
- gains for individual input channels, apply to current session only?

amtm
30th January 2012, 19:06
Center-channel boost should effectively be handled by Dynamic Range Compression (existing LAV Audio option). Otherwise, Windows Sound Device (Levels Balance). Or for bitstreaming, the external processor/receiver.

That's great but if you aren't playing ac3 (throws out option 1), you aren't bitstreaming or using an external receiver (throws out option 3) and you don't want to constantly have to be changing the levels balance before and after playing a file because that's a pain in the ass when all you want is a temporary boost when playing a file (throwing out option 2) it's quite nice to have built-in functionality that will do it for you automatically.

iSunrise
30th January 2012, 19:24
... I convert everything to 4.0, my matrix upmixes stereo to 4.0 plus downmixes 5.1 or 7.1 to 4.0 . Of course it doesn't allow for different coeffs for each conversion, but is that really needed?
Same here.

I only want to set one speaker configuration (like 2.0) for everything (because that´s how my speakers are setup) and regardless of the source channel-configuration, it should always downmix it to 2.0. If in the rare case there is a 1.0 source, it should also be upmixed to 2.0.

You should not make this too complicated at first. If some users are not satisfied with your first implementation, there´s always the possibility to give you feedback for consideration.

fairchild
30th January 2012, 19:28
I'm curious as to what is the main benefit of having the decoder do the mixing (down or upmix). For example if you are using headphones/2.1 channel speaker system, and you set Windows to 2.0 doesn't the windows sound mixing do an adequate job or is there some benefit in this case to let the decoder do the mixing first? Also, I can see the benefit for say a 5.1 speaker system and having the decoder upmix a 2.0 track to 5.1 but otherwise can someone enlighten me?

clsid
30th January 2012, 19:33
Yes, main reason is because the Windows mixer doesn't give satisfying results.

mindbomb
30th January 2012, 20:26
imo, the mixer is nice, but what would really work well with it is a new audio renderer as well.

One with ASIO and wasapi and other cool features.

That might be the start of a paradigm shift of doing audio decoding and processing on the cpu instead of the receiver in the htpc world.

mbordas
30th January 2012, 21:42
imo, the mixer is nice, but what would really work well with it is a new audio renderer as well.

One with ASIO and wasapi and other cool features.

That might be the start of a paradigm shift of doing audio decoding and processing on the cpu instead of the receiver in the htpc world.

hah, yes, I've been doing this for awhile now - and the thing that really made it possible? Lav filters.

I use a multichannel amp. No extra remote, no bitstreaming nonsense, everything controlled through the htpc. I don't even have to look at the thing, it's hidden in my basement.

But to your point: Reclock is fine but I suppose it could always be better. But having a mixer in lav audio, or even the center channel boost, would be great.

tetsuo55
30th January 2012, 22:06
So, here i was thinking about an audio mixer, mostly how to make it properly configurable, and i hit somewhat of a wall.

I'm thinking about 2 modes, a simple and a advanced mode.

The simple mode would just offer downmixing to 5.1 or stereo, with the appropriate controls (front/back balance, center level)
This really isn't a big problem, and should work out just fine.

The advanced mode should offer a full matrix to toy with.
Here is where the problems come in. To make such a full matrix truly useful, don't you need one matrix per input channel configuration?
I'm thinking, i would want a separate matrix for 7.1, which i want to downmix to 5.1, and i want a separate matrix for stereo, which i want to upmix to 5.1, and of course i want a 5.1 matrix, which shouldn't touch the audio at all. Not to mention all the configurations in between (although i could ponder ignoring them, and converting everything to a "standard" layout before putting it through the matrix)

Looking at ffdshow for example, it doesn't allow this except if you start using profiles. So is this something people really would use? Or would one matrix thats somehow applied to everything just be fine?
Maybe i'm missing something, and ffdshow somehow stores different matrices, but doesn't really explain that on the configuration dialog?What I am looking for a set and forget setting for my speakercount.
So if I have 5.1 up and downmixing would occur automatically depending on the source. same for stereo or any other physical config.

pirlouy
30th January 2012, 22:09
For me, Matrix should be linked to profiles; and profiles refer to post-processing, so not needed for LAV.
I would just add some options often asked (center boost; 2.0 -> 4.0; crossover for subwoofer at 80Hz; convert all to stereo, 7.1 -> 5.1).

For fine-tuning, it would be useful to have a post-processing plugin for LAV, which could handle profiles. Of course, the plugin would not be created by you since you already said you're not interested, but it would be nice if you add an API for that (new challenge for developers).

But maybe it's useless, and people just need a new "ffdshow processor" which would start with good bases...

JMGNYC
30th January 2012, 23:43
I'm having a problem playing back VC-1 direct BD rips to MKV (no re-encoding so high bitrate). Using all version, including .45, chapter skip, or even a 30 second skip takes close to a minute before playback resumes not using any HW acceleration. I've an ATI 5570 so I've tried DXVA copy back with .45 and now when skipping forward or back I blue screen in Win7 64. I'm using 32 bit MPC-HC and MadVR. Any suggestions?

oddball
31st January 2012, 01:30
LAV Filters is not playing nice with ReClock. If I use LAV video decoder and ReClock MPC-HC hangs 8 out of 10 times. If I use ffdshow video decoder (still using LAV splitter) I get no hangs. Any ideas apart from not using ReClock (Not an option actually I find it essential)?

Hmm. This seems to be more of an issue with fullscreen switching. In the past I used to do it manually. Now I have MPC-HC do it for me. I usually let my secondary display fallback to 60Hz and when it plays a 23,756fps source it auto switches to 23/24Hz on the Nvidia secondary head. Likewise for PAL @ 50Hz etc. However if I leave the secondary display at 23/24Hz instead it does not hang on most sources (because it is already at the correct refresh rate) and 50Hz switching does not hang either. I have not tried 29/30/60fps stuff yet to see if it hangs switching to 60Hz that is up next. If it does not hang I will leave the display at 23/24Hz. But I stress this only occurs if I use LAV Video decoder in conjunction with ReClock so is a weird bug indeed.

oddball
31st January 2012, 02:38
eng:eng|f;eng:off;*:eng;und:und
or maybe
eng:eng|f;eng:off;*:eng;*:*

The second one seems to be the one. Thanks.

omarank
31st January 2012, 08:26
Feature request: An LFE crossover setting in audio mixer would be great.


:thanks:

nevcairiel
31st January 2012, 08:30
Feature request: An LFE crossover setting in audio mixer would be great.


Not going to happen anytime soon, doing a low-pass filter for LFE is quite a bit more work then just mixing things.

Aspeh
31st January 2012, 09:23
nevcairiel

Don't stop at LAV mixer , Team up with madshi and code a new Media Player. :)

Though what you would call it might start arguments.

LavMad Player , MadLav Player or just LM player. :D

nevcairiel
31st January 2012, 09:41
No thanks, i rather focus my time on the projects i already have. :p

aufkrawall
31st January 2012, 10:32
Please forgive me for being pushy, nevcairiel, but what about the FRAPS issue? :)
It always has to be RGB output, it can't fail and the official FRAPS decoder by Beepa does the same.

Wouldn't it be possible to simply force automatic RGB output if source is any FRAPS codec?

nevcairiel
31st January 2012, 10:36
I will never force one output format for any codec, if the decoder outputs YUV, then i'll offer YUV to the renderer.
However, i did change one thing. In absence of any other information, LAV will now assume that FRAPS content is RGB, which means if you're using EVR, it'll stick with RGB because EVR is unable to change its input format. With EVR, this will basically ensure that its always getting RGB with FRAPS video. This was the cause of failure before, because it was converting the RGB to YUV

aufkrawall
31st January 2012, 10:58
I will never force one output format for any codec, if the decoder outputs YUV, then i'll offer YUV to the renderer.

I understand that you don't want to implement such "hacks" or fixes for other's fault. :)


However, i did change one thing. In absence of any other information, LAV will now assume that FRAPS content is RGB, which means if you're using EVR, it'll stick with RGB because EVR is unable to change its input format. With EVR, this will basically ensure that its always getting RGB with FRAPS video. This was the cause of failure before, because it was converting the RGB to YUV
Great. :)
Does this also apply to FRAPS YV12?
It looks wrong too without converting to RGB because it keeps PC levels, unfortunately.
I don't think that in this particular case it would ever have any kind of bad effects to future FRAPS videos.

Alabanda
31st January 2012, 12:31
Great to hear about the audio mixer!

I hope you also implement 'Headphone Virtual Spatialization' along the way. It works much better for (as per name) headphones when it comes to downmixing than simple stereo output. It's alike to Dolby Headphone. (Some say HRTF is even more powerful when it comes to in-ear headphones, but I don't have any and hate them actually :P Virtual Spatialization works for all.)

Thanks.

roytam1
31st January 2012, 13:20
I understand that you don't want to implement such "hacks" or fixes for other's fault. :)


Great. :)
Does this also apply to FRAPS YV12?
It looks wrong too without converting to RGB because it keeps PC levels, unfortunately.
I don't think that in this particular case it would ever have any kind of bad effects to future FRAPS videos.

if you have those samples, try them out with nightly build:
http://roy.orz.hm/lavf-w32-nightlies/lavf-my120131-b26ac02.7z

nevcairiel
31st January 2012, 13:32
I hope you also implement 'Headphone Virtual Spatialization' along the way. It works much better for (as per name) headphones when it comes to downmixing than simple stereo output. It's alike to Dolby Headphone. (Some say HRTF is even more powerful when it comes to in-ear headphones, but I don't have any and hate them actually :P Virtual Spatialization works for all.)


I have no plans to implement anything beyond what can be done with a default matrix mixer.

rsd78
31st January 2012, 16:51
Hi Nevcairiel,

I know this is a longshot request, but I thought I'd at least ask :). I am a long time user of Albain's Media Control plugin (I'm a WMC user). It is only compatible with Haali and Gabest splitter. I believe he open sourced the code, but I believe Albain has been MIA for some time. Is there any chance it'd be possible to make LAV Splitter compatible with Media Control? I don't have a sense of how hard it is, but I recall Albain adding Gabest support a while back in a relatively short period of time. I know you offer subtitle control via a regex of sorts, but I really like the GUI Media Control offers via the WMC player. I completely understand if its either not feasible or you are not interested however. I'd also accept an alternative GUI via WMC instead of Media Control that works with WMC instead as well :) (Kidding, but I thought I'd throw that out there).

I'd love to use Lav Splitter as it seems much more stable than Haali/Gabest when I played with it, but I really like the Media Control plugin/GUI.

Thanks Again

Andy o
31st January 2012, 17:20
I hope you also implement 'Headphone Virtual Spatialization' along the way. It works much better for (as per name) headphones when it comes to downmixing than simple stereo output. It's alike to Dolby Headphone. (Some say HRTF is even more powerful when it comes to in-ear headphones, but I don't have any and hate them actually :P Virtual Spatialization works for all.)


To me sounds better too (DH is based on HTRFs though, they mean different things). I'm a fan of DH, and would love it if LAV had it (but alas), cause right now there's no way for a DirectShow player to offer it, except commercial ones. foobar2000 uses the dolby dll for it, the same way LAV uses the dts dll for DTS-HD decoding.

A better option though, that will work with everything, not just the player, is a card that offers it. The Xonar U3 looks like the perfect one for the price. I have a DX on my desktop and U1 on my laptop and like them for it, but bear in mind that many people hate the Xonar cards.

mindbomb
31st January 2012, 17:28
yea, the dolby headphone plugin for foobar is great, although ironically, since its for a medium that is mostly stereo, it is not really necessary there.
Arcsoft also has a dtsheadphone.dll that could probably be used similiarly.

Also, although not alot of people use software deinterlacing anymore probably, I was wondering if there was any plans to multithread the yadif software deinterlacer, like in ffdshow?

e-t172
31st January 2012, 18:21
So, here i was thinking about an audio mixer

Can you please make it so that your mixer honors Dolby Metadata (http://www.dolby.com/uploadedFiles/Assets/US/Doc/Professional/18_Metadata.Guide.pdf) (most notably downmixing coefficients)? This would be a nice feature. It doesn't have to override the user specified matrix, just combine the two.

nevcairiel
31st January 2012, 18:24
Can you please make it so that your mixer honors Dolby Metadata (http://www.dolby.com/uploadedFiles/Assets/US/Doc/Professional/18_Metadata.Guide.pdf) (most notably downmixing coefficients)? This would be a nice feature. It doesn't have to override the user specified matrix, just combine the two.

I can ask the decoder to downmix using the Dolby metadata, or i can downmix myself, i cannot however combine the two.

NikosD
31st January 2012, 18:32
http://www.mediafire.com/?id7cd8aujnmwslf

What do you think is the problem with that file ?
Why did you post it ?

Because with PotPlayer 1.5.31590 and Intel Drivers 2559 I don't see a problem.

e-t172
31st January 2012, 18:38
I can ask the decoder to downmix using the Dolby metadata, or i can downmix myself, i cannot however combine the two.

In that case it would probably be best to have an option to let the decoder downmix, and use user-specified parameters if the decoder doesn't support downmixing. Personally, I don't care about specifying matrixes but I do care about correct downmixing which complies with the Dolby specifications.

Alabanda
31st January 2012, 20:05
To me sounds better too (DH is based on HTRFs though, they mean different things). I'm a fan of DH, and would love it if LAV had it (but alas), cause right now there's no way for a DirectShow player to offer it, except commercial ones. foobar2000 uses the dolby dll for it, the same way LAV uses the dts dll for DTS-HD decoding.

A better option though, that will work with everything, not just the player, is a card that offers it. The Xonar U3 looks like the perfect one for the price. I have a DX on my desktop and U1 on my laptop and like them for it, but bear in mind that many people hate the Xonar cards.

I'm not sure if you already know it, but I'll post for others anyway. ffdshow audio decoder already has such a functionality which is akin to Dolby Headphone. (I meant that in my previous post.)

It's in the Mixer submenu. For 'Output speakers configuration:' choose 'headphone virtual spatialization' and it'll be on. Use this only for 5.1+ audio input though, it's no good for stereo :P

Sad to hear you won't implement this nevcairiel. I wonder if I can get a VST host to work system-wide.

kalston
31st January 2012, 21:03
I'm not sure if you already know it, but I'll post for others anyway. ffdshow audio decoder already has such a functionality which is akin to Dolby Headphone. (I meant that in my previous post.)

It's in the Mixer submenu. For 'Output speakers configuration:' choose 'headphone virtual spatialization' and it'll be on. Use this only for 5.1+ audio input though, it's no good for stereo :P

Sad to hear you won't implement this nevcairiel. I wonder if I can get a VST host to work system-wide.

I tried it and didn't like it tbh. I'm not the biggest fan of Dolby Headphones either (annoying bass boost among other things).

I use JRiver MC with Isone Surround Free VST ( http://www.toneboosters.com/tb-isonesurround/ ) and I like it much better.
I think it's possible to use it in ffdshow (AFAIK there's a way of using VST plugins in ffdshow) and it's usable in foobar2000 too. It works not only with a 5.1 source but also with a 2.0 one, making it sound like it's played on two speakers in front of you (i.e. it doesn't upmix it).

Back on topic I don't really need the mixer to do anything other than "proper" down/upmixing, if that even means anything. I just want it to convert everything to a specified format (like upmixing mono to stereo and downmixing everything else to 2.0).

One thing I don't understand is that downmixing (to stereo) with ffdshow results in a higher volume (and clipping on some tracks) while when downmixing with JRiver MC's thing or my Asus drivers the volume is lower but it is also clipping-free on those same tracks. Does that mean they do a better dowmixing or that they simply do it differently?

aufkrawall
31st January 2012, 23:00
if you have those samples, try them out with nightly build:
http://roy.orz.hm/lavf-w32-nightlies/lavf-my120131-b26ac02.7z
Both RGB and YV12 videos look right with that LAV Decoder version at default settings (and EVR), so nevcairiel has fully solved it. :)

fairchild
31st January 2012, 23:34
@kalston and Alabanda, ffdshow also has a feature called Head-related transfer function (HRTF) which seems to do a better job than headphone virtual spatialization and of course can be used in conjunction with Lav Audio on headphones. It sounds superior to myself than the former but still debateable if it sounds better than just a proper 2.0 downmix/upmix.

CruNcher
1st February 2012, 04:12
@Nev
seek stall problem

http://www.mediafire.com/?4n6j8ut3gv72tql

MPC-HC Splitter = work
Av Splitter = work
VLC = work
mplayer = work
ffplay = work
avplay = work
Lav Splitter = fail (stalls after seek)
Potplayer = fail (cant seek @ all)

i guess i continue with seek issues (especially stalls) as i have no playback ones left for .ts/.mpg ;)

zerowalker
1st February 2012, 04:37
LavFilter "Splitter" set to AVI and Thumbnail on AVI causes COM Surrogate to crash.

CruNcher
1st February 2012, 04:42
you have to find out @ which file it's crashing @ parsing, or do you mean any .avi file parsing causes it to crash ?

Ok im through with seek testing amazing how stable .ts/.mpg is not 1 crash how hard i tried it (H.264,Mpeg1,Mpeg2,VC-1) :D seeking was always no problem sync also was always keeped nicely after seek the only stream that showed some strange effect after seek was a old known one Hilary.0000.ts it shows a strange acceleration effect to keep the sync something not visible when using Av Splitter on it though MPC-HC splitter fails rather hard on it sometimes stalls after seek actually the first seek always stalls.

But the rest was absolutely no problem back /forth long/short and partially scrubbing seeks very stable :) also fast forward is possible upto 4x times it seems though restricted to 4x times for everything ;)

This is no rating on Performance as the current setup wont allow it and also performance is imho with something like that not a priority :P

so i leave *.ts/.mpg with the last 2 issues behind me now ;)

https://forum.doom9.org/showpost.php?p=1554629&postcount=8628 <- no audio until seek else stall
https://forum.doom9.org/showpost.php?p=1555128&postcount=8700 <- no seek possible (stalls after seek doesn't recover)

@Nev
Why did you leave High Profile predictive 4:4:4 active for DXVA2 Copy Back, you should let it fall back to Libavcodec like it does with Quicksync,Cuvid already, unless of course you developed something to reconstruct the Intel Decoder output (the Grey shadow chroma reconstruction algorithm from the nothing) ;)

nevcairiel
1st February 2012, 07:47
@Nev
Why did you leave High Profile predictive 4:4:4 active for DXVA2 Copy Back, you should let it fall back to Libavcodec like it does with Quicksync,Cuvid already, unless of course you developed something to reconstruct the Intel Decoder output (the Grey shadow chroma reconstruction algorithm from the nothing) ;)

Fallback works fine for me.
Its quite clever, ffmpeg itself will only ever try to use the DXVA2 decoder on 4:2:0 8-bit content.

In addition to that, i of course also have my media type checks that all decoders share.

CruNcher
1st February 2012, 08:07
Strange wonder why it fails on this lossless stream .mp4 then here
anyways :)

http://www.mediafire.com/?co6fcj11mvaxfw4

Lav Video = Microstutter
ffdshow = smooth

http://www.mediafire.com/?e0v46wp9kzvi09n

Lav Splitter + Lav Video = Black Screen (Audio Plays)
vlc = works
ffplay = works
avplay = works

aufkrawall
1st February 2012, 08:58
A seeking issue with LAV Decoder and MPC HC Ogg Splitter:
http://tagesschau.vo.llnwd.net/d3/video/2012/0131/TV-20120131-0713-2501.webm.ogv
The file doesn't start playing without some manual seeking.
It's fixed with the newest MPC HC and its internal decoder, but not with LAV.

nevcairiel
1st February 2012, 09:29
A seeking issue with LAV Decoder and MPC HC Ogg Splitter:
http://tagesschau.vo.llnwd.net/d3/video/2012/0131/TV-20120131-0713-2501.webm.ogv
The file doesn't start playing without some manual seeking.
It's fixed with the newest MPC HC and its internal decoder, but not with LAV.

Technically, thats a bug in the MPC-HC Ogg Splitter.
The only reason it works is that MPC-HCs own decoder has a check to disable multi-threaded Theora decoding when its own Ogg splitter is used (because it doesn't work). If you disable multi-threading in LAV, it'll also start working.

I'll add the same check, i suppose.
It would of course never have been a problem if you just used LAV Splitter. :)

Aleksoid1978
1st February 2012, 12:09
http://www.mediafire.com/?e0v46wp9kzvi09n

Lav Splitter + Lav Video = Black Screen (Audio Plays)


Bug in parser - incorrect detect JPEG codec as MPEG-4(mp4v)

Aleksoid1978
1st February 2012, 12:19
nevcairiel - can you help fix bug in MPC Ogg Splitter - to support MT in Theora ???

CruNcher
1st February 2012, 12:25
@Nev
line error

http://www.mediafire.com/?5xufemxwx0fwpcn

Lav Video = fail
ffdshow = ok

Fourcc RAW cannot be decoded

http://www.mediafire.com/?lzqbdjhbp61mrjz

vlc = fail
avplay = ok
ffplay = ok

pirlouy
1st February 2012, 13:11
Would be easier to help you integrating LAV with MPC HC...

CruNcher
1st February 2012, 13:25
Wow just as a side note with Lav Splitter + Cyberlinks Decoder in DXVA mode you can scrub ultra fast through .mkv with VC-1 it's so fast that the position marker in MPC-HC cant update with it and lags behind ;)
It's not possible with Arcsofts Decoder

Have to make a Video of this this is crazy :D

http://www.mediafire.com/?qr6uh0x689t61xl

unfortunately fast forward doesn't reflect the manual seeking performance :(

Pat357
1st February 2012, 15:50
Hi Nev,

I went through a part of my test-video collection and found some files that play OK with CUVID, but give a green screen with DXVA copy-back :

1) mpg : http://www.mediafire.com/?s9tcp2fjjjrs178
2) x264 in MP4 : http://www.mediafire.com/?d92xafh7c4dda54
3) .TS http://www.mediafire.com/?fr5g7jahhvha5fj
4) x264 in MP4 : http://www.mediafire.com/?htqq7672qz7jttr
5) x264 in MKV : http://www.mediafire.com/?1b8565dr7rfv7y7
6) .ts http://www.mediafire.com/?3z15d95b7j0j3jm
This file crashed my MPC-HC when using DVXA-CB
7) .ts : http://www.mediafire.com/?vg2dpo17r8c7apy
This is a very strange behavior : to get the green screen, disable all CSP's in Lav-video except RGB32, set the output to MadVR
(MadVR should support RBG, right ? Well, have a look then :p
8) WMV3 http://www.mediafire.com/?x4efqsmy1m2bpsw

For the first 7 files, the NV CUVID decoder was effectively used, without problems in playback. (cuvid state was "active").
This should rule out that the green screens come from limitations of my HW (ie unable to decode using HW acceleration).
For the last file (WMV3), the cuvid decoder appeared not being used ....... Should CUVID not use the VC1 "engine" for WMV3 ??

System used :
Core i7 - 970 (6+6 cores) @ 3,7 GHz
24GB DDR3 RAM 2000Mhz
NVidia GTX 570 / VP4 decoder / driver : 285.62
Win7 x64 Prof including all updates
MPC-HC 32bit v1.5.3.4008
Lav filters 0.45 0094f45 from the git-repo

I still have some "stalling" files, ie these won't start playing (hang at the start of the video) and I end up with a black screen.
If I use another splitter, AV splitter or Sonic HD Demuxer instead of Lav-splitter, these play perfect (still using together with Lav-audio & Lav-video)

nevcairiel
1st February 2012, 16:45
I started testing your files, but of course i cannot reproduce a single issue with any of them (and i also have a GTX 570)
Maybe i should downgrade to 285.62 drivers, hm? (Been using the 290 beta)

Also, what exactly does "Green Screen" mean? It stays constantly green, one green frame at the beginning?

PS:
CUVID does not like WMV3, so its not available at all.

Edit:
I downgraded drivers and i get green screens, finally something to work with. ;)

nevcairiel
1st February 2012, 17:00
Hmmm.
I tested the files with the Microsoft DXVA decoder, and it also shows no image (a black screen in its case)

I'm inclined to think its related to this issue, introduced in the 285 driver, fixed in 290.36:


[GeForce 500 series]: Playback of videos or live TV using Windows Media Center results in a black screen. [893436]


I'll see if there is something trivial to fix, but maybe its just a bug in the driver, and not worth my time.
Its sad NVIDIA hasn't released a final driver yet to fix this, but already have a 295 beta now ....

First i'll try the crashing sample, see what happens.

nevcairiel
1st February 2012, 17:25
http://www.mediafire.com/?co6fcj11mvaxfw4

Lav Video = Microstutter
ffdshow = smooth


Either that file is muxed wrong, or MP4 behaves differently with MPEG4/XVID then any other container...
Anyone have some more MPEG4/XVID in MP4 samples to confirm? :d (Needs to be complex enough to have B-frames)

I fixed it assuming that my assumption is right... :)

nevcairiel
1st February 2012, 18:24
Bug in parser - incorrect detect JPEG codec as MPEG-4(mp4v)

LAV properly detects it as MJPEG, but still doesn't play properly.
Must be something the decoder can't deal with. I recommend reporting it to ffmpeg (if ffplay also fails)

clsid
1st February 2012, 19:34
Nev, the rmvb freeze I reported yesterday is related to frame threading. No problems with 1 thread.

Anima123
1st February 2012, 19:38
A little bit OT, but is the motion-vector prediction bug with libav rv decoder still there? If it has been fixed then I can consider getting away with Realplayer completely.

kolak
1st February 2012, 19:56
LavFilter "Splitter" set to AVI and Thumbnail on AVI causes COM Surrogate to crash.

LAV Splitter in case of AVI crashes many of my software.
LAV Decoder crashes graphedit, but works with graphstudio.

Pat357
1st February 2012, 20:11
Edit:
I downgraded drivers and i get green screens, finally something to work with. ;)
Glad you see them too !
I started thinking that something else was wrong with my system...:D

clsid
1st February 2012, 20:12
You need to post sample files if you want a fix.

In case of GraphEdit, that is a bug of GraphEdit itself and can't be fixed. It does not support filters that have many mediatypes. You should use GraphStudio(Next).

Asmodian
1st February 2012, 20:14
LAV Decoder crashes graphedit, but works with graphstudio.

The graphedit crash is its fault. No filter is allowed to have over 99 supported video types in graphedit.

Pat357
1st February 2012, 20:28
LAV Splitter in case of AVI crashes many of my software.
LAV Decoder crashes graphedit, but works with graphstudio.
The Lav filters make Graphedit always crash because graphedit is limited to 100 media-types per filter; the Lav-filters have more then 100 different media-types.

You can only use GraphStudio or the the more recent GraphStudioNext with the lav-filters. (=based on Graphstudio, but more featutes).
Have a look at http://code.google.com/p/graph-studio-next/

nevcairiel
1st February 2012, 20:43
Glad you see them too !
I started thinking that something else was wrong with my system...:D

Yes i do, but i still blame the driver. I tried earlier and newer drivers, and all work just fine, only that one fails.
I've been using the 290 beta driver for a month or so already, and it works just fine, i recommend upgrading. :p

dann23
1st February 2012, 20:50
Hi Nev,

I went through a part of my test-video collection and found some files that play OK with CUVID, but give a green screen with DXVA copy-back :

1) mpg : http://www.mediafire.com/?s9tcp2fjjjrs178
2) x264 in MP4 : http://www.mediafire.com/?d92xafh7c4dda54
3) .TS http://www.mediafire.com/?fr5g7jahhvha5fj
4) x264 in MP4 : http://www.mediafire.com/?htqq7672qz7jttr
5) x264 in MKV : http://www.mediafire.com/?1b8565dr7rfv7y7
6) .ts http://www.mediafire.com/?3z15d95b7j0j3jm
This file crashed my MPC-HC when using DVXA-CB
7) .ts : http://www.mediafire.com/?vg2dpo17r8c7apy
This is a very strange behavior : to get the green screen, disable all CSP's in Lav-video except RGB32, set the output to MadVR
(MadVR should support RBG, right ? Well, have a look then :p
8) WMV3 http://www.mediafire.com/?x4efqsmy1m2bpsw

For the first 7 files, the NV CUVID decoder was effectively used, without problems in playback. (cuvid state was "active").
This should rule out that the green screens come from limitations of my HW (ie unable to decode using HW acceleration).
For the last file (WMV3), the cuvid decoder appeared not being used ....... Should CUVID not use the VC1 "engine" for WMV3 ??

System used :
Core i7 - 970 (6+6 cores) @ 3,7 GHz
24GB DDR3 RAM 2000Mhz
NVidia GTX 570 / VP4 decoder / driver : 285.62
Win7 x64 Prof including all updates
MPC-HC 32bit v1.5.3.4008
Lav filters 0.45 0094f45 from the git-repo

I still have some "stalling" files, ie these won't start playing (hang at the start of the video) and I end up with a black screen.
If I use another splitter, AV splitter or Sonic HD Demuxer instead of Lav-splitter, these play perfect (still using together with Lav-audio & Lav-video)

I had this problem but with an ati card. driver 11.5 worked but newer drivers display green image. after reinstalling windows it works with the latest driver.

And I have another problem. using wmp12 with lav doesn't work. the problem is with the splitter. using matroska splitter from mpchc works fine with lav video decoder and audio decor. but when I use lav splitter all I can do is seek within movie but it doesn't play. I see the images but it doesn't play. maybe nev can help. I'm using wmp x64 and lav is also 64bit

nevcairiel
1st February 2012, 22:01
@Nev
line error

http://www.mediafire.com/?5xufemxwx0fwpcn

Lav Video = fail
ffdshow = ok

The file plays fine.
What does "line error" even mean? o.o

CruNcher
1st February 2012, 22:10
scaling it up causes a strange colored line on the top (rounding errror ?)

nevcairiel
1st February 2012, 22:11
scaling it up causes a strange colored line on the top

Not for me.
I even made you a screenshot!
http://images.gammatester.com/pics/2dea80d038dca9000ad9dbfb02c0a51c.png

Playing with LAV Splitter + LAV Video + EVR-CP

CruNcher
1st February 2012, 22:16
did you tried rendering on Nvidia or SB, though i guess it could be indeed the first issue with Intels scaling, but why doesn't it happen with ffdshow ? ;) ?

the major difference would be egurs NV12 vs your NV12 i guess ;)


lav video = http://img821.imageshack.us/img821/5746/greenlinescalinglavvide.png
ffdshow = http://img62.imageshack.us/img62/3374/ffdshowscalingok.png

i tried royatam and xhmikosrs build both the same, both EVR intel scaling

nevcairiel
1st February 2012, 22:19
I suppose it doesn't like that the video has a odd number of rows, which technically is invalid for 4:2:0 content.
madVR will also produce a green line, although at the bottom.

I guess i can try pad it with one extra line of blackness, but not sure how easy that will be.
Not a high priority, odd-numbered 4:2:0 content is invalid by definition (can't have half a row of chroma), and looks fine on nvidias scaler, or if you just use RGB output.

PS:
What does egur have to do with anything? Its flash video, has nothing to do with QuickSync. :p
The only thing you can do is complain to him that his upscaler fails on this. :D

Aleksoid1978
1st February 2012, 22:50
LAV properly detects it as MJPEG, but still doesn't play properly.
Must be something the decoder can't deal with. I recommend reporting it to ffmpeg (if ffplay also fails)

All play ok - with frame rate = 1fps.

NikosD
1st February 2012, 23:01
Hi Nev,

I went through a part of my test-video collection and found some files that play OK with CUVID, but give a green screen with DXVA copy-back :

1) mpg : http://www.mediafire.com/?s9tcp2fjjjrs178
2) x264 in MP4 : http://www.mediafire.com/?d92xafh7c4dda545 from the git-repo[/B]
.
.
.
.
.
I still have some "stalling" files, ie these won't start playing (hang at the start of the video) and I end up with a black screen.
If I use another splitter, AV splitter or Sonic HD Demuxer instead of Lav-splitter, these play perfect (still using together with Lav-audio & Lav-video)

All the files are perfectly decoded by the system at signature and LAV video 0.45 at DXVA2 copy-back mode.

All but MPEG-2 of course, because UVD2.x supports only assisted (MPEG2_IDCT) acceleration.

Even WMV3 is HW accelerated via ModeVC1_VLD, the first 30 sec of the clip of course.

With the help of PotPlayer even MPEG-2 is HW accelerated.

Nowadays, I think ATI take revenge on everyone questioned the hardware/drivers.
I haven't found even a single file that is not perfectly decoded in DXVA mode.

Hardware/ drivers combination on a very high level of maturity.

Nevcariel, do you have any complaints from ATI users using DXVA2 copy-back mode reagrding artifacts and playback quality ?

For me the only complaint could be regarding performance of copy-back mode on ATI HW and nothing else.

CruNcher
1st February 2012, 23:07
UVD was from the beginning more Powerfull then VPx and had more features (though i had endless discussions with a Ateme Engineer back then about it and it had tons of bugs driver compared to Nvidia the first UVD in the 2000 series he was going crazy with how bad it handled his Broadcast streams ;)) and as said AMD is getting everything from the lab now ultra fast into the drivers see the Deshaking realtime (Motion Estmation) stuff :P
With Fusion becoming more mature everything gets more defined now hopefully well see their SDK and Ecosystem becoming also mature now :)
Though as you know i left for Intel and Intel has some great potential as well Haswell will rock so badly and Intels GFX will mature fast the improvements they made in a short time are awesome work from where they came from, though also Software has made big improvements from the OS Compiler to Applications,Engines everything got more efficient :)

nevcairiel
1st February 2012, 23:52
Nevcariel, do you have any complaints from ATI users using DXVA2 copy-back mode reagrding artifacts and playback quality ?


Dunno, most people forget to mention what GPU they use.
All the samples play fine for me as well, its just one broken driver version for NVIDIA that caused the problem. Not like that hasn't happened to ATI before (only every other driver or so. :P)

If you want to compare, on the driver maturity side, i would always give NVIDIA the win.
If ATI would just try to give back to the community, like NVIDIA and Intel do with their truly easy to use APIs, and the driver support and performance to go with it, it would go along way, but because that never happened, i'll stick to questioning their drivers. :)

Anyhow, this is OT for this thread, there are already numerous threads here and at AVSForum that discuss the pros and cons of NVIDIA, ATI and Intel. :d
Please take it there, if you want to continue.

CruNcher
1st February 2012, 23:59
Dunno, most people forget to mention what GPU they use.
All the samples play fine for me as well, its just one broken driver version for NVIDIA that caused the problem. Not like that hasn't happened to ATI before (only every other driver or so. :P)

If you want to compare, on the driver maturity side, i would always give NVIDIA the win.
If ATI would just try to give back to the community, like NVIDIA and Intel do with their truly easy to use APIs, and the driver support and performance to go with it, it would go along way, but because that never happened, i'll stick to questioning their drivers. :)

Yeah though innovation wise they where always the first i don't forget that especially as their where in the 90s (ATI) the Kings of Video, though the 3D race with Nvidia changed a lot and shifted things, ultimately ending up in Nvidia buying 3dfx ;)
And AMD/ATI still show that see the shift to NT 6 with MLAA they where the first later Nvidia overpowered them though with FXAA then they introduced Realtime Deshaking in the driver and don't forget their Adaptive Denoising way before and now they come with Realtime Deshaking part 2 (improved Realtime Motion Estimation) ;)
ANd lets be true Nvidia didn't pointed in 3D with Performance or Efficiency but more with Quality and features (CUDA) of the Rendered result (and of course their excellent Ecosystem (APIs) ideas all the conferences the Summit GTC and stuff just awesome and yeah the all time hated by ATI users "TWIMTBP") (And starting adapting Driver to Game releases (3D Engines) with sometimes questionable enhancements) ;)
Though it seems slowly AMD/ATI are their now too and still keeped their Efficiency, though Nvidia has all the ARM research and low power improvement going on (Tegra and "Project Denver") that can be ported back to the Desktop GPU core that Kepler and Maxwell will show how those where used in improving power consumption on the Desktop :)

Xaurus
2nd February 2012, 00:01
Nvidia 295.51 drivers are out for those who don't know.

http://www.geforce.com/News/articles/nvidia-geforce-295-51-beta-drivers-released

6233638
2nd February 2012, 02:04
Nvidia 295.51 drivers are out for those who don't know.

http://www.geforce.com/News/articles/nvidia-geforce-295-51-beta-drivers-releasedI seem to be getting massive framerate problems with CUVID enabled when using these. DXVA2 Copy-back seems to work fine.

So far I have only tested with VC-1 content. Initially the film played fine, but after 15–20 minutes, it started constantly dropping frames. Quit MPC-HC, started the video from where I left off, and 15–20 minutes later, it started happening again. There were some points though, where it would happen much quicker. It may just have been a coincidence that these points were spaced 15–20 minutes apart.

Redemption80
2nd February 2012, 02:40
Getting the same thing with MediaPortal and the new drivers.

Also seeing a roughly 20min thing as it happened half way through a TV show last night, and again around 20mins into a movie tonight.

Pat357
2nd February 2012, 03:21
Hmmm.
I tested the files with the Microsoft DXVA decoder, and it also shows no image (a black screen in its case)
You're talking about the WMV file I uploaded ? ("Step_into_Liquid_1080_cut.wmv")
Because with the NV drivers I have (285.62), it just plays fine with the "WMV decoder DMO".
Maybe there's still something good in the 285.62 drivers ;)

nevcairiel
2nd February 2012, 08:16
You're talking about the WMV file I uploaded ? ("Step_into_Liquid_1080_cut.wmv")
Because with the NV drivers I have (285.62), it just plays fine with the "WMV decoder DMO".
Maybe there's still something good in the 285.62 drivers ;)

No, i'm talking about MPEG2/H264 files, using the MS DTV-DVD Decoder. The WMVideo Decoder DMO does not use DXVA by default (its somewhat tricky to get it to do it)

dead_screem
2nd February 2012, 08:28
No, i'm talking about MPEG2/H264 files, using the MS DTV-DVD Decoder. The WMVideo Decoder DMO does not use DXVA by default (its somewhat tricky to get it to do it)

Really? because on XP the WMVideo Decoder DMO uses DXVA by default and I have no idea how to stop that. It seems pretty unstable and causes hangs. Whether that's a decoder issue or video drivers, who knows.

CruNcher
2nd February 2012, 08:56
Really? because on XP the WMVideo Decoder DMO uses DXVA by default and I have no idea how to stop that. It seems pretty unstable and causes hangs. Whether that's a decoder issue or video drivers, who knows.

http://www.citizeninsomniac.com/WMV/ <- use wmv power toy to disable it if you need

@Nev
Karls Matroska RV9 Demo has problems

http://www.mediafire.com/?9jfyr5sy0f9wp4p

vlc = ok
ffplay = ok
avplay = ok
lav video = stutter
ffdshow = stutter

ryrynz
2nd February 2012, 09:46
Nev, I saw your post in the QuickSync forum regarding the performance regression and also read Eric's reply, will you be looking any further into it?

kolak
2nd February 2012, 13:21
The Lav filters make Graphedit always crash because graphedit is limited to 100 media-types per filter; the Lav-filters have more then 100 different media-types.

You can only use GraphStudio or the the more recent GraphStudioNext with the lav-filters. (=based on Graphstudio, but more featutes).
Have a look at http://code.google.com/p/graph-studio-next/

Thanks for explanation- yes, I'm using graphstudio.

nevcairiel
2nd February 2012, 13:33
Nev, I saw your post in the QuickSync forum regarding the performance regression and also read Eric's reply, will you be looking any further into it?

Its still plenty fast as-is, so no, i won't. I'll let Eric look into it.
The main performance difference has already been fixed by adjusting the settings i use. I'm not sure there still is a difference, none that i could measure anyway.

rsd78
2nd February 2012, 13:49
Hi Nevcairiel,

I know this is a longshot request, but I thought I'd at least ask :). I am a long time user of Albain's Media Control plugin (I'm a WMC user). It is only compatible with Haali and Gabest splitter. I believe he open sourced the code, but I believe Albain has been MIA for some time. Is there any chance it'd be possible to make LAV Splitter compatible with Media Control? I don't have a sense of how hard it is, but I recall Albain adding Gabest support a while back in a relatively short period of time. I know you offer subtitle control via a regex of sorts, but I really like the GUI Media Control offers via the WMC player. I completely understand if its either not feasible or you are not interested however. I'd also accept an alternative GUI via WMC instead of Media Control that works with WMC instead as well :) (Kidding, but I thought I'd throw that out there).

I'd love to use Lav Splitter as it seems much more stable than Haali/Gabest when I played with it, but I really like the Media Control plugin/GUI.

Thanks Again

Hi Nev, not to bug you too much because I know you get a lot of requests/posts a day, but I just wanted to see your thoughts on the above? Thanks

nevcairiel
2nd February 2012, 13:52
Hi Nev, not to bug you too much because I know you get a lot of requests/posts a day, but I just wanted to see your thoughts on the above? Thanks

I have no interest in such a thing.
If someone wants to work on that plugin (or something comparable) and needs some information how to interface with LAV, i'll be glad to provide it, but i will not work on it myself.

SSuzaku
2nd February 2012, 14:14
hello everyone, i used LAV Filter 0.45 with MPC-HC (30-01-2012) recently with my Hybrid Vaio VPCCA16FG

this is my specs

i7 2620M
8GB RAM Crucial
AMD Radeon 6630M (Intel HD 3000)

it's wierd that i can get video with Intel [power saving mode], but i can't get video with the AMD [high performance] ( audio and subtitle were fine though )

i used the settings here http://www.elysiumsubs.com/docs/mpc-hc-guide
nothing changed.

if i used madVR i can get the video with AMD

what can i do to fix this?

thanks before

Pat357
2nd February 2012, 17:14
Nev,

Could you please add SNOW to the supported decoders ?
It's supported by ffmpeg, so it should not be too much work.

With LAV I got :
LAV Splitter Source::Video

Media Type 0:
--------------------------
Video: SNOW 848x480 23.976fps

AM_MEDIA_TYPE:
majortype: MEDIATYPE_Video {73646976-0000-0010-8000-00AA00389B71}
subtype: Unknown GUID Name {574F4E53-0000-0010-8000-00AA00389B71}
formattype: FORMAT_VideoInfo {05589F80-C356-11CE-BF01-00AA0055595A}
bFixedSizeSamples: 0
bTemporalCompression: 1
lSampleSize: 1
cbFormat: 88

With ffmpeg :
Input #0, matroska,webm, from 'K:\film\new_2012-01\13_snow.mkv':
Metadata:
ENCODER : Lavf53.24.0
Duration: 00:00:12.73, start: 0.000000, bitrate: 984 kb/s
Stream #0:0: Video: snow, yuv420p, 848x480, SAR 1:1 DAR 53:30, 23.98 fps, 23.98 tbr, 1k tbn, 1k tbc (default)
Stream #0:1(jpn): Audio: aac, 48000 Hz, stereo, s16 (default)
24.50 A-V: 12.735 fd= 22 aq= 0KB vq= 50KB sq= 0B f=1/1 f=0/0

nevcairiel
2nd February 2012, 17:25
Could you please add SNOW to the supported decoders ?


Done, also added FFV1 because someone asked for that some time ago.


@Nev
Karls Matroska RV9 Demo has problems

http://www.mediafire.com/?9jfyr5sy0f9wp4p


I tried to improve it a bit and its running better now, but its no full fix. RV30/40 in MKV is just muxed with terrible timestamps. :(

MasterNobody
2nd February 2012, 17:59
Done, also added FFV1 because someone asked for that some time ago.
Thanks. That was me who asked it.

Portioli
2nd February 2012, 22:53
does lav dxva2 works with madvr de-interlaced?

which is the optimal configuration of lav filters & madVR for 1080i50/60 files?

plew
2nd February 2012, 23:04
I can't get sound from Swedish - Girl with the Dragon Tattoo BD movie using MPC-HC and LAV Audio Decoder set to bit stream. This movie has been saved using DVDFab main movie. I have no other problems playing all other BD, folder or MKV. Going to Filters I do not see Lav Audio at all. No options for audio or subs. No other Audio decoder.

This plays with TMT5 just fine. I get subs and both audio tracks.

Thanks.

CruNcher
3rd February 2012, 00:33
Btw did someone experienced that Cyberlinks VC-1 Decoder doesn't connect with every VC-1 .mkv but denies connection with certain .mkvs and lav splitter ?
I wonder whats going on as those .mkv connect fine with Arcsofts (but Decoding is very bad timestamp issues in general with .mlv) or Microsofts Decoder it works much better with Cyberlinks or Microsofts Decoder and Cyberlinks even allows Super fast seek :)
Its not a Profile issue as some have the same Profile and one connects the other doesn't very strange behavior

@Nev
Not sure about this

http://www.mediafire.com/?y4czq3rkmcns19l

vlc = fail
mirillis = ok
mplayer2 = ok
mplayer = ok
avplay = not sure
ffplay = not sure
powerdvd = fail
lav splitter + lav video = fail
haali = fail
mpc-hc = fail
av splitter = fail
lav splitter + cyberlink = fail
lav splitter + arcsoft = ok

Arcsoft rarely likes to work with anyone (see VC-1 mkvs that Cyberlink doesn't accept and Arcsoft plays but wrong) this is one rare sample where it does work fine with the others and the others fail :P

ryrynz
3rd February 2012, 07:56
Its still plenty fast as-is, so no, i won't. I'll let Eric look into it.
The main performance difference has already been fixed by adjusting the settings i use. I'm not sure there still is a difference, none that i could measure anyway.

The reason why I ask is because Eric's benchmarks showed no difference with the file you posted and it looks like he dismissed it and has no plans to look any further into it.

Your change in MT settings did fix part of the difference I observed but I'm still seeing a slight difference in LAV Filters in favor of 0.22 if I use it to replace QuickSync 0.24.

In FFDShow 4234 which features Rev18 (0.22) of QuickSync, I get average performance with my anime intro file.
Replacing 0.24 in the latest version of LAV Filters with 0.22 gives me consistently 10% performance increase over 0.24 which doesn't stack up vs my FFDShow results.

I'll include the intro sample here and if it interests you or Eric as to why this is happening you can look into it.

These are my bench results with this file, all tests run 5 times and averaged.

FFDShow 4234 QuickSync Rev18 (0.22) 550FPS
FFDShow 4238 QuickSync Rev20 (0.23) 550FPS
FFDShow 4294 QuickSync Rev30 (0.24) 2050FPS

LAV Filters 0.45 QuickSync 0.24 2040FPS
LAV Filters 0.45 QuickSync 0.22 2300FPS

http://www.mediafire.com/?iak3fxc4pdc2rir

Thanks Nev.

nevcairiel
3rd February 2012, 07:58
Like i said before, those numbers are too high to be real FPS values. 550fps might be valid for a SD file, however 2300 is most likely a measuring error because of too short files.
Pick a file thats much longer, maybe 10 minutes or so. Sure, benchmarking might take a while longer, but the result will be much more accurate and in line with other performance results.

Also, if you're really complaining about 2050 vs 2300 fps, i'll dismiss this as well.
Find a HD file that shows a performance different. :p

ryrynz
3rd February 2012, 08:59
No worries, I'll take a look at longer files. Not complaining just thought it interesting, I'll get back to you with some results later.

ryrynz
3rd February 2012, 10:15
Find a HD file that shows a performance different. :p

You win 0.22 and 0.24 give similar results.

I tested with a 20 minute H.264 848x480 episode and compared some.

With Xhmikosr's builds I get around 315 FPS =/
With release 0.45 I get 980 FPS :D
FFDShow 4292 (Rev 30) 690 FPS =|
Also tried latest nightly and again 315 FPS.

Some strange results.

I run into an error in MPC after registering all three filters in the nightlies that the video cannot be rendered, LAV installed from an installer works fine (Xhmikosr and release builds) what's happening there? Graphedit works with it fine except slowly.

nevcairiel
3rd February 2012, 10:18
Xhmikosr builds don't include the Intel decoder, last i checked (so it might be software decoding at work there), or it might be an older version.
In general, i only give support for my own builds, i cannot control if anyone just messed up their build (which happened quite a bit in the past)

I can provide a recent git build when i get home from work.

ryrynz
3rd February 2012, 10:27
Yeah I can understand the support for your builds only. I keep the Intel decoder in the directory so perhaps the nightlies aren't linking to it or something. I'll take you up on your gracious offer of the git build, thanks :D

egur
3rd February 2012, 14:49
Version 0.25 beta is out with the following changes:
* Fixed handling of CCV1 streams (Haali splitter custom fourCC).
* Support for H264 AVI files.
* Optimized memory copy further. Removed ASM code. Code now uses intrinsic for both 32 and 64 bit as intrinsic code reached 32 bit efficiency.
* Code cosmetics.
* FFDShow rev4295

Download from SourceForge home page (http://sourceforge.net/projects/qsdecoder/)

nevcairiel
3rd February 2012, 14:51
Great thanks, i'll update the binary i ship :)

nevcairiel
3rd February 2012, 16:26
As promised, here is the latest Git build, including the latest QuickSync decoder (0.25 - r31)

http://files.1f0.de/lavf/LAVFilters-0.45-29-gbf1e69c.zip

No really major changes since 0.45, mostly bug fixes and a few new video formats supported.

omarank
3rd February 2012, 17:37
Not going to happen anytime soon, doing a low-pass filter for LFE is quite a bit more work then just mixing things.

ok.. I was wondering if there is a way to use J River Media Center's audio post-processing in mpc hc. I mean is it possible to hook LAV audio decoder to J River DSP outputting to ReClock? Sorry, this is a little OT.

nevcairiel
3rd February 2012, 17:44
ok.. I was wondering if there is a way to use J River Media Center's audio post-processing in mpc hc. I mean is it possible to hook LAV audio decoder to J River DSP outputting to ReClock? Sorry, this is a little OT.

Thats not possible, you have to use JRiver as the player to use their DSP.

goldie
3rd February 2012, 18:51
I have do some test with LAV Filter test3 build on my laptop.
...
Finally, I got a new graphic driver update for my laptop from Sony, then the out-of-order decoded frame problem of the QuickSync in LAV Filters had been fixed and the video became watchable! :p
And, ... the QuickSync performance is almost two times than DXVA2 (copy-back) mode in this benchmark, very nice! :eek:
That is so great! Now my laptop is very happy with it. :D
( Video clip used: http://xhmikosr.1f0.de/samples/2160p/DucksTakeOff/DucksTakeOff_1080p50.x264.CRF23.mkv )

None (Software) mode ~ 55 fps
DXVA2 (copy-back) mode ~ 98 fps
Intel QuickSync mode ~186 fps

None (Software) mode:
http://i.imgur.com/5oLqf.png

DXVA2 (copy-back) mode:
http://i.imgur.com/VjOXI.png

Intel QuickSync mode:
http://i.imgur.com/lFnco.png


Hardware info:
Intel HM67 Chipsset
Intel Core i7-2640M 4MB-L3cache 2.8GHz (Turbo 3.5GHz)
Intel HD 3000 (driver version 8.15.10.2418 (8.862.4.0))
Dual-channel DDR3-1333 8GB

Software info:
Windows 7 Professional 64-bit
MPC-HC 1.6.0.4014 (Official) 32-bit
LAVFilters 0.45 (Official) 32-bit

:thanks:

wanezhiling
3rd February 2012, 19:29
Btw did someone experienced that Cyberlinks VC-1 Decoder doesn't connect with every VC-1 .mkv but denies connection with certain .mkvs and lav splitter ?
Try lastest PotPlayer, it has fixed the compatibility with CyberLink 12 for almost every VC-1 container(.ts .mkv .avi .evo .wmv) I have collected, and Lav splitter benefits from this too.

http://i.imgur.com/HkCTY.png mkv
http://i.imgur.com/KoDPv.png mkv

http://i.imgur.com/Tcdkh.jpg ts
http://i.imgur.com/CvIuo.jpg evo
http://i.imgur.com/se4T8.png avi
http://i.imgur.com/hf973.jpg wmv
http://i.imgur.com/tSUZS.png wmv(wmv3)

Thunderbolt8
3rd February 2012, 21:20
tested dxva2 copy back with gpu-z and noticed that when using coreavc, the core clock jumps from 250MHz from being idle to 750MHz under load with load being 25-33%, while when using dxva2 copy back, the core clock only jumps from 250MHz to 500MHz and load is ~50%.
could this be a reason why copy back is slower than coreavc?

nevcairiel
3rd February 2012, 21:27
Those two decoders operate completely different, a comparison is pointless.

ryrynz
3rd February 2012, 23:06
@thunderbolt

You could test that theory yourself by disabling powermizer. Watch your temps go up though :-D I would think that disabling powermizer will do little for you. Different methods require different processing and powermizer should know best.

kasper93
4th February 2012, 00:39
@nevcairiel: Could you check this sample http://aleksoid.tosei.ru/Test/Sample/m2ts.VC-1i+DTS.m2ts It won't work with LAV Video Decoder, but works fine with ffdshow.

IanD
4th February 2012, 04:39
Can someone explain the best way to play DVD with MPC and LAV?

Whilst I can open a DVD, I can't navigate using the onscreen menus (only using navigate->Jump To) and video decoding only uses MPEG2 Decoder (low merit), never LAV decoder.

It might be useful to add a small guide for DVD playback to the first page.

CruNcher
4th February 2012, 04:43
Try lastest PotPlayer, it has fixed the compatibility with CyberLink 12 for almost every VC-1 container(.ts .mkv .avi .evo .wmv) I have collected, and Lav splitter benefits from this too.

http://i.imgur.com/HkCTY.png mkv
http://i.imgur.com/KoDPv.png mkv

http://i.imgur.com/Tcdkh.jpg ts
http://i.imgur.com/CvIuo.jpg evo
http://i.imgur.com/se4T8.png avi
http://i.imgur.com/hf973.jpg wmv
http://i.imgur.com/tSUZS.png wmv(wmv3)

Nice no VC-1 Adapter needed anymore *.wmv it seems hope its the same for intel :)

Now Potplayer just needs to get Subtitles working on DXVA and EVR and it would be the most efficient Player for Intel Hardware currently (leaving EVR Custom) for EVR default :)

wanezhiling
4th February 2012, 06:29
Nice no VC-1 Adapter needed anymore *.wmv it seems hope its the same for intel :
If using ASF reader, VC-1 Adapter still will occur at some time though PotP's bulit-in source is ok.
For connecting CyberLink VC-1, korean have at least workaround for 5 times as I see.
They tested almost every commercial codec which is not a secret among East Asia area. ;)
Also, mpc-hc and lav, they focus on all the time.:cool:
They are known as a "bad guy" though PotP is really a great player.

Arcsoft VC1 is friendly to be used with any splitter/source though a little low quality compared with CyberLink, with PotP, CyberLink HAM mode works all the time even if now CyberLink PDVD13,14,15 released, but for dxva and sw mode, a workaround is needed.


Now Potplayer just needs to get Subtitles working on DXVA and EVR and it would be the most efficient Player for Intel Hardware currently (leaving EVR Custom) for EVR default :)
http://i.imgur.com/5ezkY.png ass
http://i.imgur.com/ZCeA7.jpg srt
:confused:

Edit: http://forum.doom9.org/showpost.php?p=1555409&postcount=668
Ok, I try to let korean know this.
Maybe I cant describe it clearly cuz of my poor english, so you also could contact them(ahahlive@hanmail.net) if you wish.:p


PS: I have confirmed that PotP's cuda decoder must "borrowed" some from LAV, lastest 1.5.31908 has support mpeg4 asp as well.....
http://i.imgur.com/qQNLR.png
And I intentionally ask them why didn't add H/W deinterlacing like LAV CUVID,they replied: Because CUDA' deinterlace is not recommand. It use too much CPU or GPU and quality is not good. So I don't want use that function. So use PotPlayer's deinterlace function(H/W or S/W).:eek:

ryrynz
4th February 2012, 06:48
As promised, here is the latest Git build, including the latest QuickSync decoder (0.25 - r31)


Thanks, I uninstalled 0.45 and restarted, copied over the git build, ran command prompt under Admin, and installed all three filters via the batch files then restarted again.
Loaded Graphedit edited the LAV video properties for QuickSync decoding.. and I have the same issue with the other nightlies. MPC says cannot render the file and Graphedit benches at 300 odd FPS instead of 950 like with 0.45.
QuickSync DLL 0.25 is in there.. stumped. :(

nevcairiel
4th February 2012, 07:43
@nevcairiel: Could you check this sample http://aleksoid.tosei.ru/Test/Sample/m2ts.VC-1i+DTS.m2ts It won't work with LAV Video Decoder, but works fine with ffdshow.

Thats VC-1 interlaced, its not supposed to work with software decoding. Should work with the hardware decoders, though.
ffdshow uses the microsoft decoder for VC-1, which can decode these.


And I intentionally ask them why didn't add H/W deinterlacing like LAV CUVID,they replied: Because CUDA' deinterlace is not recommand. It use too much CPU or GPU and quality is not good. So I don't want use that function. So use PotPlayer's deinterlace function(H/W or S/W).:eek:

Thats nonsense, it doesn't use more GPU then any other H/W deinterlacer, and it also looks exactly the same as doing H/W deinterlacing with EVR for example.... :) (Unless they did it wrong. :p)
The only thing it uses "more" is of course the bit of speed it needs to copy the frame from the GPU to the system.

nevcairiel
4th February 2012, 07:56
Thanks, I uninstalled 0.45 and restarted, copied over the git build, ran command prompt under Admin, and installed all three filters via the batch files then restarted again.
Loaded Graphedit edited the LAV video properties for QuickSync decoding.. and I have the same issue with the other nightlies. MPC says cannot render the file and Graphedit benches at 300 odd FPS instead of 950 like with 0.45.
QuickSync DLL 0.25 is in there.. stumped. :(

Sounds like your system is majorly screwed up for some reason.
Stuff really hasn't changed.

In any case, since tehre are no changes, just installing 0.45 with the installer and then replacing all the files should be fine (without doing any re-registering)

Aspeh
4th February 2012, 08:46
Can someone explain the best way to play DVD with MPC and LAV?

Whilst I can open a DVD, I can't navigate using the onscreen menus (only using navigate->Jump To) and video decoding only uses MPEG2 Decoder (low merit), never LAV decoder.

It might be useful to add a small guide for DVD playback to the first page.

Shows that you are not a regular reader of this thread :D as sadly there is no community spirit here as anybody that asks for help or for the correct settings guides will either get ignored or get hostility and attacked like a new users a few pages back got just recently.

bexy i think it was.

golagoda
4th February 2012, 09:09
Can someone explain the best way to play DVD with MPC and LAV?

Whilst I can open a DVD, I can't navigate using the onscreen menus (only using navigate->Jump To) and video decoding only uses MPEG2 Decoder (low merit), never LAV decoder.

It might be useful to add a small guide for DVD playback to the first page.
https://code.google.com/p/lavfilters/issues/detail?id=47

nevcairiel
4th February 2012, 09:29
Can someone explain the best way to play DVD with MPC and LAV?

Whilst I can open a DVD, I can't navigate using the onscreen menus (only using navigate->Jump To) and video decoding only uses MPEG2 Decoder (low merit), never LAV decoder.

It might be useful to add a small guide for DVD playback to the first page.

DVD playback really isn't all that different from normal file playback - with one problem: LAV Video doesn't support dvd playback (yet).
You'll have to stick with another video decoder for the time being, and guessing why they don't work is really beyond me. :)

Shows that you are not a regular reader of this thread

Apparently, neither are you.
That one case really was the exception, and not because he was a new user, but because of his attitude and behavior. I for one can't stand the over-entitled users that behave as if they are paying me and i should bow down before them, oh well.

Also, your comment really wasn't all that helpful now, was it.

Aspeh
4th February 2012, 09:42
Apparently, neither are you.
That one case really was the exception, and not because he was a new user, but because of his attitude and behavior. I for one can't stand the over-entitled users that behave as if they are paying me and i should bow down before them, oh well.

Also, your comment really wasn't all that helpful now, was it.

A prime example of a hostile abrasive attitude that really wasn't all that helpful now, was it.

wanezhiling
4th February 2012, 10:17
Thats nonsense, it doesn't use more GPU then any other H/W deinterlacer, and it also looks exactly the same as doing H/W deinterlacing with EVR for example.... :) (Unless they did it wrong. :p)

But on my GT240 card, using LAV CUVID's H/W deinterlacing is always terrible, frame drops all the time... 60 --> 34,33..
The only way to avoid is just using an old driver below 285.62(I now use 285.38b) and untick "HQ Processing"(this's the key point)

On my other GTS450 card, everything is fine with lastest 295.51b and CUVID's H/W deinterlacing.

I know you'll say that GT240 is old and low performance, but at least it's a VP4 card and stronger than GT430 in terms of 3D performance, why GT430 is ok but GT240 is poor with CUVID.:(:(
In fact, CoreAVC cuda(hardware deinterlaing) and any other DXVA deinterlacer(PotP mpc-hc cyberlink arcsoft ffdshow...) works all fine on GT240, they all could keep 60 and smooth playback even I play a such heavy interlaced clip.:(

Nev, if you have a gt240 card, you'll know my pain.:(

nevcairiel
4th February 2012, 10:20
The GT430 actually has more 3D performance then the GT240 (same number of CUDA cores, but 430 is based on Fermi and higher clocked)
Anyway, thats what options are for. You can turn it off and let the renderer do deinterlacing if that works better for you.

PS:
EVR for example has automatic checks, and if the deinterlacing is running too slow, it'll turn down the quality. LAV will always use the highest quality.

wanezhiling
4th February 2012, 10:35
PS:
EVR for example has automatic checks, and if the deinterlacing is running too slow, it'll turn down the quality. LAV will always use the highest quality.
I know this, I like CUVID and its H/W deinterlaing, but its not available on GT240, sigh.:(

Anyway, I have a GTS450 card, I hope it never has a bottleneck with LAV CUIVD.:)

aufkrawall
4th February 2012, 11:08
There seems to be a memory leak in the splitters.
When I open a 23 GB MKV lossless x264 RGB file, MPC HC's memory usage exceeds 2 GB (independent from used a/v decoders).
With MPC's internal splitter it's just 650 MB.

dead_screem
4th February 2012, 11:14
The highest quality deinterlacer is not available on Windows XP (which i guess from your previous comment you're using).
Enabling it caused severe image degredation for alot of people, so i opted to just completely disable it for an overall better experience.If you were referring to the HQ Processing option I never used that. However it looks like the problem is with using 25p/30p mode, I switched to 50p/60p and the issue went away. Although I coulda swore it used to work fine with 25p/30p mode. I'll have to test some older versions later.

Also DXVA2 still shows as available on XP, it doesnt bother me that it does, but I was curious why it even does so in the first place? what is it that it checks for?

Also any news on the audio mixer? have you made any specific decisions on how/when youll do the matrix (up/downmix)?
And just curious how does ac3filter do it? Does it have a one-for-all matrix?

nevcairiel
4th February 2012, 11:38
There seems to be a memory leak in the splitters.
When I open a 23 GB MKV lossless x264 RGB file, MPC HC's memory usage exceeds 2 GB (independent from used a/v decoders).
With MPC's internal splitter it's just 650 MB.

Using a lot of memory does not equal a memory leak. It could just be using a lot of memory!

I'll try converting a movie to RGB lossless and see what happens, maybe i need to limit memory usage as well as queue size based on frames.
Wonder how long it'll hog all my CPU for this :d

aufkrawall
4th February 2012, 11:54
Using a lot of memory does not equal a memory leak. It could just be using a lot of memory!

Since I'm not an expert I often use the verb "seems". ;)
It just looked a little high to me, playback is smooth though.


I'll try converting a movie to RGB lossless and see what happens, maybe i need to limit memory usage as well as queue size based on frames.
Wonder how long it'll hog all my CPU for this :d
I grabbed the raw PNGs of Big Buck Bunny and converted it with ultrafast preset, didn't take long. :)
Please don't use placebo, or I won't witness the fix. :D

DragonQ
4th February 2012, 11:57
The GT430 actually has more 3D performance then the GT240 (same number of CUDA cores, but 430 is based on Fermi and higher clocked)
Anyway, thats what options are for. You can turn it off and let the renderer do deinterlacing if that works better for you.
Yeah, in my pre-HTPC-purchase research it basically came down to the GT430 vs the GT520 (the GT240 was discounted early). The latter is able to handle higher bit rates due to the newer VDPAU Feature Set D but the former has more CUDA cores so handles postprocessing better (important for MadVR). Since the GT430 only falls down with bit rates higher than 100 Mbps (24p) or 65 Mbps (60p), it's not really a problem right now, so if you want MadVR then the GT430 is the card to get :). It's also available in passively cooled editions, which are perfect for HTPCs. A GT530 would be the best of both worlds but doesn't currently exist...

EVR for example has automatic checks, and if the deinterlacing is running too slow, it'll turn down the quality. LAV will always use the highest quality.
One of the best things about LAV is the nVidia hardware deinterlacing :). It's so good on the later models that do vector adaptive deinterlacing (like the GT430/520 but not on my older GTS250 unfortunately). The only time I've seen it struggle (and by that I mean there's noticeable bob) is with content that has been upconverted from SD to HD by ITV...but then their upscaling has always been questionable and differs between region so it's blatantly a problem their end.

nevcairiel
4th February 2012, 12:07
Please don't use placebo, or I won't witness the fix. :D

just --qp 0 --profile high444 --output-csp rgb, nothing fancy
Going at around 15-20 fps, so just 2 hours for this full movie i picked. :p

VipZ
4th February 2012, 12:11
Thanks for the streaming support Nev :)

I have two questions on this, is this the intended behaviour of this,
File Source (URL) -> LAV Splitter -> Lav Audio

The above seems to work with 0.45 and latest compile, so thinking the above may be wrong

With the above there is no stream info being displayed, is this due to the required interfaces not being supported yet or the graph is wrong?

Thanks

nevcairiel
4th February 2012, 12:14
I don't know what you mean with "stream info"?

Anyhow, the way its supposed to work is that LAV is the source filter (no File Source (URL)). Also, it only works with RTSP/RTP/MMS streams, not with HTTP (yet).
Its also not hooked up to automatically be used for streaming, you can test it in graphstudio by just giving it such a url though.

VipZ
4th February 2012, 12:18
I don't know what you mean with "stream info"?

Anyhow, the way its supposed to work is that LAV is the source filter (no File Source (URL)). Also, it only works with RTSP/RTP/MMS streams, not with HTTP (yet).

Thanks, I just realised that I was trying to test something that wasn't yet supported (SHOUTcast).

And by steam info I am referring to the info passed such as track name currently playing etc.

nevcairiel
4th February 2012, 12:23
Since I'm not an expert I often use the verb "seems". ;)
It just looked a little high to me, playback is smooth though.


I limited the queues to 512MB soft limit with a 768MB hard limit now, so it shouldn't use all that much memory. I had mine using way over 2GB with my test movie now.

I also added the IBufferInfo interface, so in MPC-HC you can activate the "Statistics" in the menu, and it'll show you how the queues fare (number of frames and size in bytes)

PS:
My file looked kinda funky, i think i did something wrong in the encode. :p

Thanks, I just realised that I was trying to test something that wasn't yet supported (SHOUTcast).

And by steam info I am referring to the info passed such as track name currently playing etc.

Don't think either of those things ever will be supported.
I just use whatever methods ffmpeg gives me.

aufkrawall
4th February 2012, 12:53
I limited the queues to 512MB soft limit with a 768MB hard limit now, so it shouldn't use all that much memory. I had mine using way over 2GB with my test movie now.

I also added the IBufferInfo interface, so in MPC-HC you can activate the "Statistics" in the menu, and it'll show you how the queues fare (number of frames and size in bytes)

Cool, thanks a lot. :)


PS:
My file looked kinda funky, i think i did something wrong in the encode. :p

Oh yes, it can be struggling to use other colorspaces than I420 with x264. :scared:
For me Avisynth with ffms and CSP/range conversion works great as a source for x264.
ffmpeg included in x264 can't even decode x264 RGB video properly, I had to use LAV via DirectShowSource for this. :cool:

VipZ
4th February 2012, 12:54
Don't think either of those things ever will be supported.
I just use whatever methods ffmpeg gives me.

No problem, thanks for the info.

ryrynz
4th February 2012, 13:18
Sounds like your system is majorly screwed up for some reason.


:D Did some investigation, it's the splitter. If I take 0.45 and just unregister the splitter then reinstall it still doesn't work.

I wonder what the installer does to register the splitter that the batch file does not do?

I looked through GraphStudio and the splitter is there when installed via the batch file, it's values are all the same as when it's installed via the installer..
oh well guess I'm up for a fresh install sometime, I wish Windows really was a lot more robust I think I've manage to corrupt every version I've run except Windows 98. :p Strangely though this is my HTPC and is kept quite clean.

nevcairiel
4th February 2012, 13:23
I wonder what the installer does to register the splitter that the batch file does not do?

It hooks it up to the file extensions, while the batch doesn't do that.

For the record, in the (near) future, the batch files won't do anything but register the .ax in the system, they will no longer mess around with the registry - i'll leave that to the installer.
That means using the batch file will in most cases no longer be enough to get it to play stuff properly (at least for the splitter) - on the other hand, if you used the installer once, you are free to re-register the .ax file without it stopping to work. ;)

I'll probably also separate x86 & x64 installers to make it easier.

Pepp
4th February 2012, 13:23
If it wasn't for Bexey i would have never of known you was supposed to change any settings in mpc-hc and would still be running it on the default install settings.

It would be nice if a *.ini / mpc-hc.ini file option was added to the LAV installer to set mpc-hc to the basic needed settings.

madshi
4th February 2012, 13:47
[...]
My ignore list is growing...

roytam1
4th February 2012, 13:47
I limited the queues to 512MB soft limit with a 768MB hard limit now, so it shouldn't use all that much memory. I had mine using way over 2GB with my test movie now.

I also added the IBufferInfo interface, so in MPC-HC you can activate the "Statistics" in the menu, and it'll show you how the queues fare (number of frames and size in bytes)

PS:
My file looked kinda funky, i think i did something wrong in the encode. :p



Don't think either of those things ever will be supported.
I just use whatever methods ffmpeg gives me.

Don't know if the patch still applies but it is here:
http://wiki.multimedia.cx/index.php?title=Talk:FFmpeg_Summer_Of_Code_2009#Shoutcast_support
http://web.archiveorange.com/archive/v/yR2T400567mWEyyZHg3k

IanD
4th February 2012, 14:41
DVD playback really isn't all that different from normal file playback - with one problem: LAV Video doesn't support dvd playback (yet).
You'll have to stick with another video decoder for the time being, and guessing why they don't work is really beyond me. :)

I think it would be really useful to have a small section on the first page that is kept updated with what LAV will handle and what is best handled by another method, to avoid repeated questions by those who aren't in a position to read 500 pages. I don't recall any mention of DVD playback not being supported.

DVD playback is the most important thing for me right now: I have a great Bluray player, but want the ability to get the best out of my DVD collection without top of the line CPU.

I tried using ffdshow as decoder, but playback freezes on the first frame.

There are so many players, splitters, decoders, renderers and tweaks that it is confusing finding something that works for DVD and maximises the quality and adjustments necessary to cope with varying quality.

Andy o
4th February 2012, 17:27
You don't need a "top of the line" CPU to software-decode SD MPEG-2 at all!

dann23
4th February 2012, 18:33
My ignore list is growing...

perhaps you can add another one :D

mindbomb
4th February 2012, 19:49
on the subject of dvd playback, this is what I do, and it works for some reason:
lav video and lav splitter don't connect
use microsoft dtv-dvd video decoder and dvd navigator instead
add ffdshow audio processor (you can even disable mixer,add 32 float output for it, so it isn't actually doing anything)

voila, the open dvd function of mpc hc now works
for some reason, having it go through another audio filter prevents macrovision errors.

renethx at avsforums detailed all the possibilities for dvd playback, including this method
http://www.avsforum.com/avs-vb/showpost.php?p=21568337&postcount=654

kalston
4th February 2012, 20:39
I think it would be really useful to have a small section on the first page that is kept updated with what LAV will handle and what is best handled by another method, to avoid repeated questions by those who aren't in a position to read 500 pages. I don't recall any mention of DVD playback not being supported.

DVD playback is the most important thing for me right now: I have a great Bluray player, but want the ability to get the best out of my DVD collection without top of the line CPU.

I tried using ffdshow as decoder, but playback freezes on the first frame.

There are so many players, splitters, decoders, renderers and tweaks that it is confusing finding something that works for DVD and maximises the quality and adjustments necessary to cope with varying quality.

This sounds like a bug because DVD decoding really does not need a fast CPU. Even when rescaling with madVR and whatever it should be pretty light.

Otherwise, how could I have watched smooth DVDs back in the days on my 600mhz laptop?

clsid
4th February 2012, 22:02
on the subject of dvd playback, this is what I do, and it works for some reason:
lav video and lav splitter don't connect
use microsoft dtv-dvd video decoder and dvd navigator instead
add ffdshow audio processor (you can even disable mixer,add 32 float output for it, so it isn't actually doing anything)

voila, the open dvd function of mpc hc now works
for some reason, having it go through another audio filter prevents macrovision errors.

renethx at avsforums detailed all the possibilities for dvd playback, including this method
http://www.avsforum.com/avs-vb/showpost.php?p=21568337&postcount=654
A bit of topic, but it doesn't work here with madVR. I get the Copy-Protect Fail error. Also when using ffdshow as video decoder. With and without ffdshow audio processor in the graph.

SamuriHL
4th February 2012, 22:05
A bit of topic, but it doesn't work here with madVR. I get the Copy-Protect Fail error. Also when using ffdshow as video decoder. With and without ffdshow audio processor in the graph.

Some people can get it to work. Others can't. Kinda frustrating.

Reino
5th February 2012, 00:16
...That means using the batch file will in most cases no longer be enough to get it to play stuff properly (at least for the splitter)...Could you be a bit more specific?

mindbomb
5th February 2012, 00:33
A bit of topic, but it doesn't work here with madVR. I get the Copy-Protect Fail error. Also when using ffdshow as video decoder. With and without ffdshow audio processor in the graph.

it's pretty on topic imo, how to deal with dvds when you are primarily using madvr and lav filters, which many are.

i think you have to decrypt the dvds first? idk

nevcairiel
5th February 2012, 00:49
Could you be a bit more specific?

I cannot be specific if the specifics haven't been specified yet :p
The main deal is that the .ax files will no longer mess with the registry in "unsafe" ways. This makes no difference at all for the decoders, but the splitter will need extra help to get going, which the installer will provide instead.

Pepp
5th February 2012, 01:03
My ignore list is growing...

Does the board need to know your private settings. ?

You posted that for no other reason than to be antagonistic and instigate more negative posts which is on the same steps as passive agressive bullying.

clsid
5th February 2012, 01:16
it's pretty on topic imo, how to deal with dvds when you are primarily using madvr and lav filters, which many are.

i think you have to decrypt the dvds first? idkIt is very much possible that ripping to harddrive makes a difference in avoiding the Copy-Protect error. I tested with an actual DVD disc. Depending on the filterset I get either Macrovision error or Copy-Protect error.

suanm
5th February 2012, 02:10
the new DXVA2 "copy back" function works very perfectly on my PC,thank you ,nevcairiel
,i like very much DXVA2 "copy-back" decoding mode

CruNcher
5th February 2012, 02:51
Just wait for Native ;)

@Nev
Runtime wrong Detected 2:20 min (no seek possible)

http://www.mediafire.com/?7611h5nk120y9cm

VLC = 1:37 ??? (seeking ok)
Mpc-HC Splitter = ok
Mainconcept Splitter = ok
Av splitter = fail (0:01)
Lav Splitter = fail (0:01)
Microsoft Splitter = fail (0:01)

ryrynz
5th February 2012, 04:50
Does the board need to know your private settings. ?

You posted that for no other reason than to be antagonistic and instigate more negative posts which is on the same steps as passive aggressive bullying.

Pepp, I think you misunderstand. This could perhaps be related to your short period of time here on this or perhaps any forum, so I'd like to make a quick post about socializing here.

There is a degree of respect that needs to be held for developers who share their work with others for nothing other than appreciation and interest. This is important Pepp, it makes people feel valued when you treat them with respect, they in turn will reciprocate this back to you, everyone wins.

What Madshi said was neither antagonistic nor negative IMO.

In fact I never in my time on the forum seen him as being either of those two things! Even when directed with somewhat harsh criticisms in what I have considered to be in a rude manner, this always spoke volumes to me as a good judge of his character.

What was said merely serves as a warning to others and was just basic socializing, I actually found the post quite humorous :p

Remember, these projects are created for their enjoyment, thankfully we can share in it by discussing ideas and testing new features. Please place yourself in their shoes before commenting, thanks.

Pepp
5th February 2012, 05:21
Pepp, I think you misunderstand. This could perhaps be related to your short period of time here on this or perhaps any forum, so I'd like to make a quick post about socializing here.

There is a degree of respect that needs to be held for developers who share their work with others for nothing other than appreciation and interest. This is important Pepp, it makes people feel valued when you treat them with respect, they in turn will reciprocate this back to you, everyone wins.

What Madshi said was neither antagonistic nor negative IMO.

In fact I never in my time on the forum seen him as being either of those two things! Even when directed with somewhat harsh criticisms in what I have considered to be in a rude manner, this always spoke volumes to me as a good judge of his character.

What was said merely serves as a warning to others and was just basic socializing, I actually found the post quite humorous :p

Remember, these projects are created for their enjoyment, thankfully we can share in it by discussing ideas and testing new features. Please place yourself in their shoes before commenting, thanks.

To be blunt your post amounts to nothing more than brown nosing and and attempt to make obnoxious behaviour acceptable.Just because someone codes something it doesn't give them antisocial privileges over others infact they should be humble and gracious that users are intrested.

The fact that you find the instigation of negative posts towards users humorous is in fact a sad reflection on your own social defects.

IanD
5th February 2012, 05:49
renethx at avsforums detailed all the possibilities for dvd playback, including this method
http://www.avsforum.com/avs-vb/showpost.php?p=21568337&postcount=654
Thanks for that.

I used to be able to play DVDs with MPC, but it didn't support post-processing; I could play DVDs with ffdshow, but it didn't properly support menus; with the appreciated release of LAV, I was hoping to get better playback for all codecs including DVD with an opportunity to also do processing with GPU and CPU.

One must understand that DVD playback is far more than decoding: to get the best results when upscaling to an HDTV, one must de-interlace well and post-process the video for noise and other artifacts.

There is a thread on AVSforum I believe about getting the best results for DVD upscaling, using ffdshow, but it requires a top of the line CPU due to the supersampling and avisynth processing routines included. Eventually I would like to achieve similar quality effects, but using both GPU and CPU to reduce the hardware requirements.

My Oppo 93 met all my requirements for DVD and Bluray, because it uses a Qdeo VP, but now that file playback is being eroded by the studios, I need a replacement and currently an HTPC is the likeliest candidate as none of the media players do Video Processing.

CruNcher
5th February 2012, 08:47
@Nev

Snow playback issue

http://www.mediafire.com/?c4j4l6xoooae6c2

ffplay = ok
lav splitter + lav video = fails connection

Andy o
5th February 2012, 09:30
Bleh, some moderator should take a look at the IPs of those "Jan 2012" concern trolls who suspiciously appeared after the first guy went MIA knowing that others have put that name under their ignore lists.

roytam1
5th February 2012, 09:57
@Nev

Snow playback issue

http://www.mediafire.com/?c4j4l6xoooae6c2

ffplay = ok
lav splitter + lav video = fails connection

Did you unregister the filters and reregister them again?

nevcairiel
5th February 2012, 10:01
@Nev

Snow playback issue

http://www.mediafire.com/?c4j4l6xoooae6c2

ffplay = ok
lav splitter + lav video = fails connection

Works just fine.

madshi
5th February 2012, 10:55
Bleh, some moderator should take a look at the IPs of those "Jan 2012" concern trolls who suspiciously appeared after the first guy went MIA knowing that others have put that name under their ignore lists.
Agreed. I've been about 6 years on this forum. And my ignore list contained exactly 1 user for 5.5 years. Now in 2012 I've added three new users to my ignore list. We're being overrun by trolls atm. Or maybe it's just one and the same person all over again.

RBG
5th February 2012, 11:23
As promised, here is the latest Git build, including the latest QuickSync decoder (0.25 - r31)

http://files.1f0.de/lavf/LAVFilters-0.45-29-gbf1e69c.zip

No really major changes since 0.45, mostly bug fixes and a few new video formats supported.

Nevcairiel, please, can you upload updated x64 version too.

Some testing, system Windows 7 x64, graph studio process started with high priority:

LAV video x86 0.45 QS 0.24
http://10pix.ru/img1/1308/6751364.png

LAV video x86 0.45 QS 0.25
http://10pix.ru/img1/4786/6751362.png

FFdshow x64 QS 0.25
http://10pix.ru/img1/2080/6751361.png

CruNcher
5th February 2012, 11:53
Did you unregister the filters and reregister them again?

nope indeed that fixes it thx :)

Works just fine.

yup it does, my fault :)

nevcairiel
5th February 2012, 12:44
Here is another test version.

New in this version is support for "Random Dithering", that instead of a very static pattern used by Ordered Dithering now uses random data to dither. This has the advantage that it avoids creating any visible "pattern" in the video, especially if you use a cheap LCD display with a 6-bit panel that applys dithering itself. Its the new default mode.

x86: http://files.1f0.de/lavf/LAVFilters-0.45-59-g0fba6a6.zip
x64: http://files.1f0.de/lavf/LAVFilters-0.45-59-g0fba6a6-x64.zip

fastplayer
5th February 2012, 13:09
I'm glad you're still improving the RGB converter! :)
There are still too many systems with decent CPUs crippled by crappy GPUs...

nevcairiel
5th February 2012, 13:10
This also affects playback of 10-bit material, not only RGB output. ;)
Granted, even then it only affects playback with EVR/VMR, where one might potentially prefer RGB output ...

fastplayer
5th February 2012, 13:14
I would've known that if it was mentioned in the tooltip... :D

no1d
5th February 2012, 14:19
Hello.
MPC-HC crashes when opening Lagarith RGB video (avi). It's ok with internal splitter and lagarith's own decoder. All combinations with Lav filters dont work.
Tried 0.45 and latest git version.
What am I doing wrong?

RBG
5th February 2012, 14:26
Here is another test version.


Thanks. I wonder why x86 is always a little bit ahead than x64 even on a native x64 system?

LAV video x86
http://img827.imageshack.us/img827/373/lavvideo86.png

LAV video x64
http://img717.imageshack.us/img717/8306/lavvideox64.png

roytam1
5th February 2012, 14:27
Hello.
MPC-HC crashes when opening Lagarith RGB video (avi). It's ok with internal splitter and lagarith's own decoder. All combinations with Lav filters dont work.
Tried 0.45 and latest git version.
What am I doing wrong?

Please provide sample.
And try unregister and reregister the filters if you updated them recently.

egur
5th February 2012, 15:06
Thanks. I wonder why x86 is always a little bit ahead than x64 even on a native x64 system?


I actually got slightly better results in x64 while benchmarking the QS decoder in ffdshow. Anyway they were very close (<1%).
At these frame rates (~3ms per frame) the splitter performance is also matters.

RBG
5th February 2012, 16:13
I actually got slightly better results in x64 while benchmarking the QS decoder in ffdshow. Anyway they were very close (<1%).
At these frame rates (~3ms per frame) the splitter performance is also matters.

I used the same splitter(LAV) and the same file, and tried to maintain accuracy as much as it possible, that means no software running in the background, and high priority to graph studio process.

Here are the results:

FFdshow x86 4296 ICL12
http://img850.imageshack.us/img850/5765/ffdshowx86.png

FFdshow x64 4296 MSVC 2010
http://img59.imageshack.us/img59/1323/ffdshowx64.png

LAV x86
http://img12.imageshack.us/img12/460/lavx86.png

LAV x64
http://img256.imageshack.us/img256/7575/lavx64.png

Indeed ffdshow x64 is slightly faster on native x64 system, but in the case of LAV the situation is opposite, x86 version is faster and that result is repeatable on different samples.

nevcairiel
5th February 2012, 16:22
The speed difference is well below 1%, not statistically significant. :)

RBG
5th February 2012, 17:08
nevcairiel

Btw, do you intend to add WMV9 simple and main profile QS acceleration support to LAV. It would be great. :)

nevcairiel
5th February 2012, 17:20
Btw, do you intend to add WMV9 simple and main profile QS acceleration support to LAV. It would be great. :)

It would seem its not supported by the hardware and/or Intels driver, it only accepts advanced profile.

RBG
5th February 2012, 17:46
It would seem its not supported by the hardware and/or Intels driver, it only accepts advanced profile.

You're right, I tried it and even ffdshow shows QS decoder usage, the decoder itself falls back to super slow software mode. And what about CUVID or dxva copy back acceleration for nvidia and ati hardware?

nevcairiel
5th February 2012, 17:53
DXVA already supports WMV9.
CUVID doesn't like it, but i never really tried that hard because WMV9 is a fairly useless format. :p

RBG
5th February 2012, 18:25
DXVA already supports WMV9.


Pure DXVA supports it, DXVA(copy-back) shows only green screen.

nevcairiel
5th February 2012, 18:27
Pure DXVA supports it, DXVA(copy-back) shows only green screen.

As you might've guessed, works for me.
Make sure you don't use the 285 WHQL driver on NVIDIA, its bugged.

RBG
5th February 2012, 18:30
WMV9 is a fairly useless format. :p

From my point of view it's better to have acceleration support than not to have it.;) I am dreaming of being able to combine hardware decode of most common formats with madVR. :)

Make sure you don't use the 285 WHQL driver on NVIDIA, its bugged.

Yes I am on 285th driver.:p What version can you recommend as most stable for video playback?

nevcairiel
5th February 2012, 18:34
From my point of view it's better to have acceleration support than not to have it.;) I am dreaming of being able to combine hardware decode of most common formats with madVR. :)

H264, VC1 and MPEG2 are the most common formats in modern media, covering everything from broadcasts, DVDs, Blu-rays, and a fair part of streaming too. ;)


Yes I am on 285th driver.:p What version can you recommend as most stable for video playback?

The 290 beta worked perfectly for me ever since it came out.
Apparently the 295 beta that was just released has some new issues, though, so stick with 290

Shark007
5th February 2012, 18:41
The 290 beta worked perfectly for me ever since it came out.

290.36 to be more precise. 290.53 also has issues.

RBG
5th February 2012, 18:46
H264, VC1 and MPEG2 are the most common formats in modern media, covering everything from broadcasts, DVDs, Blu-rays, and a fair part of streaming too. ;)


That is true if you're speaking about big media, but among amateurs I pretty often encounter WMV9 simple and main, probably made by something like windows movie maker.:)


The 290 beta worked perfectly for me ever since it came out.


290.36 to be more precise. 290.53 also has issues.


Thank you.

SamuriHL
5th February 2012, 19:04
I think I'm gonna go back to 290.36, as well. Re-setup my custom profiles and leave it alone until someone says there's a better driver.

no1d
5th February 2012, 19:43
Hello.
MPC-HC crashes when opening Lagarith RGB video (avi). It's ok with internal splitter and lagarith's own decoder. All combinations with Lav filters dont work.
Tried 0.45 and latest git version.
What am I doing wrong?

Sample:
http://www.mediafire.com/?y8uf88pd33sxx75

Della
5th February 2012, 20:02
I'm trying to get a hanndle on LAV filters. I have replaced Haalli with LAV & finding some videos play without sound.
Specifically, the file below will play without sound, but, if I disable AC3 under Lav Audio formats, it plays just fine.

Am I missing something?

Thanks

Player WMP
Windows 7 Ultimate
No other codecs installed.
Video Card nVidea Gt460
No Bitstreaming selected.



General

Complete name : C:\Videos\Luck.S01E01.720p.HDTV.x264-IMMERSE.mkv
Format : Matroska
Format version : Version 2
File size : 2.59 GiB
Duration : 1h 0mn
Overall bit rate : 6 177 Kbps
Writing application : x264.exe
Writing library : mkv2rls x264-tv version built on 2011. febr. 17

Video
ID : 1
Format : AVC
Format/Info : Advanced Video Codec
Format profile : High@L4.1
Format settings, CABAC : Yes
Format settings, ReFrames : 5 frames
Codec ID : V_MPEG4/ISO/AVC
Duration : 1h 0mn
Bit rate : 5 670 Kbps
Width : 1 280 pixels
Height : 720 pixels
Display aspect ratio : 16:9
Frame rate : 23.976 fps
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Progressive
Bits/(Pixel*Frame) : 0.257
Stream size : 2.38 GiB (92%)
Writing library : x264 core 120 r2119 0637cd6
Encoding settings : cabac=1 / ref=5 / deblock=1:-1:-1 / analyse=0x3:0x113 / me=umh / subme=8 / psy=1 / psy_rd=1.00:0.15 / mixed_ref=1 / me_range=16 / chroma_me=1 / trellis=1 / 8x8dct=1 / cqm=0 / deadzone=21,11 / fast_pskip=1 / chroma_qp_offset=-3 / threads=36 / sliced_threads=0 / nr=0 / decimate=1 / interlaced=0 / bluray_compat=0 / constrained_intra=0 / bframes=3 / b_pyramid=2 / b_adapt=2 / b_bias=0 / direct=3 / weightb=1 / open_gop=0 / weightp=2 / keyint=250 / keyint_min=23 / scenecut=40 / intra_refresh=0 / rc_lookahead=50 / rc=crf / mbtree=1 / crf=18.0 / qcomp=0.60 / qpmin=0 / qpmax=69 / qpstep=4 / ip_ratio=1.40 / aq=1:1.00
Language : English
Default : Yes
Forced : No

Audio
ID : 2
Format : AC-3
Format/Info : Audio Coding 3
Mode extension : CM (complete main)
Codec ID : A_AC3
Duration : 1h 0mn
Bit rate mode : Constant
Bit rate : 384 Kbps
Channel(s) : 2 channels
Channel positions : Front: L R
Sampling rate : 48.0 KHz
Bit depth : 16 bits
Compression mode : Lossy
Stream size : 165 MiB (6%)
Language : English
Default : Yes
Forced : No

Text
ID : 3
Format : UTF-8
Codec ID : S_TEXT/UTF8
Codec ID/Info : UTF-8 Plain Text
Language : English
Default : No
Forced : No

nevcairiel
5th February 2012, 20:11
Sample:
http://www.mediafire.com/?y8uf88pd33sxx75

Looks like i can reproduce the crash, i'll see what i can do.

Edit:
And should be fixed

dead_screem
5th February 2012, 22:00
290.36 to be more precise. 290.53 also has issues.

what issues are you guys referring to? because I tried 295 on my 9800GTX and VMR9 becomes unstable after awhile and crashes. I reverted to 257 release drivers (which is also the last version that CUVID works right with interlaced VC-1 (havn't tested MPEG-2 or H.264)
Was this one of the known issues with 295 and 290.53?

Shark007
5th February 2012, 22:06
Was this one of the known issues with 295 and 290.53?

I'm sorry to inform you that i did not document exactly what driver had what problem.
I just kept track of what driver version worked properly, and which didnt.

Della
5th February 2012, 22:07
Well, after not being able to play the file on two Win 7 64 machines I found the solution was to uninstall LAV & re-install Haali Media Splitter. Playback works normally.

I couldn't get the video portion to play (it would remain on a fixed position as if paused) at all with LAV, but, interestingly works fine with Haali.

Della



[QUOTE=Della;1556128]I'm trying to get a hanndle on LAV filters. I have replaced Haalli with LAV & finding some videos play without sound.
Specifically, the file below will play without sound, but, if I disable AC3 under Lav Audio formats, it plays just fine.

Am I missing something?

Thanks

Player WMP
Windows 7 Ultimate
No other codecs installed.
Video Card nVidea Gt460
No Bitstreaming selected.


Audio
ID : 2
Format : AC-3
Format/Info : Audio Coding 3
Mode extension : CM (complete main)
Codec ID : A_AC3
Duration : 1h 0mn
Bit rate mode : Constant
Bit rate : 384 Kbps
Channel(s) : 2 channels
Channel positions : Front: L R
Sampling rate : 48.0 KHz
Bit depth : 16 bits
Compression mode : Lossy
Stream size : 165 MiB (6%)
Language : English
Default : Yes
Forced : No

nevcairiel
5th February 2012, 22:12
I couldn't get the video portion to play (it would remain on a fixed position as if paused) at all with LAV, but, interestingly works fine with Haali.


Wasn't your problem with having no audio just two hours ago...?
You should decide what your problem is. :p

There is a known issue with WMP being "stuck" like that when there is subtitles involved. I'm working on it.

Della
5th February 2012, 22:26
Wasn't your problem with having no audio just two hours ago...?
You should decide what your problem is. :p

Also, i cannot reproduce a single issue here. :) I'm looking at WMP playing a MKV with all 3 LAV components just now.
But then, if you use WMP, i leave to your own fate, unless someone actually gives me a problem i can reproduce.

Actually I should have been more clear. The audio not working on the first machine led me to trying it on a second computer, used solely as a HTPC, with a previous version of LAV installed. On this second machine, I ran into both the audio and video issues.

I believe the problem with the audio is limited to 2 channel AAC as everything else seemed OK. I'd be happy to provide the sample, I can assure you there is a problem in the lav splitter or audio.

Thanks for taking the time to respond.

I see you edited the post, yes there are sub titles involved.

Thank you.

Della

nevcairiel
5th February 2012, 22:35
The video stalling issue should be fixed for the next version.
If a specific file doesn't work, a sample would be good.

CruNcher
6th February 2012, 00:07
Writing library : mkv2rls x264-tv version built on 2011. febr. 17

isn't there something known about wrong sampling frequencies ;)

Della try to remux it with mkvmerge, does it fix the issue ?

no1d
6th February 2012, 00:54
Looks like i can reproduce the crash, i'll see what i can do.

Edit:
And should be fixed

Thanks.
Lav splitter works well with Lagarith decoder.
LAV Video seems to be very slow (~12fps on that sample)
Lav video + internal splitter = black video.

ryrynz
6th February 2012, 03:53
Nev, my HTPC was not splitting M4V videos (as well as not setting the splitter with git builds) I had to set MPC's internal MP4/MOV splitter or install Haali's splitter to get it to work.

I've fixed both of those issues by using a neat program called CodecTweakTool. Even after uninstalling Haali I still had references to it and there were numerous Cyberlink references also which were cleaned out.

I can play with nightlies now! I'm a happy chappy.

Della
6th February 2012, 04:40
Writing library : mkv2rls x264-tv version built on 2011. febr. 17

isn't there something known about wrong sampling frequencies ;)

Della try to remux it with mkvmerge, does it fix the issue ?

Will do, and will report back.
Thank you.

Della

boyumeow
6th February 2012, 04:49
what issues are you guys referring to? because I tried 295 on my 9800GTX and VMR9 becomes unstable after awhile and crashes. I reverted to 257 release drivers (which is also the last version that CUVID works right with interlaced VC-1 (havn't tested MPEG-2 or H.264)
Was this one of the known issues with 295 and 290.53?

I will assume 295.51 is having memory leak if I'm not wrong (eats up my ram). 290.53 works better in my machine than 290.36 (screen refresh after screen saving mode). Thanks.

nevcairiel
6th February 2012, 08:25
LAV Video seems to be very slow (~12fps on that sample)


I never claimed it was fast, but its Intra-only, so multi-threading it would be quite easy (and it should scale nearly linear)

Aspeh
6th February 2012, 09:19
Nev, my HTPC was not splitting M4V videos (as well as not setting the splitter with git builds) I had to set MPC's internal MP4/MOV splitter or install Haali's splitter to get it to work.

I've fixed both of those issues by using a neat program called CodecTweakTool. Even after uninstalling Haali I still had references to it and there were numerous Cyberlink references also which were cleaned out.

I can play with nightlies now! I'm a happy chappy.

Do you know if this tweak tool has been fully checked to be clean of adware and such and do they have an official webpage for it. ?

Thanks

dlmh
6th February 2012, 09:44
Hey nev,

Is there any PQ benefits from using the DXVA2 (copy back) instead of just the software decoder on an AMD GPU (HD 6870)? I'm assuming the onboard post-processors will kick in as soon it's hardware accelerated? It's still a bummer we can't user MadVR on XBMC ;)

nevcairiel
6th February 2012, 09:47
Is there any PQ benefits from using the DXVA2 (copy back) instead of just the software decoder on an AMD GPU (HD 6870)?

In general, quality is always the same no matter which decoder you use. The post-processor should not be used in this case.

In the end, you're the one with the hardware at hand, so just set the post-processing to maximum and check if you see a difference.

ryrynz
6th February 2012, 10:18
Do you know if this tweak tool has been fully checked to be clean of adware and such and do they have an official webpage for it. ?

Thanks

http://www.free-codecs.com/download/codec_tweak_tool.htm

No adware, very clean looking program from a reputable site.

Aspeh
6th February 2012, 11:11
Wonder if LAV filters will work with the new raspberry computer. :p

nevcairiel
6th February 2012, 11:16
Wonder if LAV filters will work with the new raspberry computer. :p

They are not x86 based, so obviously not.
ARM support might be of interest some day if Win8 is out and in widespread use on ARM (and i have such a device), but then again, i doubt those things can run Win8. ;)

aufkrawall
6th February 2012, 15:35
I found some strange behavior/maybe bug:
Somehow brightness of a x264 I444 10bit video doesn't look fully right without converting to RGB when decoding while playing.

This is how the original FRAPS RGB looks like:
http://www.ld-host.de/uploads/thumbnails/e494e8bd47438960d33d94b080e5ad46.png (http://www.ld-host.de/show/e494e8bd47438960d33d94b080e5ad46.png)

x264 I444 10bit with default LAV decoding settings (CUVID) (a very little too dark):
http://www.ld-host.de/uploads/thumbnails/cea7ebe79549ed195ece54bc87a69e2f.png (http://www.ld-host.de/show/cea7ebe79549ed195ece54bc87a69e2f.png)

The same video with forced RGB32 output (to me it looks exactly like the original FRAPS RGB video):
http://www.ld-host.de/uploads/thumbnails/a7ae5ad2f56418b1b46bd017bee44509.png (http://www.ld-host.de/show/a7ae5ad2f56418b1b46bd017bee44509.png)

Download a test sample (it's not crf 0 like the pictures but this shouldn't make any difference):
http://www.mediafire.com/?n5ymnpdasada4gg
Video has TV range (Rec709).

It would be nice if it looked right without having to force RGB32 output. :)
I don't know what's about 8bit or I420, didn't test it.
Tested with latest MPC HC, LAV nightly and EVR custom.

Btw: Memory consumption with huge x264 RGB file and LAV splitter is now reduced by a half (~1.2 GB), but that's still twice as much than with MPC's internal splitter.

nevcairiel
6th February 2012, 15:50
Somehow brightness of a x264 I444 10bit video doesn't look fully right without converting to RGB when decoding while playing.

I take the YUV 4:4:4 data and deliver it to the renderer (as untouched as possible), its most likely the renderer screws up something during its own RGB conversion.
I don't have any control over what the renderer does with it. Especially 4:4:4 data is usually very badly supported by "standard" renderers like EVR, so RGB is the recommended output format.

It might be a result of the dithering, afterall you're converting a 10-bit file to 8-bit, which will loose data, and then convert it to RGB in the renderer, which again is a lossy process.
The best course of action is to convert in LAV Video to RGB (which will retain the 10-bit all through the processing path up to the end), or use a renderer which understands 10-bit input (only madVR a this time)


Btw: Memory consumption with huge x264 RGB file and LAV splitter is now reduced by a half (~1.2 GB), but that's still twice as much than with MPC's internal splitter.

Especially with 10-bit 4:4:4 material, the decoder itself also needs quite a bit of memory (especially with multi-threading)
The splitter limits itself to 512MB now, i guess i could cut that down a bit still, however 1.2GB really seems to be well within acceptable limits for playback of such formats.

The queues that cache all those frames in memory serve a real purpose, so lowering them isn't always a good idea.

aufkrawall
6th February 2012, 17:04
It might be a result of the dithering, afterall you're converting a 10-bit file to 8-bit, which will loose data, and then convert it to RGB in the renderer, which again is a lossy process.
The best course of action is to convert in LAV Video to RGB (which will retain the 10-bit all through the processing path up to the end), or use a renderer which understands 10-bit input (only madVR a this time)

Is there any drawback when forcing RGB output (with EVR custom or VMR-9)?
Somebody stated some sites back in this thread that there would be a quality loss because dither noise is applied before upscaling or at least it would be better to apply dither noise after upscaling like madVR does.

nevcairiel
6th February 2012, 17:06
Is there any drawback when forcing RGB output (with EVR custom or VMR-9)?
Somebody stated some sites back in this thread that there would be a quality loss because dither noise is applied before upscaling or at least it would be better to apply dither noise after upscaling like madVR does.

This is generally true, however for 10-bit this doesn't apply, because 10-bit always needs to be dithered - EVR and VMR9 don't accept 10-bit input formats. So instead of dithering the YUV, i would claim you can get a better quality by using the full 10-bit for RGB creation and dithering at the end of that process.

JarrettH
6th February 2012, 17:08
I always noticed in the past that RGB32 output looked better when I used ffdshow. Is RGB32 preferred in certain setups?

nevcairiel
6th February 2012, 17:09
I always noticed in the past that RGB32 output looked better when I used ffdshow. Is RGB32 preferred in certain setups?

If you don't use madVR, you can get better results from RGB32 output, because alot of renderers use rather crappy upscaling algorithms for the chroma, and also have no access to the matrix or range flags.
However, using RGB32 output breaks renderer deinterlacing, so it should only be used if you have progressive content or using a decoder which can deinterlace (CUVID deinterlacing or YADIF, for example)

aufkrawall
6th February 2012, 17:24
This is generally true, however for 10-bit this doesn't apply, because 10-bit always needs to be dithered - EVR and VMR9 don't accept 10-bit input formats. So instead of dithering the YUV, i would claim you can get a better quality by using the full 10-bit for RGB creation and dithering at the end of that process.
Ok, thanks for the info. :)
But I find it strange that I have to disable AYUV in LAV to get RGB output, it's only listed for 8bit video.
Are those portrayal columns simplified and AYUV can also apply to 10bit video, or is it because of the renderer's 10bit incompatibility?

I always noticed in the past that RGB32 output looked better when I used ffdshow. Is RGB32 preferred in certain setups?
I got some strange artifacts when using RGB output for a Theora 4:2:0 video.
Didn't happen with LAV.

If you don't use madVR, you can get better results from RGB32 output, because alot of renderers use rather crappy upscaling algorithms for the chroma, and also have no access to the matrix or range flags.
However, using RGB32 output breaks renderer deinterlacing, so it should only be used if you have progressive content or using a decoder which can deinterlace (CUVID deinterlacing or YADIF, for example)
Can those of us who don't use madVR have a checkbox to simply output RGB with a corresponding mouse-tooltip that says what you just said? :D

Xaurus
6th February 2012, 17:42
Slightly off-topic, but does anyone know if it is possible to run the Intel 2000 IGP at the same time as a Nvidia 450 GTS?

I mean, connect the display to the 450 and just take the audio from the Intel 2000 IGP.

I've search everywhere but I can't really find an answer.

BetA13
6th February 2012, 19:15
I found some strange behavior/maybe bug:
Somehow brightness of a x264 I444 10bit video doesn't look fully right without converting to RGB when decoding while playing.

This is how the original FRAPS RGB looks like:
http://www.ld-host.de/uploads/thumbnails/e494e8bd47438960d33d94b080e5ad46.png (http://www.ld-host.de/show/e494e8bd47438960d33d94b080e5ad46.png)

x264 I444 10bit with default LAV decoding settings (CUVID) (a very little too dark):
http://www.ld-host.de/uploads/thumbnails/cea7ebe79549ed195ece54bc87a69e2f.png (http://www.ld-host.de/show/cea7ebe79549ed195ece54bc87a69e2f.png)

The same video with forced RGB32 output (to me it looks exactly like the original FRAPS RGB video):
http://www.ld-host.de/uploads/thumbnails/a7ae5ad2f56418b1b46bd017bee44509.png (http://www.ld-host.de/show/a7ae5ad2f56418b1b46bd017bee44509.png)

Download a test sample (it's not crf 0 like the pictures but this shouldn't make any difference):
http://www.mediafire.com/?n5ymnpdasada4gg
Video has TV range (Rec709).

It would be nice if it looked right without having to force RGB32 output. :)
I don't know what's about 8bit or I420, didn't test it.
Tested with latest MPC HC, LAV nightly and EVR custom.

Btw: Memory consumption with huge x264 RGB file and LAV splitter is now reduced by a half (~1.2 GB), but that's still twice as much than with MPC's internal splitter.



hy there mate.,..
i had teh same problem with FRAPS...i switched to MSI Afterburner..it has better features in video recording ingame..RTV!/mjpeg codec and one more..good quality..no diff to fraps, but its easyer to work with the files afterwards and the colors are better, well not better. lets say just gamma right...

heres an quote from the dev of afterburner:

o Added gamma corrected video capture support

its mostly not the CODEC from ur player causing this but the recording software..

give it a try if u want...u will love it, better then fraps..also u dont have 4 gb files in teh end..afterburner can record much bigger in a row..

http://www.abload.de/img/unbenanntnjuyg.jpg

http://forums.guru3d.com/showthread.php?t=358092

greetz BetA

Thunderbolt8
6th February 2012, 21:02
In general, quality is always the same no matter which decoder you use.even though this should be the case, sometimes I get more image corruption or in general a strange kind of flickerin of parts of the image when using coreavc. happens especially in case of documentaries. (its not there all the time, but there seems to be a certain starting point and from then on it doesnt seem to disappear any more)

CruNcher
7th February 2012, 05:34
@Nev
RV4 issue

http://www.mediafire.com/?dagppuv8p8wppnr

ffplay = ok
avplay = fail
vlc = ok
mpc hc splitter + lav video = ok
av splitter + lav video = fail
lav splitter + lav video = fail

oddball
7th February 2012, 06:54
I'm having trouble with a certain video using CUVID. It appears to put excessive strain on the CPU when high quality processing is enabled in LAV Video decoder. I have tested other videos and they play fine but this one can cause dropped frames on occasion and I am not sure why. I use MPC-HC latest with latest MadVR but I sometimes put ffdshow RAW filters in the chain for delogo or levels and with this clip it just drops frames all over the place when I do this. I can play other 720p and 1080p videos and nairy a dropped frame is in sight. So I don't know why this is different. I have an OC'd E7600 and Nvidia GTX560 so it should handle most stuff. If I use DXVA copy back or Quicksync with the dll (Not sure if it's doing anything on an E7600 to be honest) then everything plays fine but then it's not using high quality processing I guess since it goes greyed out.

http://www.mediafire.com/?ab69qo3wibfmb9j

sexus
7th February 2012, 07:06
well latest lav workin nicely , thou freezes if loading up a .avi file if i disable mpeg4 in lav videos format listing it works , please fix this id appreciate it as would others too id asume

nevcairiel
7th February 2012, 07:53
@Nev
RV4 issue

http://www.mediafire.com/?dagppuv8p8wppnr


Looks like the multi-thread freeze/hang issue i already reported to ffmpeg/libav. They have to fix it. Works fine if you play with 1 thread

Aleksoid1978
7th February 2012, 08:08
Looks like the multi-thread freeze/hang issue i already reported to ffmpeg/libav. They have to fix it. Works fine if you play with 1 thread

in MPC-HC play fine with multi-thread.

sagematt
7th February 2012, 09:55
I'm getting a nasty repeated frames issue when using QuickSync decoder. This capture (http://dl.dropbox.com/u/355632/20120207_035151.mkv) easily illustrates my problem.
I reproduced the problem on LAV Filters 0.43, 0.44 and 0.45. It happens with both EVR Custom and madVR.
CPU is Core i7-2630QM

nevcairiel
7th February 2012, 09:57
I'm getting a nasty repeated frames issue when using QuickSync decoder. This capture (http://dl.dropbox.com/u/355632/20120207_035151.mkv) easily illustrates my problem.
I reproduced the problem on LAV Filters 0.43, 0.44 and 0.45. It happens with both EVR Custom and madVR.
CPU is Core i7-2630QM

Install the latest Intel drivers directly from Intel.com, not from Windows Update.
It may also be caused by Lucid Virtu if you run that.

nevcairiel
7th February 2012, 11:41
in MPC-HC play fine with multi-thread.

Its a combination of splitter and video decoder bug. MPC-HC decoder also fails with LAV Splitter.
I'll try to hack together a temporary workaround for LAV Splitter until the decoder in libav can be fixed

egur
7th February 2012, 15:37
I'm getting a nasty repeated frames issue when using QuickSync decoder. This capture (http://dl.dropbox.com/u/355632/20120207_035151.mkv) easily illustrates my problem.
I reproduced the problem on LAV Filters 0.43, 0.44 and 0.45. It happens with both EVR Custom and madVR.
CPU is Core i7-2630QM

This clip is mpeg4 (MP4V fourcc). QS doesn't support mpeg4. Problem is somewhere else...

nevcairiel
7th February 2012, 15:39
This clip is mpeg4 (MP4V fourcc). QS doesn't support mpeg4. Problem is somewhere else...

I think he just recorded that to illustrate the original problem. :p
It seems like the old problem caused by odd drivers or Virtu

ajp_anton
7th February 2012, 16:23
This clip is mpeg4 (MP4V fourcc). QS doesn't support mpeg4. Problem is somewhere else...Isn't h.264 part of mpeg4? I always get confused when only "mpeg4" is mentioned... what exactly does it mean?

DragonQ
7th February 2012, 16:31
"MPEG4" usually refers to MPEG 4 Part 2 (e.g. DivX, Xvid, H.263). AVC/H.264 is MPEG 4 Part 10.

nevcairiel
7th February 2012, 16:44
Yeah, when people say MPEG4 they mean the old MPEG4, H264/AVC1 is usually always referred to by those designations.

oddball
7th February 2012, 17:48
Its a combination of splitter and video decoder bug. MPC-HC decoder also fails with LAV Splitter.
I'll try to hack together a temporary workaround for LAV Splitter until the decoder in libav can be fixed

Cheers. I look forward to it! :)

ruinevil
7th February 2012, 18:17
Can CUVID accelerate 10 bit h264?

I thought it couldn't; LAV splitter would just dump the video to the CPU.

nevcairiel
7th February 2012, 18:38
Can CUVID accelerate 10 bit h264?


It cannot. It'll fallback to software decoding automatically.

Pat357
7th February 2012, 18:51
I reverted to 257 release drivers (which is also the last version that CUVID works right with interlaced VC-1
What problem do you have with CUVID concerning VC-i ?
It stopped working if I understand this correct ? Only on your older 9xxx card or also on your newer ?

nevcairiel
7th February 2012, 19:05
What problem do you have with CUVID concerning VC-i ?
It stopped working if I understand this correct ? Only on your older 9xxx card or also on your newer ?

There seems to be an issue regarding partial acceleration with some drivers (so it would only affect older cards)

dead_screem
7th February 2012, 20:39
What problem do you have with CUVID concerning VC-i ?
It stopped working if I understand this correct ? Only on your older 9xxx card or also on your newer ?
horrible blocking on certain parts of video when hardware deinterlacing is active. Apparently only affects "partial" acceleration. VC-1 and MPEG-2, although I havn't tested it with MPEG-2 myself. 258.96 is the last version that is not bugged.
if you really wanna see, i can upload screenshots if i can find where i saved them.

sexus
7th February 2012, 22:00
well latest lav workin nicely , thou freezes if loading up a .avi file if i disable mpeg4 in lav videos format listing it works , please fix this id appreciate it as would others too id asume

Thunderbolt8
7th February 2012, 22:41
for some reason, dxva2 copy back works better with VC-1 for me than with AVC, while its the other way round with coreavc and MS decoder dmo. I only get few dropped and delayed frames with 30mbps VC-1, but a really non smooth picture with 18mbps AVC.

with coreavc, I can play stuff up to 35-40mbps without problems (mostly).

kostasoft
8th February 2012, 01:26
2 nevcairiel
It's possible to make support of multistream video in the container mkv that splitter divided both streams and allowed to reproduce them simultaneously? It is necessary for reproduction of three-dimensional films.

roytam1
8th February 2012, 04:23
@nev: access violation at the end of GraphStudioNext decoder performance test, testing RV40 decoder with a rmvb file.

eax=00000000 ebx=02b92180 ecx=02b4baec edx=02af8480 esi=02b920c8 edi=02b920c8
eip=02af848c esp=0012fa84 ebp=0012fa84 iopl=0 nv up ei pl zr na po nc
cs=001b ss=0023 ds=0023 es=0023 fs=003b gs=0000 efl=00000246

function: LAVSplitter!CLAVSplitter__AbortOperation
02af847b cc int 3
02af847c cc int 3
02af847d cc int 3
02af847e cc int 3
02af847f cc int 3
LAVSplitter!CLAVSplitter__AbortOperation:
02af8480 55 push ebp
02af8481 8bec mov ebp,esp
02af8483 8b4508 mov eax,[ebp+0x8]
02af8486 8b80c0000000 mov eax,[eax+0xc0]
fault ->02af848c 8b08 mov ecx,[eax] ds:0023:00000000=????????
02af848e 8b5118 mov edx,[ecx+0x18]
02af8491 6a01 push 0x1
02af8493 50 push eax
02af8494 ffd2 call edx
02af8496 5d pop ebp
02af8497 c20400 ret 0x4
02af849a cc int 3
02af849b cc int 3
02af849c cc int 3
02af849d cc int 3

*----> Stack trace <---*
*** ERROR: Module load completed but symbols could not be loaded for D:\_Programs\graphstudionext.exe
WARNING: Stack unwind information not available. Following frames may be wrong.
*** ERROR: Symbol file could not be found. Defaulted to export symbols for C:\WINDOWS\system32\USER32.dll -
*** ERROR: Symbol file could not be found. Defaulted to export symbols for C:\WINDOWS\system32\kernel32.dll -
ChildEBP RetAddr Args to Child
0012fa84 02af7442 02b9216c e8c999b5 02b920c8 LAVSplitter!CLAVSplitter__AbortOperation+0xc
0012faac 02af71b3 e8c999c5 02b920c8 02b920c8 LAVSplitter!CLAVSplitter__Close+0x52
0012fadc 02af6ebb 02b920d0 0012fafc 02b25253 LAVSplitter!CLAVSplitter__~CLAVSplitter+0xc3
0012fae8 02b25253 00000001 00ef9350 00f06fb8 LAVSplitter!CLAVSplitter__`scalar deleting destructor'+0xb
0012fafc 02af6661 02b920c8 0012fb40 005809c9 LAVSplitter!CUnknown__NonDelegatingRelease+0x23
0012fb08 005809c9 02b920d4 00f06fb8 00f06fb8 LAVSplitter!CLAVSplitter__Release+0x11
0012fb40 0040108b 00000001 0012fb64 0057cfee graphstudionext+0x1809c9
0012fb4c 0057cfee 00000001 00000005 00000000 graphstudionext+0x108b
0012fb64 00581f30 3db37c25 00f039c0 00ef9350 graphstudionext+0x17cfee
0012fba4 00582d58 00ef9350 00000000 00582e47 graphstudionext+0x181f30
0012fbf0 005755d9 00f066a8 00f1b4a8 00ef91e8 graphstudionext+0x182d58
0012fc10 0056f9f1 3db37b0d 00000000 00ef91e8 graphstudionext+0x1755d9
0012fc58 00577ad2 00000000 0056d326 3db37a59 graphstudionext+0x16f9f1
0012fd0c 0056daa2 00ef9ed8 00000770 00ef9ed8 graphstudionext+0x177ad2
0012fd34 00423dfd 00000770 00000000 00ef9350 graphstudionext+0x16daa2
0012fd54 004272b9 00000770 00000000 00ef9350 graphstudionext+0x23dfd
0012fdc8 00427348 00ef9ed8 0007076a 00000770 graphstudionext+0x272b9
0012fde8 77d18734 0007076a 00000770 00000000 graphstudionext+0x27348
0012fe14 77d18816 00427312 0007076a 00000770 USER32!GetDC+0x6d
0012fe7c 77d189cd 00000000 00427312 0007076a USER32!GetDC+0x14f
0012fedc 77d18a10 00177020 00000000 0012ff1c USER32!GetWindowLongW+0x127
0012feec 004204c3 00177020 00000000 00665ab0 USER32!DispatchMessageW+0xf
0012ff1c 005409c7 80000001 00000000 7ffd4000 graphstudionext+0x204c3
0012ff30 00527404 00400000 00000000 0002096e graphstudionext+0x1409c7
0012ffc0 7c817077 80000001 036aca00 7ffd4000 graphstudionext+0x127404
0012fff0 00000000 00527457 00000000 78746341 kernel32!RegisterWaitForInputIdle+0x49

EDIT: when I revert this commit (http://git.1f0.de/gitweb?p=lavfsplitter.git;a=commitdiff;h=d1db625b604f21eab896d5f9a79686c03005b62a), it works without crashing again

Aleksoid1978
8th February 2012, 08:02
Hi Nev
Can you modify you splitter, that for RealAudio set to output "normal" subtype:
MEDIASUBTYPE_COOK, MEDIASUBTYPE_ATRC and etc. This need for support AudioDecoder that don't support speciality WAVEFORMATEXFFMPEG.

nevcairiel
8th February 2012, 08:06
It already does output more then one media type. The LAV special one is always the second one in the list.
I can look at them and see if they are "standard", but they'll still not work with the RealAudio DLL based decoder.

roytam1
8th February 2012, 08:24
It already does output more then one media type. The LAV special one is always the second one in the list.
I can look at them and see if they are "standard", but they'll still not work with the RealAudio DLL based decoder.
quoting my edited message some posts before:
when I revert this commit (http://git.1f0.de/gitweb?p=lavfsplitter.git;a=commitdiff;h=d1db625b604f21eab896d5f9a79686c03005b62a), it works without crashing again

nevcairiel
8th February 2012, 08:29
quoting my edited message some posts before:

I already fixed it properly, but i cannot push changes from work :)

betaking
8th February 2012, 09:22
NEV last lav git play some h.264 or x.264 give me carsh!
Faulting application mpc-hc.exe, version 1.6.1.4048, time stamp 0x4f321266, faulting module LAVSplitter.ax, version 0.46.0.0, time stamp 0x4f31f2a6, exception code 0xc0000005, fault offset 0x0000860c, process id 0xdb0, application start time 0x01cce63a5cdd1c60.
Git commit: ae4d0e034729731344009c97be61def690ab0864

Mercury_22
8th February 2012, 09:24
NEV last lav 0.46 git play some h.264 or x.264 give me carsh!
Faulting application mpc-hc.exe, version 1.6.1.4048, time stamp 0x4f321266, faulting module LAVSplitter.ax, version 0.46.0.0, time stamp 0x4f31f2a6, exception code 0xc0000005, fault offset 0x0000860c, process id 0xdb0, application start time 0x01cce63a5cdd1c60.

Yeah last lavsplitter x64 keeps crashing everything not just h264 e.g. com surrogate too (when exploring my videos folder) (not windows explorer! ):mad: Also the crash message appears after a few sec after closing e.g. MPC not immediately after closing

ageback
8th February 2012, 09:24
NEV last lav 0.46 git play some h.264 or x.264 give me carsh!
Faulting application mpc-hc.exe, version 1.6.1.4048, time stamp 0x4f321266, faulting module LAVSplitter.ax, version 0.46.0.0, time stamp 0x4f31f2a6, exception code 0xc0000005, fault offset 0x0000860c, process id 0xdb0, application start time 0x01cce63a5cdd1c60.

Confirmed.Crash caused by LAV Splitter.
Git commit: ae4d0e034729731344009c97be61def690ab0864

nevcairiel
8th February 2012, 09:36
Is it too hard to just read the last few posts? It has already been covered. Reporting the same issue 5 times on the same page doesn't help, especially if i said its already fixed. :p

Also, if you use Git builds, its at your own risk. There is no guarantee that it'll work.

PS:
There is no such thing as 0.46

betaking
8th February 2012, 09:43
Is it too hard to just read the last few posts? It has already been covered. Reporting the same issue 5 times on the same page doesn't help, especially if i said its already fixed. :p

Also, if you use Git builds, its at your own risk. There is no guarantee that it'll work.

PS:
There is no such thing as 0.46

sorry! can you fix it?

Kurtnoise
8th February 2012, 09:44
He said it's already fixed...do you understand ?

nevcairiel
8th February 2012, 09:45
sorry! can you fix it?

I cannot. It'll have to keep crashing for ever now. :rolleyes:

roytam1
8th February 2012, 09:48
sorry! can you fix it?

It is fixed but the commit doesn't push to git tree yet.
For the time being you can use my build, which reverts a commit to make it working again:
http://roy.orz.hm/lavf-w32-nightlies/lavf-my120208-revert_d1db625b6-ae4d0e0.7z

diff:
http://roy.orz.hm/lavf-w32-nightlies/lavf-my120208-revert_d1db625b6-ae4d0e0.diff

Aleksoid1978
8th February 2012, 13:16
It already does output more then one media type. The LAV special one is always the second one in the list.
I can look at them and see if they are "standard", but they'll still not work with the RealAudio DLL based decoder.

No - you set output non standart subtype. Example - COOK codec.
DEFINE_GUID(MEDIASUBTYPE_COOK,
0x4b4f4f43, 0x0000, 0x0010, 0x80, 0x00, 0x00, 0xAA, 0x00, 0x38, 0x9B, 0x71);

but LAV splitter output is different :
subtype: Unknown GUID Name {6B6F6F63-0000-0010-8000-00AA00389B71}

That's what I'm talking about.

nevcairiel
8th February 2012, 13:19
That's what I'm talking about.

No its not what you were talking about. You talked about my own special LAV type, which is completely independent from everything else.

Its not my fault that the RM files have "cook" as FourCC inside the file but someone decided to use "COOK" as FourCC in DirectShow. :p

Aleksoid1978
8th February 2012, 13:24
Its not my fault that the RM files have "cook" as FourCC inside the file but someone decided to use "COOK" as FourCC in DirectShow. :p

But - you use "COOK" as FourCC in you code :)
also - Atrac3

// RealMedia Audio
{ &MEDIASUBTYPE_COOK, CODEC_ID_COOK },
{ &MEDIASUBTYPE_ATRC, CODEC_ID_ATRAC3 },

aufkrawall
8th February 2012, 17:51
I'm sorry to bother you again, nevcairiel, but I'm very sure that there is a bug on decoder site regarding x264 I444 video. ;)
The brightness difference also occurs with 8bit video, so it can't be a dithering issue of the renderer.
I also checked if it could be a problem of Avisynth.
The script is:
LoadCPlugin("C:\Program Files (x86)\AviSynth 2.6\plugins\ffms2.dll")
FFVideoSource("trinergb.avi", colorspace="RGB32")
ConvertToYV24(matrix="Rec709")
AssumeFPS(30)
When I use this script but let x264 output only I420 the results looks absoulutely right (*bright*), it's not a decoder conversion issue.
Here's a I444 8bit sample:
http://www.mediafire.com/?msr5x58laeb0j32

Please take a look at it, probably it's a bug of ffmpeg or x264 encoder? :)

nevcairiel
8th February 2012, 17:53
If the RGB output looks correct, its not a decoder issue, but like i said before, a issue in the renderer and its handling of 4:4:4 YUV.
Even if it was a decoder issue, i didn't write the H264 decoder, thats all in the hands of ffmpeg/libav.

daWsOn_s
8th February 2012, 19:33
Hello,
I'm trying to get subtitles to work.
The files is m2ts directly from the BD.
The subs are in pgs, I have enabled lav video/audio/splitter and all works, the subs are recognized too. When I select one sub track from the stream with lav splitter they don't show up. I also tried enabling ffdshow sub fillter but no luck.
What is the way to show them?
I'm using MPC HC latest nightly of course.

PS: they work in VLC

CruNcher
8th February 2012, 21:32
@Nev

very strange dvcpro 50 decoding issue

http://www.mediafire.com/?ioa60f2np2fbife

avplay = ok
ffplay = ok
lav splitter + lav video = fail

nevcairiel
8th February 2012, 22:01
@Nev

very strange dvcpro 50 decoding issue

http://www.mediafire.com/?ioa60f2np2fbife


Fixed

Pat357
9th February 2012, 00:46
The script is:
LoadCPlugin("C:\Program Files (x86)\AviSynth 2.6\plugins\ffms2.dll")
FFVideoSource("trinergb.avi", colorspace="RGB32")
ConvertToYV24(matrix="Rec709")
AssumeFPS(30)
When I use this script but let x264 output only I420 the results looks absoulutely right (*bright*), it's not a decoder conversion issue.
I'm afraid it's not that simple to make conclusions like that.

1.What codec is used for your source "trinergb.avi".
If your source is not directly decoded in RGB32, FFMS2 will do a YUV conversion to RGB32 because that's what you ask it to do in the second line of your script.
You will surely get RGB32, but the question is : has it been converted from YUV-> RGB32 first by FFMS2 or not and if it was, how was it converted (matrix ? levels ?)

2.What version of FFMS2 are you using ? The previous version, FFMS v2.16, had major problems with the levels and did always use "full range" (0-255).
The recently released FFMS v2.17 should fix this, but I didn't check it yet. There might be still something not 100% in the first "fixed" release....
I didn't folow the FFMS2 thread, but you can read it here on Doom9 and on Doom10.

3. What version and/or MOD of x264 are you using ?
There have been a lot of discusions about 10bit -> 8bit colorspace/level problems and incorrect levels (not follwing BT709 matrix) with the "official" x264 version.
There are also some x264 MOD's that claim to fix some colorspace conversions/levels. (see for example http://x264.fushizen.eu/ ).
I would recommend the safest way : this is not letting x264 do any 10-> 8 or 8 -> 10 bit conversions and always feeding it "16-235" (lim. level).
I suspect that when feeding it RGB32, it will start with a RGB32 -> YUV conversion, so here could it already go wrong.

4. To check the color visually, I trust MadVR far more than EVR if you talk about 4:4:4 input.
Of course if you use LAV to convert to RGB32, than even EVR or VMR should be able to handle that:p

Note : about x264 vanilla; I just checked with the --version command, the latest build 2146
x264_32b_8b__2146_unpatched --version
x264 0.120.2146 bcd41db
(libswscale 2.1.0)
(libavformat 53.20.0)
(ffmpegsource 2.16.4.0)
built on Jan 19 2012, gcc: 4.6.2
configuration: --bit-depth=8 --chroma-format=all
x264 license: GPL version 2 or later
libswscale/libavformat/ffmpegsource license: GPL version 3 or later
Not sure if this is related to the above mentioned "x264 issues", but it seems like this x264 version was build with the known faulty FFMS v2.16. Maybe the x264 guys patched something to get around this, this should be traceable in the logs.

Reino
9th February 2012, 01:13
If his source is yuvj420p (normal Fraps recording method), FFVideoSource(colorspace="RGB32") + ConvertToYV24(matrix="Rec709") produce an entirely different image compared to AVISource (Fraps through VFW) and the (almost) equivalent FFVideoSource() + ConvertToRGB(matrix="PC.709").
If his source is bgra (forced lossless RGB in Fraps), and although the recording of FPS1(bgra) files on my computer has issues (http://forum.doom9.org/showthread.php?p=1556898#post1556898), I don't think FFVideoSource would need any matrix or level conversion at all, because, although faulty, opening one of these FPS1(bgra) files with AVISource or FFVideoSource, without any conversion, produces the same result here (png screenshots are identical in filesize).
Or does Fraps record in yet even more colourspaces, with different ranges and matrixes, I don't know of?...

sexus
9th February 2012, 01:25
well latest lav workin nicely , thou freezes if loading up a .avi file if i disable mpeg4 in lav videos format listing it works , please fix this nevcairiel id appreciate it as would others too id asume

Asmodian
9th February 2012, 01:37
lol @ sexus

A mpeg4 .avi file? I believe Nevcairiel will need a bit more information. ;)

Sebastiii
9th February 2012, 07:05
Hi Nevcairiel,

With that sample all player freeze : https://rapidshare.com/files/1846637945/test_low.flv
It didn't happen on version 0.42 :)
Thanks :)

nevcairiel
9th February 2012, 07:42
well latest lav workin nicely , thou freezes if loading up a .avi file if i disable mpeg4 in lav videos format listing it works , please fix this nevcairiel id appreciate it as would others too id asume

You posted the exact same message 4 times or so, which already in itself is considered very rude.

In addition, you don't provide a single piece of useful information.
Try again with a sample file, and i might not ignore you.


With that sample all player freeze : https://rapidshare.com/files/1846637945/test_low.flv
It didn't happen on version 0.42 :)


This is probably already fixed in ffmpeg, but i'm not going to update until the new version is released.

sexus
9th February 2012, 08:47
my bad shouldve given you the mediainfo xD:

General
Complete name : H:\Movies library III\Yaar Anmulle\Yaar Anmulle - DVDScr - XviD - 1CDRip - [DDR].avi
Format : AVI
Format/Info : Audio Video Interleave
File size : 691 MiB
Duration : 2h 25mn
Overall bit rate : 665 Kbps
Writing application : VirtualDubMod 1.5.10.3 | www.virtualdub-fr.org || (build 2550/release)
Writing library : VirtualDubMod build 2550/release

Video
ID : 0
Format : MPEG-4 Visual
Format profile : Advanced Simple@L5
Format settings, BVOP : 1
Format settings, QPel : No
Format settings, GMC : No warppoints
Format settings, Matrix : Custom
Codec ID : XVID
Codec ID/Hint : XviD
Duration : 2h 25mn
Bit rate : 580 Kbps
Width : 624 pixels
Height : 256 pixels
Display aspect ratio : 2.40:1
Frame rate : 23.976 fps
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Progressive
Compression mode : Lossy
Bits/(Pixel*Frame) : 0.152
Stream size : 603 MiB (87%)
Writing library : XviD 1.2.1 (UTC 2008-12-04)

Audio
ID : 1
Format : MPEG Audio
Format version : Version 2
Format profile : Layer 3
Mode : Joint stereo
Mode extension : Intensity Stereo
Codec ID : 55
Codec ID/Hint : MP3
Duration : 2h 25mn
Bit rate mode : Constant
Bit rate : 80.0 Kbps
Channel(s) : 2 channels
Sampling rate : 24.0 KHz
Compression mode : Lossy
Delay relative to video : 13ms
Stream size : 83.2 MiB (12%)
Alignment : Split accross interleaves
Interleave, duration : 1000 ms (23.98 video frames)
Interleave, preload duration : 1008 ms


as said mpeg4 enabled in file list in lav video makes mpchc crash with that file type, not all .avi files but the ones with these properties, thanks

roytam1
9th February 2012, 08:58
my bad shouldve given you the mediainfo xD:

General
Complete name : H:\Movies library III\Yaar Anmulle\Yaar Anmulle - DVDScr - XviD - 1CDRip - [DDR].avi
Format : AVI
Format/Info : Audio Video Interleave
File size : 691 MiB
Duration : 2h 25mn
Overall bit rate : 665 Kbps
Writing application : VirtualDubMod 1.5.10.3 | www.virtualdub-fr.org || (build 2550/release)
Writing library : VirtualDubMod build 2550/release

Video
ID : 0
Format : MPEG-4 Visual
Format profile : Advanced Simple@L5
Format settings, BVOP : 1
Format settings, QPel : No
Format settings, GMC : No warppoints
Format settings, Matrix : Custom
Codec ID : XVID
Codec ID/Hint : XviD
Duration : 2h 25mn
Bit rate : 580 Kbps
Width : 624 pixels
Height : 256 pixels
Display aspect ratio : 2.40:1
Frame rate : 23.976 fps
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Progressive
Compression mode : Lossy
Bits/(Pixel*Frame) : 0.152
Stream size : 603 MiB (87%)
Writing library : XviD 1.2.1 (UTC 2008-12-04)

Audio
ID : 1
Format : MPEG Audio
Format version : Version 2
Format profile : Layer 3
Mode : Joint stereo
Mode extension : Intensity Stereo
Codec ID : 55
Codec ID/Hint : MP3
Duration : 2h 25mn
Bit rate mode : Constant
Bit rate : 80.0 Kbps
Channel(s) : 2 channels
Sampling rate : 24.0 KHz
Compression mode : Lossy
Delay relative to video : 13ms
Stream size : 83.2 MiB (12%)
Alignment : Split accross interleaves
Interleave, duration : 1000 ms (23.98 video frames)
Interleave, preload duration : 1008 ms


as said mpeg4 enabled in file list in lav video makes mpchc crash with that file type, not all .avi files but the ones with these properties, thanks

Don't you think you need to upload a sample for testing and debugging?

sexus
9th February 2012, 09:04
ok shall do

fastplayer
9th February 2012, 09:20
ok shall do
Don't bother. I reported you already.

sexus
9th February 2012, 09:27
http://www.mediafire.com/?lnjean4wf2abgd8 < this i used a different format and switching off mpeg4 in format list wont change the freezing only disabling lav video makes it work

your welcome

p.s: im abit a noob sometimes ok xD

update : ill do another one with a smaller size one sec

http://www.mediafire.com/?jeor5lxxr0tl1kw < this is the one where when you turn off mpeg4 in format list it wont freeze mpchc

aufkrawall
9th February 2012, 11:32
I'm afraid it's not that simple to make conclusions like that.

1.What codec is used for your source "trinergb.avi".
If your source is not directly decoded in RGB32, FFMS2 will do a YUV conversion to RGB32 because that's what you ask it to do in the second line of your script.

Source's CSP is RGB.


2.What version of FFMS2 are you using ? The previous version, FFMS v2.16, had major problems with the levels and did always use "full range" (0-255).
The recently released FFMS v2.17 should fix this, but I didn't check it yet. There might be still something not 100% in the first "fixed" release....

I'm using 2.17, the range conversion just works fine.
I'm sure about this because if I let x264 output I420 with the same AVS script, there is no brightness issue.
It doesn't matter either if I let Avisynth convert to YV12 or YV24.


I didn't folow the FFMS2 thread, but you can read it here on Doom9 and on Doom10.

I already got there useful advices. :)


3. What version and/or MOD of x264 are you using ?
There have been a lot of discusions about 10bit -> 8bit colorspace/level problems and incorrect levels (not follwing BT709 matrix) with the "official" x264 version.

I always use the latest official version.
And it has really nothing to do with 10bit video, the issue also exists for 8bit.


There are also some x264 MOD's that claim to fix some colorspace conversions/levels. (see for example http://x264.fushizen.eu/ ).
I would recommend the safest way : this is not letting x264 do any 10-> 8 or 8 -> 10 bit conversions and always feeding it "16-235" (lim. level).
I suspect that when feeding it RGB32, it will start with a RGB32 -> YUV conversion, so here could it already go wrong.

I agree with your idea of not feeding it with risky formats, that's why I let Avisynth do all the decoding and conversion stuff.
But this doesn't help in the case of I444. :(


4. To check the color visually, I trust MadVR far more than EVR if you talk about 4:4:4 input.
Of course if you use LAV to convert to RGB32, than even EVR or VMR should be able to handle that:p

Yes, this works as a workaround.
Maybe it works only with madVR because it does a RGB conversion anyway.

nevcairiel
9th February 2012, 11:35
Yes, this works as a workaround.
Maybe it works only with madVR because it does a RGB conversion anyway.

Every renderer does a RGB conversion, its just how it works in PCs. The question is how good or bad the conversion is. ;)

Aleksoid1978
9th February 2012, 12:47
Hi Nev.
You have bug in HRESULT CLAVAudio::ParseRealAudioHeader(const BYTE *extra, const int extralen)
and some COOK fail decode.
Need
int ra_extralen = min((extra + extralen) - (fmt+4), *(DWORD*)fmt);
replace with
int ra_extralen = AV_RB32(fmt);

also here
if (codec == CODEC_ID_COOK) {
uint8_t *extra = (uint8_t *)av_mallocz(extralen + FF_INPUT_BUFFER_PADDING_SIZE);
add Atrac codec:
if (codec == CODEC_ID_COOK || codec == CODEC_ID_ATRAC3) {

Sebastiii
9th February 2012, 12:51
This is probably already fixed in ffmpeg, but i'm not going to update until the new version is released.

Thanks :)

avih
9th February 2012, 13:05
sexus, striked. Please read the forum rules. You clearly don't own this movie.

sexus
9th February 2012, 13:19
ummm actually i do own the movie i just tested it out before i purchased it on dvd thats for what torrents are for and its easier to make a sample file from a already pre-ripped movie than having to make one from a dvd from ground up -.-', anyhow if you think i own it or not isnt the issue at hand that ive mentioned , ill be waiting to see what you have to say nevcairiel , im countin on ya bro, peace

p.s: that strike was uncalled for only cause you "clearly" "asume" that i dont own a movie -.-', pro moderation here i can tell

egur
9th February 2012, 21:04
Version 0.26 beta is out with the following changes:
* Added option to disable SW decoding when HW can’t decode. Default is not to decode in SW.
* Even faster memory copy function
* FFDShow rev4313

Download from SourceForge home page (http://sourceforge.net/projects/qsdecoder/)

Cyber-Mav
9th February 2012, 22:16
ummm actually i do own the movie i just tested it out before i purchased it on dvd thats for what torrents are for and its easier to make a sample file from a already pre-ripped movie than having to make one from a dvd from ground up -.-', anyhow if you think i own it or not isnt the issue at hand that ive mentioned , ill be waiting to see what you have to say nevcairiel , im countin on ya bro, peace

p.s: that strike was uncalled for only cause you "clearly" "asume" that i dont own a movie -.-', pro moderation here i can tell

post the proof that you own it then

andyvt
9th February 2012, 22:31
Is there a way to tell which decoding method is actually being used by LAV Video? I'm trying to figure out if a clip is falling back to the software method.

Andy o
9th February 2012, 22:41
If you have HW decoding available, it should say if it's being used right next to the option, in the LAV video control panel, while you're playing the video.

andyvt
9th February 2012, 22:54
Is there a way to tell which decoding method is actually being used by LAV Video? I'm trying to figure out if a clip is falling back to the software method.

Turns out there is, I just wasn't looking close enough.

When playing a clip that can be HWA the "available" text next to the select box changes to "active". If the clip can't be HWA it stays "available".

sexus
10th February 2012, 06:49
post the proof that you own it then

post proof? lols :rolleyes:

heres your proof, the cover scan

http://i1033.photobucket.com/albums/a413/breakdown23/KGrHqRmIE8FFTiZTfBPLueRMDw60_12.jpg

XadoX
10th February 2012, 08:55
Is there a way to open the filter settings directly from windows, without going through a special programm like MPC-HC?

roytam1
10th February 2012, 08:59
Is there a way to open the filter settings directly from windows, without going through a special programm like MPC-HC?

rundll32 LAVAudio.ax,OpenConfiguration
rundll32 LAVSplitter.ax,OpenConfiguration
rundll32 LAVVideo.ax,OpenConfiguration

Andy o
10th February 2012, 09:02
post proof? lols :rolleyes:

heres your proof, the cover scan

http://i1033.photobucket.com/albums/a413/breakdown23/KGrHqRmIE8FFTiZTfBPLueRMDw60_12.jpg

A picture taken from the ebay listing? You know the first rule of digging holes, right?

sexus
10th February 2012, 09:05
atleast i tried ,yarrr!! lols xD, anyhow wont change the problems currently with lav as said , hope this can be fixed

XadoX
10th February 2012, 09:14
rundll32 LAVAudio.ax,OpenConfiguration
rundll32 LAVSplitter.ax,OpenConfiguration
rundll32 LAVVideo.ax,OpenConfiguration

:thanks:

nevcairiel
10th February 2012, 09:15
:thanks:

If you use the installer, it also creates shortcuts like that in the start menu.

pankov
10th February 2012, 11:59
Slightly off-topic, but does anyone know if it is possible to run the Intel 2000 IGP at the same time as a Nvidia 450 GTS?

I mean, connect the display to the 450 and just take the audio from the Intel 2000 IGP.

I've search everywhere but I can't really find an answer.

I think I've already posted it here but I'll do it again.
Yes, it works just fine. I'm using a Z68 mobo but I don't think that will make a difference.

wanezhiling
10th February 2012, 12:23
http://www.gokuai.com/f/xFm30Q4AHAV5dUks

General
Complete name : D:\Others\[CASO][A-Channal][02][BDRIP][1920x1080][X264xAAC][A4D07E42].mp4
Format : MPEG-4
Format profile : Base Media
Codec ID : isom
File size : 539 MiB
Duration : 24mn 36s
Overall bit rate mode : Variable
Overall bit rate : 3 062 Kbps
Encoded date : UTC 2011-08-12 11:22:12
Tagged date : UTC 2011-08-12 11:22:12

Video
ID : 1
Format : AVC
Format/Info : Advanced Video Codec
Format profile : High@L4.1
Format settings, CABAC : Yes
Format settings, ReFrames : 4 frames
Codec ID : avc1
Codec ID/Info : Advanced Video Coding
Duration : 24mn 36s
Bit rate : 2 670 Kbps
Maximum bit rate : 15.0 Mbps
Width : 1 920 pixels
Height : 1 080 pixels
Display aspect ratio : 16:9
Frame rate mode : Constant
Frame rate : 23.976 fps
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Progressive
Bits/(Pixel*Frame) : 0.054
Stream size : 470 MiB (87%)
Writing library : x264 core 112 r1867 22bfd31
Encoding settings : cabac=1 / ref=4 / deblock=1:1:1 / analyse=0x3:0x133 / me=umh / subme=7 / psy=1 / psy_rd=0.40:0.00 / mixed_ref=1 / me_range=16 / chroma_me=0 / trellis=1 / 8x8dct=1 / cqm=0 / deadzone=21,11 / fast_pskip=1 / chroma_qp_offset=-2 / threads=6 / sliced_threads=0 / nr=0 / decimate=1 / interlaced=0 / constrained_intra=0 / bframes=5 / b_pyramid=0 / b_adapt=2 / b_bias=0 / direct=3 / weightb=1 / open_gop=0 / weightp=2 / keyint=300 / keyint_min=20 / scenecut=40 / intra_refresh=0 / rc_lookahead=40 / rc=crf / mbtree=1 / crf=18.0 / qcomp=0.60 / qpmin=10 / qpmax=51 / qpstep=4 / vbv_maxrate=38000 / vbv_bufsize=30000 / crf_max=0.0 / nal_hrd=none / ip_ratio=1.40 / aq=1:1.00
Encoded date : UTC 2011-06-05 18:04:11
Tagged date : UTC 2011-08-12 11:22:31

Audio #1
ID : 2
Format : AAC
Format/Info : Advanced Audio Codec
Format profile : LC
Codec ID : 40
Duration : 24mn 36s
Bit rate mode : Variable
Bit rate : 192 Kbps
Maximum bit rate : 238 Kbps
Channel(s) : 2 channels
Channel positions : Front: L R
Sampling rate : 48.0 KHz
Compression mode : Lossy
Stream size : 34.0 MiB (6%)
Encoded date : UTC 2011-08-12 11:22:25
Tagged date : UTC 2011-08-12 11:22:31

Audio #2
ID : 3
Format : AAC
Format/Info : Advanced Audio Codec
Format profile : LC
Codec ID : 40
Duration : 24mn 36s
Bit rate mode : Variable
Bit rate : 192 Kbps
Maximum bit rate : 227 Kbps
Channel(s) : 2 channels
Channel positions : Front: L R
Sampling rate : 48.0 KHz
Compression mode : Lossy
Stream size : 34.1 MiB (6%)
Encoded date : UTC 2011-08-12 11:22:29
Tagged date : UTC 2011-08-12 11:22:31




LAV Splitter select audio 2 not 1 by default.. audio 2 is just some commentary.
http://i.imgur.com/G3Hg3.png

egur
10th February 2012, 14:46
@Nev,
I quick question.
this clip (http://www.digital-digest.com/movies/The_Avengers_1080p_Theatrical_Trailer.html) works great using LAV splitter buy Haali reports a frame rate of 90,000fps.
The SPS header indeed reports this fps:
sps.time_scale = 180000
sps.num_units_in_tick = 1

LAV splitter reports correctly a value of AvgTimePerFrame = 417225 within the video info header. Haali reports zero.
Since the SPS data is obviously wrong, how do you find the real frame rate? From looking at several samples?

nevcairiel
10th February 2012, 15:01
I have a hard-limit of i think 125 fps, if the header reported value is above that i don't trust it and use the (somewhat inaccurate) measured value (417225 isn't exact)

This is also the reason why i rather use the timestamps instead of basing any calculations on the fps. :)

egur
10th February 2012, 15:08
I have a hard-limit of i think 125 fps, if the header reported value is above that i don't trust it and use the (somewhat inaccurate) measured value (417225 isn't exact)

This is also the reason why i rather use the timestamps instead of basing any calculations on the fps. :)

Thanks.
Unfortunately your time stamp trickery is possible at the filter level only.

Pat357
10th February 2012, 21:56
post proof? lols :rolleyes:
heres your proof, the cover scan
Did this picture come from your scanner ???
Why than there this text in the image :
"Processed By eBay with ImageMagick, R1.1.1.||B2"
:p
You must have a special scanner ..:p

clsid
10th February 2012, 23:11
Nev, you need to add "IsWin64" checks in the new installer before accessing HKCR64. Otherwise there will be an error on x86.

SamuriHL
10th February 2012, 23:21
Oh great, he changed the installer again? sigh. :) Guess I'll have to rewrite mine again. :D

nevcairiel
10th February 2012, 23:43
Oh great, he changed the installer again? sigh. :) Guess I'll have to rewrite mine again. :D

Again? I never really touched it much. :p
This change now has been a long time coming though.

SamuriHL
10th February 2012, 23:56
Again? I never really touched it much. :p
This change now has been a long time coming though.

Yea, you had made some change a bunch of versions back and I had to fix mine again. Maybe when you added the intel stuff. can't remember. Not a big deal. I basically start with yours and hit delete a bunch of times. :D HEHEHE. Remember mine's only used to update the J River private copy and register it so that I only have one updated copy on my machine. And using the installer ensures it gets registered properly.

sexus
11th February 2012, 02:59
Did this picture come from your scanner ???
Why than there this text in the image :
"Processed By eBay with ImageMagick, R1.1.1.||B2"
:p
You must have a special scanner ..:p

actually i do ;), hope you can do something about the mpeg4 incompatibility you can clearly see from the second clip ive uploaded nevcairiel id appreciate alot and the first clip freezing with h264 enabled in format list;)

ryrynz
11th February 2012, 05:59
Bit of a longshot, but does anyone know how I could possibly fix a huge startup delay when requesting files over a network when File Source Async is in the chain?
Loading local files is fine, also using LAV Splitter Source over the network is fine.

RBG
11th February 2012, 06:28
ryrynz

I just tried, and I have no delay opening files over the network with File Source Async in the chain, well at least with Windows Server and gigabit ethernet.

MaxMan23
11th February 2012, 06:48
Hi, i have a problem with lav 0.45 and cuvid. i get no video (black screen) when cuvid is selected in the settings (dxva2 works). Tested with mediaportal und mpc. But why? My PC is a q6600 @ 3.6 ghz, Geforce 560 Ti

boyumeow
11th February 2012, 07:09
Xvid do not work with nvidia driver more than 285, if I'm not wrong. Thanks.

ryrynz
11th February 2012, 09:31
ryrynz

I just tried, and I have no delay opening files over the network with File Source Async in the chain, well at least with Windows Server and gigabit ethernet.

I know it's an issue limited to my HTPC but I don't know what's causing it, was just wondering if anyone had seen it before. :thanks:

Hi, i have a problem with lav 0.45 and cuvid. i get no video (black screen) when cuvid is selected in the settings (dxva2 works). Tested with mediaportal und mpc. But why? My PC is a q6600 @ 3.6 ghz, Geforce 560 Ti

What driver version you using? What is the file type are your attempting to play back? Tried a fresh install of LAV filters?

nevcairiel
11th February 2012, 09:32
I know it's an issue limited to my HTPC but I don't know what's causing it, was just wondering if anyone had seen it before. :thanks:

If you dont have issues using LAV as the source directly -> problem solved? :)

Tacio
11th February 2012, 09:57
How can I use LAV Splitter Source as file source instead File Source Async? I set Prefer merit for LAV Splitter Source and block File Source Async, but mpc shows that it can't render stream at all. In LAV Splitter Source I checked all formats.

nevcairiel
11th February 2012, 10:04
How can I use LAV Splitter Source as file source instead File Source Async? I set Prefer merit for LAV Splitter Source and block File Source Async, but mpc shows that it can't render stream at all. In LAV Splitter Source I checked all formats.

You need to install via the Installer, it'll set that up automatically.

cremor
11th February 2012, 10:39
You need to install via the Installer, it'll set that up automatically.

Could there be a bug in the installer that was introduced in one of the last few versions?
I'm using all four "LAV x" external filter options as prefer in MPC-HC since I started using LAV and previously that was everything that was needed. But some time ago I noticed that AVIs are using File Source Async (and LAV Splitter) whereas MKVs still use LAV Splitter Source.

I didn't change any of the format options in the installer or the splitter configuration.

nevcairiel
11th February 2012, 10:48
You may have installed something else that reset the extension mapping. In any case, the next version has a basically completely rewritten installer, so if there were bugs, they are no more (but i'm sure new ones instead!)

MaxMan23
11th February 2012, 15:25
Hi, i have a problem with lav 0.45 and cuvid. i get no video (black screen) when cuvid is selected in the settings (dxva2 works). Tested with mediaportal und mpc. But why? My PC is a q6600 @ 3.6 ghz, Geforce 560 Ti

My driver version is 290.36 and the files are xvid (avi).

nevcairiel
11th February 2012, 15:27
My driver version is 290.36 and the files are xvid (avi).

If you activate CUVID acceleration for XVID, you do so at your own risk. If it doesn't work, turn it off. :p
There is a reason its off by default, because it doesnt work properly.

RBG
11th February 2012, 16:34
If you activate CUVID acceleration for XVID, you do so at your own risk. If it doesn't work, turn it off. :p
There is a reason its off by default, because it doesnt work properly.

Actually on forceware 285.62 I haven't encountered any problems yet, everything plays fine. Can you upload any DivX/Xvid sample which can not be played, I wanna test it.

zerowalker
11th February 2012, 18:08
Can you fix the Splitter to prevent it from crashing Recycle bin and COM Surrogate?
Or is it something i myself can fix:)?

Sorry asked this before;O

nevcairiel
11th February 2012, 18:15
You have to find out which file it crashes on, and then send me a sample of such a file.

zerowalker
11th February 2012, 19:24
It´s pretty much every file i think, as long as i try to make a thumbnail of it;S

It may also be MKV files which i Direct Stream copied so they became "AVI" that mess up.

But it´s the thumbnail thing i think, and if i deactivate AVI splitting on LAV it´s solved, though i want to use it;S

dead_screem
11th February 2012, 20:03
I have a hard-limit of i think 125 fps, if the header reported value is above that i don't trust it and use the (somewhat inaccurate) measured value (417225 isn't exact)

This is also the reason why i rather use the timestamps instead of basing any calculations on the fps. :)

Hmmm, does the fact that this file is VFR further make calculations difficult? I notice lav sets 23.968fps should anything be added that would round the calculated fps to a more standard value such as 23.976 in this case?

Also does the fact that this was a mov>mp4 re-mux have any thing to do with that the file is broken?

nevcairiel
11th February 2012, 20:11
Hmmm, does the fact that this file is VFR further make calculations difficult? I notice lav sets 23.968fps should anything be added that would round the calculated fps to a more standard value such as 23.976 in this case?


Well as i've stated numerous times before, the fps in the mediatype is just a hint, its not to be taken as absolute truth.
The MSDN even says that for VFR is perfectly valid for this value to be zero, however that seems to not work with alot of decoders/renderers.

In any case, i try to supply a somewhat valid value, but if the header information are totally screwed up, i just give it what ffmpeg measured from the first few frames, which may or may not be very accurate.

How the actual frames are timed is entirely up to the timestamps, not the fps value - at least if the decoder is doing a proper job, which is why the fps aren't all that important.
Just calculating from fps would most likely not work properly with VFR anyway, not to mention cases of missing frames like with Live TV (or recordings of that)

PS:
The header info inside the H264 data seems to be broken, don't think a remux would cause that, but i guess everything is possible.

entrecour
11th February 2012, 20:44
nevcariel, First of all Thank You for LAV!

I think I have found a problem with LAV Video on certain titles.

I get video corruption (pixelation? incomplete decoding?) occasionally on X-Men BD's (I have seen the problem on 2 or 3 different titles).

Here is a screenshot (http://img194.imageshack.us/img194/9316/lavp1.jpg)

I do not get this problem with Microsoft Dtv-Dvd decoder.

This is on Windows 7 32bit, Intel i3 (no external video card, no hardware acceleration).

General
Complete name : C:\Users\Public\Rip\XMEN_WOLVERINE\BDMV\PLAYLIST\00010.mpls
Format : Blu-ray Playlist
File size : 786 Bytes
Duration : 1h 47mn
Overall bit rate mode : Variable
Overall bit rate : 0

Video
ID : 4113 (0x1011)
Menu ID : 1 (0x1)
Format : AVC
Format/Info : Advanced Video Codec
Format profile : High@L4.1
Format settings, CABAC : Yes
Format settings, ReFrames : 4 frames
Codec ID : 27
Duration : 1h 47mn
Bit rate mode : Variable
Maximum bit rate : 26.7 Mbps
Width : 1 920 pixels
Height : 1 080 pixels
Display aspect ratio : 16:9
Frame rate : 23.976 fps
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Progressive
Color primaries : BT.709-5, BT.1361, IEC 61966-2-4, SMPTE RP177
Transfer characteristics : BT.709-5, BT.1361
Matrix coefficients : BT.709-5, BT.1361, IEC 61966-2-4 709, SMPTE RP177
Source : 00074.m2ts

I have the LAV Splitter pin info if that helps.

I can upload a clip if needed (how long? what tool do you recommend to extract?).

Let me know if you need anything else.

:thanks:

clsid
11th February 2012, 23:52
Writing to HKCR\Media Type\{E436EB83-524F-11CE-9F53-0020AF0BA770} fails here for the AVI and mpeg-ps subkeys. The other keys are fine.

nevcairiel
11th February 2012, 23:53
Probably protected by Windows, nothing i really want to override automatically. Even with UAC elevation those are hard to change.
Does the installer actually complain? If so, do you know how to shut it up? :p

For MPEG-PS, the installer is basically only writing the default windows values, so if its still protected, it should be fine.
For AVI, it writes the default windows values plus 2 extra values for more files. But overall, it should also not hurt if the write fails - as long as it doesnt complain to the user.

clsid
12th February 2012, 00:14
There is no user visible error. The functions will simply return false. It is just something I noticed and found strange since it used to work fine afaik when it was done by the filter. Error code is 5, which is access denied, so it indeed seems to be some kind of protection 'feature' of Windows.

nevcairiel
12th February 2012, 00:27
Those keys are owned by "TrustedInstaller", which i assume only signed .msi installers are allowed to access.
Anyhow, no issues if those cant be written. I doubt the filter really managed to do it.

Midzuki
12th February 2012, 00:35
Those keys are owned by "TrustedInstaller",
<OT>

Micro$oft and their "infinite wisdom". :rolleyes:
I'm pretty sure that a worm can easily dress itself as a "Trusted Installer" :p

</OT>

dukey
12th February 2012, 00:38
I got a weird problem that is showing up.
The text pin on the LAV splitter seems to be connecting to a filter called
Internal Script Command Renderer

Unfortunately this seems to totally break playing of the video. LAV is spitting out MEDIATYPE_Text, MEDIASUBTYPE_NULL
and this seems to be what this filter accepts.

http://i.imgur.com/r3W8g.png

Edit:
I forgot to mention this filter is part of directshow (quartz.dll)
http://msdn.microsoft.com/en-us/library/windows/desktop/dd390349%28v=vs.85%29.aspx

nevcairiel
12th February 2012, 00:45
I got a weird problem that is showing up.
The text pin on the LAV splitter seems to be connecting to a filter called
Internal Script Command Renderer

Unfortunately this seems to totally break playing of the video.

This is already fixed for the next version, it'll no longer cause playback to fail.
I suppose i could also change the media type and avoid it connecting completely, its not used for any active subtitle type, but there was some reason why i used this one before....

clsid
12th February 2012, 02:03
It´s pretty much every file i think, as long as i try to make a thumbnail of it;S

It may also be MKV files which i Direct Stream copied so they became "AVI" that mess up.

But it´s the thumbnail thing i think, and if i deactivate AVI splitting on LAV it´s solved, though i want to use it;S
Follow these steps:
1) Temporarily disable thumbnailing. Reboot.
2) Move one file that you suspect might be triggering this issue into a separate folder. That folder should not contain any video files, nor should any of its parent/child folders.
3) Re-enable thumbnailing. Reboot.
4) Load Explorer and browse to the folder.
5) If it shows the error you have a sample, otherwise repeat with a different file.

Once you have a sample, upload it to:
http://www.mediafire.com

If a file is very large, you can cut a small part it with DGSplit.

Mangix
12th February 2012, 04:59
is it possible to do bitstreaming when possible and fall back to software decoding when not a la ac3filter?

nevcairiel
12th February 2012, 12:50
LAV Filters 0.46

LAV Splitter
- Experimental support for streaming protocols (RTSP, RTP, MMS)
- Support for IBufferInfo (Buffer Status in MPC-HC "Statistics" pane)
- Added a memory limit to avoid too high memory use from the frame queues
- Fixed playback of files with Subtitles in WMP/WMC in some cases

LAV Audio
- Improved playback of files with bad audio frame interleaving
- Improved support of COOK, ATRAC and SIPR (RealAudio codecs)

LAV Video
- Updated and improved Intel QuickSync decoder (0.26, r38)
- Fixed DXVA2 software fallback on some lossless H264 clips
- Improved Fraps decoding with EVR
- Improved playback of MPEG4 in MP4
- Multi-threaded decoding for Lagarith
- Fixed playback of Theora with the MPC-HC Ogg Splitter
- Dithering Enhancements:
- New "Random Dithering" option (new default)
- Improved "Ordered Dithering" in the RGB converter
- Support for decoding v210/v410 raw formats
- Support for Snow and FFV1


Download: Installer (both x86/x64) (http://files.1f0.de/lavf/LAVFilters-0.46.exe) -- Zips: 32-bit (http://files.1f0.de/lavf/LAVFilters-0.46.zip) & 64-bit (http://files.1f0.de/lavf/LAVFilters-0.46-x64.zip)

Re-written Installer
I finally got around to cleaning up the Installer and basically re-creating it from scratch.
As a consequence, it now does all the special tasks required to properly register a Splitter in the system, and the LAVSplitter.ax itself does not do it anymore.

This means that just registering LAVSplitter.ax will no longer change any registry settings and is perfectly safe to do. The only thing it still does is allow playback of Blu-ray bdmv/mpls files, but in a entirely non-destructive way.
This also means that just registering LAVSplitter.ax may not allow you to play back all files, at least not in all players. MPC-HC is probably fine (if you add it to preferred filters), other players may not be. I recommend using the Installer! (unless you really know what you're doing)

I may also separate the Installer into 32-bit and 64-bit in the future, i have not decided on that yet (this would also mean that the settings get separated between the two versions)

Support for Streaming Protocols
I added experimental support for a few streaming protocols, specifically RTSP, RTP and MMS.
I tested the feature on alot of streams i could find around the web, as well as against VLC streaming over my local network, and it was working OK'ish (better for some formats, worse for others). I still consider this feature highly experimental, so try it if you wish, but don't expect too much. ;)

To use it, you need to use the Installer (or resort to manual registry hackery), even with MPC-HC.
I have not tested it in any other players.

To use it, just install it and open a rtsp://, rtp:// or mms:// URL.

Random Dithering
In the past, the complaint occasionally came up that ordered dithering has a too static pattern, which makes it visible in certain situations. This is especially bad with cheap LCD displays (6-bit TN panels), which typically also use ordered dithering - and doing ordered twice results in serious artifacting.

To avoid these issues, i implemented random dithering, which instead of a static pattern uses a randomly generated buffer to dither the pixels, resulting in no visible patterns (unless the RNG is having fun with you)
I decided to make this the new default mode for dithering. Performance should be the same as ordered.

Other changes
Everything else is pretty standard changes, a bit improved format support, some new formats, bugfixes across the board.

PS:
v210 decoding is off by default because its a raw format, and needs to be turned on.

Upcoming is a big ffmpeg update. I haven't updated for a while now because i didn't want to "inherit" any new bugs before the release, so a lot of changes have already piled up, hoping nothing breaks.

As always, please report regressions as detailed as you can, with sample files or other relevant proof.

Have fun!

Edit:
The files have been updated for a minor options-related crash.
If you get a crash when closing the Audio configurations, download again!

sneaker_ger
12th February 2012, 13:02
Thx.

Why are newer versions of LAV Filters asking me to restart my PC after installing and de-installing?

nevcairiel
12th February 2012, 13:03
Why are newer versions of LAV Filters asking me to restart my PC after installing and de-installing?

It'll only ask that if it couldn't overwrite or delete some files, indicating they are still open by something.

My money is on the explorer thumbnail generator. It seems to get stuck sometimes - especially if you have "DirectShowSpy.dll" registered - without it, it seems fine.

sneaker_ger
12th February 2012, 13:07
I see.

What is the current status of raw video files splitting? I see that it is deactivated by default and activating makes it attempt to open raw h.264 files, but it will just stop at the start with a black screen.

Sebastiii
12th February 2012, 13:18
Nice one Nevcairiel ;)
Happy :)

fastplayer
12th February 2012, 13:36
As always: :thanks:

wanezhiling
12th February 2012, 13:42
I'm late..:(:(

fairchild
12th February 2012, 13:49
I installed the new version 0.46 and I'm noticing a strange behavior. I noticed it after the first install when opening up all the Lav Filters configuration pages (Splitter, Video, Audio) and going through the different options etc to see what had changed and if I needed to change anything. Well after doing all this and hitting OK on the last page (in my case it is Lav Audio, I get a windows error:

Windows host process (Rundll32)
Windows host process (Rundll32) has stopped working

A problem caused the program to stop working correctly. Please close the program.

I was able the reproduce this again by going to the Start Menu and opening every Configuration page for all 3 (Splitter, Video, Audio) and it happened again after closing the last option. I am able to reproduce this randomly by opening in this order or in different orders, and switching tabs between the configurations, then hitting OK on each window without making any changes. It doesn't seem to occur if I simply open the configuration windows and don't do anything then hit OK or Cancel. It only happens when I switch to the formats tabs or hover over some of the options and get it to become a bit busy. I can see the little windows hourglass going for a bit which is new since I have a relatively fast computer and Lav Filters usually didn't do this and was very fast at loading it's configuration tabs.

I also just tested this through MPC-HC by double clicking each configuration option under External Filters and it made MPC-HC disappear. (albeit through MPC-HC you can only have 1 configuration tab open at a time). I dunno if this is affecting other users but I thought I'd report it none-theless. Perhaps the install changes you made had some change on the rundll32 windows file which is being used to call up the configuration options. Haven't tested yet if playback is affected. If this is only happening to me then perhaps my system is not as stable as I believed, but it's odd because my system should be 100% stable as I do folding@home 24/7 where my CPU is constantly at 100% load and also have no other stability problems with gaming or encoding or anything else on my PC.

cremor
12th February 2012, 13:53
You may have installed something else that reset the extension mapping. In any case, the next version has a basically completely rewritten installer, so if there were bugs, they are no more (but i'm sure new ones instead!)

Whatever it was, after using the new installer it is fixed. Thanks!

SamuriHL
12th February 2012, 14:03
Nev, is it recommended to clean out the old reg settings before installing this one?

Sent from my Xoom using Tapatalk

entrecour
12th February 2012, 14:04
I have the same problem as fairchild above - with LAV Audio Configuration. I uninstalled rebooted and reinstalled - same problem. When I retry LAV Audio Config I get the error message again. I am running Windows Visa Ultimate 32bit.

nevcairiel
12th February 2012, 14:14
Hitting OK on the last page (in my case it is Lav Audio, I get a windows error

I have the same problem as fairchild above - with LAV Audio Configuration. I uninstalled rebooted and reinstalled - same problem. When I retry LAV Audio Config I get the error message again. I am running Windows Visa Ultimate 32bit.

Should be fixed again. This has no impact on playback, and even saving options should work fine. Also, MPC-HC didn't seem to suffer here, just the standalone config.
Thats what you get when you try to be smart and do some last-minute changes =p

I'm uploading a fixed build, taking the liberty to just replace 0.46 and not changing the version, its been only 1.5 hours or so.... :)

nevcairiel
12th February 2012, 14:26
Ok, as a reminder for everyone: 0.46 had a small bug in the audio configuration sheet that caused a crash when closing it.
I fixed the bug and replaced 0.46, so if you downloaded 0.46 before this post, i urge you to re-download it. I didn't feel like bumping the version just an hour later for such a small fix.

nevcairiel
12th February 2012, 14:27
Nev, is it recommended to clean out the old reg settings before installing this one?
It never hurts, but its not necessary.