LAV Filters - DirectShow Media Splitter and Decoders

[Alexey1975]

Strange, I've never got a problem with CUVID on H264 with my GTX-570 : it plays everything even heavy H264 with ref16.
The only limitation is very high bitrate, but that's because the VP4 is too slow and not a limitation from CUVID.

So, and I have not had any problems with my GTX560 Ti, before something was changed in the nVidia ForceWare drivers.
nVidia ForceWare 285.62 driver and subsequent versions cause the aforementioned problem. That is, if you use earlier ForceWare (eg 285.38 or previous) – there is no any problem! Everything is OK!

In general, I had some experimenting and got the conclusion:
Thus, the above problem occurs only when watching HDTV with the DVB-card (no matter which software you are using). Play back the HD-content from HDD or etc. – cause no any problem (no matter which video driver is installed).
And that's what I think about it: LAV Video Codec, CoreAVC, etc. is not the cause of the issue.

The fact is (I think) that the new nVidia drivers has increased priority of using system bus resources for graphic card or the requested minimum of system bus bandwidth was increased for graphics. Because of this, with new video card drivers, there is a shortage the capacity of system bus bandwidth for the DVB-card stream. Perhaps there is a conflict video card driver and DVB-card (Omicom S2 rev.3) driver. As a consequence, there is a problem with HD-channels. While SD-channels continued to work more or less fine – for it there is the sufficient resources.
By the way, I'm not the only one in this situation with such a problem.

Anyway, the problem is not in the LAV Video Codec (rather a matter of nVidia driver programmers) – therefore the question is closed for this thread.
Thank you all for your attention.

[DragonQ]

Hmm that is strange. My HTPC has a GT 430 with the latest drivers and plays back HDTV fine using my BGT3630 (since Nev released that post-0.48 fix for MBAFF issues).

[marsovac]

I tried and managed to implement the necessary interfaces and use the filters in C#.

However there are small issues with the Filters themselves.

- On blurays the video is choppy (while CPU usage is not high)
- The IMediaSeeking does not support negative rate (playing backwards)

The first is maybe in progress as it is new, but the second is something that should be supported to enable fast forward and backward in players. As it is now without negative rate, there is no much need for a positive rate different than 1 (fast forward without backward does not make sense).
I'll attach a screenshot of what I mean - why positive rate change doesn't help without negative.

http://i1095.photobucket.com/albums/i468/marsinjo/player.jpg

However very nice project you have!
Keep up the good work!

Edit: Adding interfaces for C#, complete with FilterProvider which loads the COM through the ICLassFactory of the *.ax (without registration). Remove the '.txt' extension to make it a c# file.

[wanezhiling]

:) Did anyone find cuvid decoding not available on win8.

[DragonQ]

I was gonna try Windows 8 Server tonight but in a virtual machine. I doubt CUVID will work in a virtual machine regardless of the guest OS.

[dukey]

I tried and managed to implement the necessary interfaces and use the filters in C#.

However there are small issues with the Filters themselves.

- On blurays the video is choppy (while CPU usage is not high)
- The IMediaSeeking does not support negative rate (playing backwards)

The first is maybe in progress as it is new, but the second is something that should be supported to enable fast forward and backward in players. As it is now without negative rate, there is no much need for a positive rate different than 1 (fast forward without backward does not make sense).
I'll attach a screenshot of what I mean - why positive rate change doesn't help without negative.

http://i1095.photobucket.com/albums/i468/marsinjo/player.jpg

However very nice project you have!
Keep up the good work!

Edit: Adding interfaces for C#, complete with FilterProvider which loads the COM through the ICLassFactory of the *.ax (without registration). Remove the '.txt' extension to make it a c# file.

you'll find very few decoders will support a negative rate. You can play backwards on mpeg2/dvd under Vista/7 but not XP.

[nevcairiel]

Reverse playback of files is really not possible, because the codecs are not meant to decode backwards. The DVD Navigator implements this by decoding parts of the file in forward, and then playing the images in reverse, but thats not a concept a general purpose decoder can adopt. So no, will not happen.

Blu-rays also play just fine for me. "xxx playback is choppy" really doesn't contain any useful information though, you should specify with which source filter, which player, which decoder in LAV is active, which renderer, OS, hardware, which codecs in the file, etc...

[italospain]

Reverse playback of files is really not possible, because the codecs are not meant to decode backwards. The DVD Navigator implements this by decoding parts of the file in forward, and then playing the images in reverse, but thats not a concept a general purpose decoder can adopt. So no, will not happen.


would it be possible though the renderer (madVR) ?

i am missing this feature too.

it would be so cool.

[marsovac]

you'll find very few decoders will support a negative rate. You can play backwards on mpeg2/dvd under Vista/7 but not XP.

That's too bad...

Is there some way to simulate that that would not be ugly and error prone?

I am targeting onyl vista+ because i have a custom EVR presenter.

And the player uses only LAV with a custom built graph SplitterSource + Video + Audio + EVR.
(I'm not registering the COMs (using PInvoke to get IClassFactory and create the filters, so I'm not able to give you a screenshot of the graph in graphstudio, but believe it is only LAV since I'm programmatically connecting the Pins)

Edit:
Today It is working a bit better (seems almost fluent), so maybe it has something to do with the kernel/driver load.

More info about decoder etc...

About the video that goes a little faster a little slower (just a bit choppy, not very fluent):

General
ID : 0 (0x0)
Complete name : F:\BDMV\STREAM\00000.m2ts
Format : BDAV
Format/Info : Blu-ray Video
File size : 11.3 GiB
Duration : 42mn 45s
Overall bit rate : 37.9 Mbps
Maximum Overall bit rate : 48.0 Mbps

Video
ID : 4113 (0x1011)
Menu ID : 1 (0x1)
Format : MPEG Video
Format version : Version 2
Format profile : Main@High
Format settings, BVOP : Yes
Format settings, Matrix : Default
Format settings, GOP : M=3, N=12
Codec ID : 2
Duration : 42mn 34s
Bit rate mode : Variable
Bit rate : 32.0 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
Compression mode : Lossy
Bits/(Pixel*Frame) : 0.644
Stream size : 9.96 GiB (88%)

Audio #1
ID : 4352 (0x1100)
Menu ID : 1 (0x1)
Format : AC-3
Format/Info : Audio Coding 3
Mode extension : CM (complete main)
Codec ID : 129
Duration : 42mn 44s
Bit rate mode : Constant
Bit rate : 640 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 : 196 MiB (2%)

Audio #2
ID : 4353 (0x1101)
Menu ID : 1 (0x1)
Format : DTS
Format/Info : Digital Theater Systems
Format profile : MA / Core
Muxing mode : Stream extension
Codec ID : 134
Duration : 42mn 43s
Bit rate mode : Variable
Bit rate : 1 576 Kbps / 1 510 Kbps
Channel(s) : 6 channels
Channel positions : Front: L C R, Side: L R, LFE
Sampling rate : 96.0 KHz / 48.0 KHz
Bit depth : 24 bits
Compression mode : Lossless / Lossy

Audio #3
ID : 4354 (0x1102)
Menu ID : 1 (0x1)
Format : AC-3
Format/Info : Audio Coding 3
Mode extension : CM (complete main)
Codec ID : 129
Duration : 42mn 44s
Bit rate mode : Constant
Bit rate : 640 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 : 196 MiB (2%)

Text #1
ID : 4608 (0x1200)
Menu ID : 1 (0x1)
Format : PGS
Codec ID : 144

Text #2
ID : 4609 (0x1201)
Menu ID : 1 (0x1)
Format : PGS
Codec ID : 144

Text #3
ID : 4610 (0x1202)
Menu ID : 1 (0x1)
Format : PGS
Codec ID : 144

[mastan]

How do I use the QuickSync hardware decoder?

I have a Z68 motherboard, my CPU is an i5-2500k, my discrete GPU is an AMD Radeon HD 6970, I'm running LAV Filters 0.48 with MPC-HC and madVR on Windows 7 x64. I have the newest Intel HD graphics driver installed and the Lucid Logix Virtu control panel (do I need this thing?) along with the newest CCC drivers. When I try to select QuickSync in LAV Video, it still says "not available." Am I missing something?

I don't really need it, I was just curious to see how well it performs.

Which output your monitor is plugged to? If AMD's, force video player to use Intel in Virtu cp.
Also you should use latest graphics drivers from Intel site, not from windows update or mb manufacturer.

[CruNcher]

would it be possible though the renderer (madVR) ?

i am missing this feature too.

it would be so cool.


as nev said it's possible todo clever workarounds like ISVs use them in their frameworks but cool isn't usable for everyone and it takes time to implement such things which could be efficiently spend on other stuff like stability and support which are the key areas where no ISV can holdup with currently, though the source is open for a reason ;)

[marsovac]

as nev said it's possible todo clever workarounds like ISVs use them in their frameworks but cool isn't usable for everyone and it takes time to implement such things which could be efficiently spend on other stuff like stability and support which are the key areas where no ISV can holdup with currently, though the source is open for a reason ;)

Do you have tips on how that can be done. What to read and where to find general information to teach myself. I'm not deep into this stuff, so I don't know where to start. I have knowledge of C++ and COM programming but lack the specific knowledge.

[Reith]

Which output your monitor is plugged to? If AMD's, force video player to use Intel in Virtu cp.
Also you should use latest graphics drivers from Intel site, not from windows update or mb manufacturer.
Yeah, my monitor is plugged into the AMD card.

How do I force the video player to use the Intel one? All I see in the Virtu control panel is an applications list. No way to do anything special with them after I add them. madVR said it failed to create a Direct3D device when I added MPC-HC to it.

Also, I did get the Intel drivers directly from their site.

[mastan]

Yeah, my monitor is plugged into the AMD card.

How do I force the video player to use the Intel one? All I see in the Virtu control panel is an applications list. No way to do anything special with them after I add them. madVR said it failed to create a Direct3D device when I added MPC-HC to it.
Try some quick sync transcoding application instead to see if QS works at all, add it to application list in Virtu cp too.

It is possible that you can transcode videos using QS, but can't use it in player, since the latter one needs to output video to another video card.

[6233638]

I can't seem to seek some .caf files when opened with LAV Splitter:

Sample (http://www.mediafire.com/?de0jga23yk04qez)

[SamuriHL]

Hey, Nev, your code builds and runs on w8 cp quite nicely. :) I had to clean install my laptop which meant rebuilding my dev env. No problems at all.

Sent from my Xoom using Tapatalk

[Xaurus]

nevcairiel or anyone else,

Why is lav splitter reporting a frame rate of 24.000 to madvr (CTRL+J) when mediainfo says it's 25.000 for this file?
Sample (http://home.halden.net/mordor/test_fps.zip).

[nevcairiel]

nevcairiel or anyone else,

Why is lav splitter reporting a frame rate of 24.000 to madvr (CTRL+J) when mediainfo says it's 25.000 for this file?
Sample (http://home.halden.net/mordor/test_fps.zip).

Thats quite simple. The video stream thinks its 24 fps, and the matroska info claims its 25 fps - with one problem, Matroska doesn't actually contain a "FPS" info, its just a guess based on some other informations. A guess LAV does not trust, so it uses the information the video stream provides.
We've been over this before, thats just the way it works, and will remain to work.

[Xaurus]

Thats quite simple. The video stream thinks its 24 fps, and the matroska info claims its 25 fps - with one problem, Matroska doesn't actually contain a "FPS" info, its just a guess based on some other informations. A guess LAV does not trust, so it uses the information the video stream provides.
We've been over this before, thats just the way it works, and will remain to work.
Okay, so you are saying that LAV is correct and that the stream is actually 24 fps.

This clip is unwatchable with my 23.976xx refresh rate setup (lots of dropped frames), but it's definitely watchable in 50.000xx fps. And with Nvidia I can't create a pure 24.000 resolution because it simply says that the resolution already exists (my 23.976 resolution).

Thanks.

[nevcairiel]

Okay, so you are saying that LAV is correct and that the stream is actually 24 fps.

No, the stream is 25 fps.
All i'm saying is that LAV is behaving correctly, the file is just not 100% correct.

[wanezhiling]

http://www.gokuai.com/f/wJ76h8iO3VHH0Bd6
nev, LAV Splitter(Source) failed.

[nevcairiel]

http://www.gokuai.com/f/wJ76h8iO3VHH0Bd6
nev, LAV Splitter(Source) failed.

The file seems broken.
When i open it with Haali, it outputs 0 streams

[pururin]

May I ask something regarding audio output, is converting between different audio bitdepths a lossy conversion at any case?

when up-convert, like 16->24 bit, is it some complex math recalculation (lossy)
or just adding zero bits, which is a lossless operation, like reclock does (it does says "bit exact" when 16->24 whilst 24->16 doesn't, am I correct that this means lossless?).

I don't know about converting to (32 bit)floating point though.

OTOH when down-convert, is it doing any dithering(+noise shaping maybe) or just simple rounding / truncating? (lossy anyway but differ in quality)

. . .
Is it a good idea to disable 16-bit int output, or even force 32-bit fp only, when connecting to ffdshow to do few post-processing like regain volume and mixer?
So that maybe it can help ffdshow to produce more accurate rounding by processing at higher bitdepth level, especially if it up-convert losslessy. ->> Is this worth doing?

Or is this totally irrelevent as ffdshow will up-converts by itself automatically when feels the need? :o
(a little question, is ffdshow converting quality about the same as LAV? They doing it quite the same way?)


BTW, this has been bugged me for a while, does swap channel filter doing any processing to the soundwave?
From what I understand it should be just remapping channels hence no need any pre-convert aforementioned.

Any advice/speculation is appreciated. Thanks!

[wanezhiling]

The file seems broken.
When i open it with Haali, it outputs 0 streams
Thanks for reply.

Did you try MPC-HC, it seemed ok.:)

[nevcairiel]

May I ask something regarding audio output, is converting between different audio bitdepths a lossy conversion at any case?

when up-convert, like 16->24 bit, is it some complex math recalculation (lossy)
or just adding zero bits, which is a lossless operation, like reclock does (it does says "bit exact" when 16->24 whilst 24->16 doesn't, am I correct that this means lossless?).

I don't know about converting to (32 bit)floating point though.

OTOH when down-convert, is it doing any dithering(+noise shaping maybe) or just simple rounding / truncating? (lossy anyway but differ in quality)



16->24 is just adding zeros.
24->16 in LAV is just a truncation. Its not meant to be high quality, its just a fallback operation in case its the only way you can get audio.

Anyway, i would suggest to just leave all formats activated in LAV, that way you avoid any unnecessary conversions. If ffdshow needs some converion, it should do it on its own.

[oddball]

I have discovered a bug in LAV Audio decoder when decoding DTS that is encoded at 44.1Khz instead of 48Khz. I have a few audio DTS files that are like that and I mux them into MKV containers so I can play them back on my receiver. Now with LAV Audio filter it sends it out as bitstream audio and the light goes blue on my receiver indicating it is decoding DTS. However LAV seems to be upsampling it somehow to 48Khz. This results in everything being speeded up (squeaky voices etc). If I use ffdshow's inbuilt DTS passthru the receiever lights up blue and plays it at the correct pitch. Any idea why LAV Audio filter is speeding it up? I can only assume it's not actually bitstreaming but reencoding the audio 'on-the-fly' as surely this would not be possible if it was just passing through the bitstream signal?

[pururin]

16->24 is just adding zeros.
24->16 in LAV is just a truncation. Its not meant to be high quality, its just a fallback operation in case its the only way you can get audio.

Anyway, i would suggest to just leave all formats activated in LAV, that way you avoid any unnecessary conversions. If ffdshow needs some converion, it should do it on its own.Thanks for the fast reply!

About this point. So 16->32 is adding (16 more) zeroes too? Am I correct this method is lossless?
I agree with the suggestion, though I'm not so sure whether ffdshow is trustable at these stuffs. :p

[nevcairiel]

Thanks for reply.

Did you try MPC-HC, it seemed ok.:)

I made it ignore the problem in the file now, so the file may or may not work, its really just luck! :)

I have discovered a bug in LAV Audio decoder when decoding DTS that is encoded at 44.1Khz instead of 48Khz. I have a few audio DTS files that are like that and I mux them into MKV containers so I can play them back on my receiver. Now with LAV Audio filter it sends it out as bitstream audio and the light goes blue on my receiver indicating it is decoding DTS. However LAV seems to be upsampling it somehow to 48Khz. This results in everything being speeded up (squeaky voices etc). If I use ffdshow's inbuilt DTS passthru the receiever lights up blue and plays it at the correct pitch. Any idea why LAV Audio filter is speeding it up? I can only assume it's not actually bitstreaming but reencoding the audio 'on-the-fly' as surely this would not be possible if it was just passing through the bitstream signal?

Its not possible to reencode the audio on the fly because there is no DTS encoder.
Also, what makes you think i would ever code something to do that? o.O

You can do two things.
1) Try with this version: http://files.1f0.de/lavf/LAVFilters-0.48-14-g9ce1413.zip
2) If 1 doesn't help, give me a sample of such a file.

[oddball]

That version just causes MPC-HC to crash. I've reinstalled .48 for now which works apart from the audio speedup issue. I will try and upload a test file for you to see. You need to have a DTS receiver though to hear this happening. Switching between LAV Audio and ffdshows DTS passthru should reveal it hopefully to you. If it plays at the same speed on both of those for you I really don't know what the issue is. The receiver lights up blue for DTS but one is higher pitched than the other which you would think is impossible yet it is happening here.

Pinout info just in case.
Input.
Filter : LAV Audio Decoder - CLSID : {E8E73B6B-4CB3-44A4-BE99-4F7BCB96E491}

- Connected to:

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

- Connection media type:

Audio: DTS 44100Hz 6ch 1411kbps

AM_MEDIA_TYPE:
majortype: MEDIATYPE_Audio {73647561-0000-0010-8000-00AA00389B71}
subtype: Unknown GUID Name {00002001-0000-0010-8000-00AA00389B71}
formattype: FORMAT_WaveFormatEx {05589F81-C356-11CE-BF01-00AA0055595A}
bFixedSizeSamples: 1
bTemporalCompression: 0
lSampleSize: 256000
cbFormat: 18

WAVEFORMATEX:
wFormatTag: 0x2001
nChannels: 6
nSamplesPerSec: 44100
nAvgBytesPerSec: 176400
nBlockAlign: 12
wBitsPerSample: 16
cbSize: 0 (extra bytes)

pbFormat:
0000: 01 20 06 00 44 ac 00 00 10 b1 02 00 0c 00 10 00 . ..D¬...±......
0010: 00 00 ..


Output.
Filter : LAV Audio Decoder - CLSID : {E8E73B6B-4CB3-44A4-BE99-4F7BCB96E491}

- Connected to:

CLSID: {18C16B08-6497-420E-AD14-22D21C2CEAB7}
Filter: Audio Switcher
Pin: Audio

- Connection media type:

Audio: 0x0092 48000Hz stereo 1536kbps

AM_MEDIA_TYPE:
majortype: MEDIATYPE_Audio {73647561-0000-0010-8000-00AA00389B71}
subtype: MEDIASUBTYPE_PCM {00000001-0000-0010-8000-00AA00389B71}
formattype: FORMAT_WaveFormatEx {05589F81-C356-11CE-BF01-00AA0055595A}
bFixedSizeSamples: 1
bTemporalCompression: 0
lSampleSize: 1
cbFormat: 18

WAVEFORMATEX:
wFormatTag: 0x0092
nChannels: 2
nSamplesPerSec: 48000
nAvgBytesPerSec: 192000
nBlockAlign: 4
wBitsPerSample: 16
cbSize: 0 (extra bytes)

pbFormat:
0000: 92 00 02 00 80 bb 00 00 00 ee 02 00 04 00 10 00 ’...€»...î......
0010: 00 00 ..

- Enumerated media type 0:

Set as the current media type



LAV Audio reports 44100 6 channel bitstreaming

Testfile.

http://www.mediafire.com/?2hc0vofdkmso2mp

[wanezhiling]

I made it ignore the problem in the file now, so the file may or may not work, its really just luck! :)



https://code.google.com/p/lavfilters/source/detail?r=be067e7ebf0e9aac7d1301fea0ada9e9e84f3972
:)

[oddball]

OK this is weird. If I turn off SPDIF bitstreaming and just let LAV Audio decode to 6 analog channels it plays at the correct pitch/speed. But if I tick the DTS bitstreaming so the receiver lights up blue it plays back speeded up. Using ffdshow DTS passthru it also lights up but plays correctly. Please note I have nothing in the MPC-HC chain like ReClock etc it's stripped down to bare LAV Audio and ffdshow audio as external filters. Everything else is pretty much defaults.

So what the heck is going on? I'm going to try my external soundcards SPDIF output and see if it plays any different. Strangely if I run SPDIF test it says my motherboards soundcard (Asus P5Q Deluxe with SoundMax ADI chipset) does not support DirectSound SPDIF. Even though it does work for most everything I have played through it.

Nope! Same issue on a totally different soundcard. I am now going to grab another player that supports LAV and try it just in case it's MPC doing something weird.

[nevcairiel]

That version just causes MPC-HC to crash. I've reinstalled .48 for now which works apart from the audio speedup issue. I will try and upload a test file for you to see. You need to have a DTS receiver though to hear this happening. Switching between LAV Audio and ffdshows DTS passthru should reveal it hopefully to you. If it plays at the same speed on both of those for you I really don't know what the issue is. The receiver lights up blue for DTS but one is higher pitched than the other which you would think is impossible yet it is happening here.

You know whats funny, if i try to bitstream that with ffdshow, i don't get any audio at all and LAV sounds fine.
Just from the definition of how bitstreaming works, it either works or it doesn't, you really don't communicate any extra information over the link that specify sample rate or whatnot.

The sample rate is encoded in the bitstream, in the DTS headers. It really doesn't make any sense to output anything else.
It somehow sounds like your receiver is doing something odd, and ffdshow somehow does the same oddity. Since i cannot get ffdshow to successfully bitstream on that sample for me at all, it surely isn't working 100% correctly either. ;)



So what the heck is going on? I'm going to try my external soundcards SPDIF output and see if it plays any different. Strangely if I run SPDIF test it says my motherboards soundcard (Asus P5Q Deluxe with SoundMax ADI chipset) does not support DirectSound SPDIF. Even though it does work for most everything I have played through it.
What have you been playing it through, if not your soundcards SPDIF? o.o

Anyway, the player has no influence on this (as long as it doesn't do something incredibly stupid). You can try switching audio renderers, in MPC-HC switch between DirectSound renderer and the "normal" renderer, maybe it makes some difference.

[oddball]

Have you checked you can get DTS passthru with ffdshow with normal movies? Anyhow I play it through SPDIF for DTS audio but I run multi-channel analog out for PAL speed down via ReClock. I can't understand why ffdshow plays DTS passthru for audio these files fine and LAV Audio doesn't. It's all a bit peculiar.

[VipZ]

Testfile.

http://www.mediafire.com/?2hc0vofdkmso2mp

Regarding this file, normally I use use DTS-HD framing for all my DTS bitstreaming so that I get DTS-ES etc. I have noticed bitstreaming fails when this option is enabled but works when disabled.

Not sure how this relates to your issue though.

Nev, maybe DTS-HD framing should never be used even if ticked with 44.1khz DTS?

[oddball]

That option has no effect on my issue just in case you are thinking it. Anyhow I am just going to continue using ffdshow to decode audio for now since it works for everything I throw at it. Plus it bypasses bitsreaming for AAC etc which allows me to use ReClock automagically for that type of content. I only posted this info in passing really to help you bugfix. But it might just be something odd with my setup.

[marsovac]

Hi guys. Since I'm unable to find people to test a player I'm working on I thought that somebody from the forum would be willing to give it a check.

I'm not seeking for SQA, just some basic functionality check on different machines and OSes.

download player x86 (zip 6.31MB) (http://www.megafileupload.com/en/file/342967/MediaPoint-x86-zip.html)

download player x64 (zip 6.61MB) (http://www.megafileupload.com/en/file/342968/MediaPoint-x64-zip.html)

The requirements should be:
- .NET 3.5 (probably SP1)
- CRT runtime (VCredist 2010) (will add in future versions)

So please let me know which configuration fail to play a video file or even open the player.

The player is based on LAV Filters and does nothing to your registry (no COM registration and uses LAV with SetRuntime(true)).

Please do not report bugs or performance issues. This is just a proof of concept with fast written code to get the functionality and a debug build.

I'm only interested if it plays as a portable player without installing any filters on the machine. - and especially in which configurations it doesn't.

Thanks!

[upv]

Does not work NVIDIA CUVID on Windows 8 CP

[Pat357]

The file seems broken.
When i open it with Haali, it outputs 0 streams
I know it looks broken because the small file size, but it is not broken...
It plays fine with MPC-HC build in splitter + LAV video + Lav-audio.
It seems LAV-splitter stalls at the start.

Dam.. I must read faster..:D

Anyway FFplay plays this file without displaying any errors.
Maybe the FFMpeg guys fixed this already and we have to wait, or they came up with the same solution as you did.:D

[SamuriHL]

Does not work NVIDIA CUVID on Windows 8 CP

nVidia will have to release a W8 CP driver. AMD already did. :D

[dansrfe]

Are there any mixer options in LAV Audio such as a prologic downmix?

[husmoaschta]

nVidia will have to release a W8 CP driver. AMD already did. :D

one week old:
http://www.geforce.com/Drivers/Results/41694

[SamuriHL]

one week old:
http://www.geforce.com/Drivers/Results/41694

Nice. And do they work with CUDA? I would assume they must.

[Mangix]

i was watching a .trp file (MPEG2 audio and video) and i got this error when seeking.

http://postimage.org/image/szchgtdzv/

[nevcairiel]

I know it looks broken because the small file size, but it is not broken...

Yes it is broken.
The fact that everyone else just ignores the problem in the file doesn't make it go away. :p

[e-t172]

About this point. So 16->32 is adding (16 more) zeroes too? Am I correct this method is lossless?

If you're talking about 32-bit IEEE float, converting to floating point is much more complicated than adding zeroes. However, if done correctly, 24 bit → 32 bit FP → 24 bit is lossless (bit-perfect). Of course most software will probably dither when converting back to integer, so it won't be bit-perfect, but the difference is absolutely negligible.

[husmoaschta]

Nice. And do they work with CUDA? I would assume they must.

No clue. Trying to get sorted out the last issues with AMD and Win7 for the moment. Win8 is planned for next life ...

[red5goahead]

Two questions:

1) Is there any possibility to have a simplified way to setup any parameter of a lav video, audio or splitter filter by code? I know there are interfaces for C in the documentation but fo example I would it with Delphi interfaces.

2) Is there any possibility to have a basic upscaling feature directly into the lav codec video?

:thanks:

[ranpha]

A question, if the VC-1 format is unchecked in the 'Format' tab of LAV Video Decoder, will it also disable the VC-1 hardware acceleration features in DXVA/CUVID mode?

[nevcairiel]

A question, if the VC-1 format is unchecked in the 'Format' tab of LAV Video Decoder, will it also disable the VC-1 hardware acceleration features in DXVA/CUVID mode?

If the format is unchecked, LAV will refuse any connection to a VC-1 movie.
I always wanted to try to build a "Only Hardware" kind of function, but its not as easy as it sounds.

[pururin]

If the format is unchecked, LAV will refuse any connection to a VC-1 movie.
I always wanted to try to build a "Only Hardware" kind of function, but its not as easy as it sounds.
I heard that microsoft decoder is still faster at VC-1.
But not in case of lav HW accelerated VC-1 decoding?