madVR - high quality video renderer (GPU assisted) [Archive] - Page 187

QBhd

17th August 2011, 23:17

I have the same problem when opening a lot of files in seamless playback mode, I thought it was due to the seamless parser which is quite buggy from what I was told on the official forum...or possibly ffdshow. If it's indeed due to mVR, it'd be great if you could provide samples to madshi so he could fix it :)

Samples for this are not needed. All you need is to have PotPlayer. Alternate SD XviD AVI's with HD x264 MKV's in a playlist. Use a decoder of your choice for the XviD's and use madVR as a decoder for the x264's. Also run madVR in Fullscreen Exclusive. When the x264 file finishes it's playback and moves on to an XviD file, madVR crashes.

QB

ForceX

17th August 2011, 23:47

That's very weird. Does it occur with every video files, or just with one? I've never seen anything like that yet. Not sure, it must either be specific to that one video file, or maybe it's a GPU driver bug? In any case, I need to be able to reproduce it, otherwise a fix will be difficult.
The video cutoff thing happens on all files, but it's just a driver bug in my laptop's Geforce. Doesn't happen with the integrated Intel one. Damn OEM drivers... RGB32 used to work before, but oh well. Not like I need it.

And it turned stuttery again.
The queues were something like this:...

Your graphics card doesn't seem to be powerful enough. Try using Bicubic resizers, and maybe increase the "present frames in advance" number.

kasper93

17th August 2011, 23:53

I don't even know what subresync is?

view->Subresync or crtl+6 in MPC-HC :) It's really usefull to apply little changes to subtitles without using any external apps or converting subtitle format. To see the problem you should run movie with frame rate other than 25.0FPS(default) and load subtitles in format like this one: {2677}{2778}line1
{2816}{2893}line2

BTW. Look at those screens good (http://dl.dropbox.com/u/16282309/MPC-HC/good.png), bad (http://dl.dropbox.com/u/16282309/MPC-HC/bad.png) (timestamps).

QBhd

18th August 2011, 00:11

As to my other problem with 720p60 x264 files... here is a small sample. and I have also included the debug log for a few minutes of playback with madVR on default settings.

720p60 Sample MKV (http://www25.zippyshare.com/v/29544471/file.html)

madVR - log (http://www25.zippyshare.com/v/78560075/file.html)

QB

BatKnight

18th August 2011, 00:47

@madshi, congratulations on your 5000 posts...
Keep up the great work. Your madVR is wonderful. I can watch an entire movie with 0 dropped frames, 0 delayed and 0 glitches up until the the end!

Thanks.

BatKnight

18th August 2011, 01:16

nand chan

18th August 2011, 01:57

Screenshots of the aforementioned banding with 4:4:4 video:

http://www.mediafire.com/?ksr2c70nn74frdr

The madVR only version isn't truly perfect, but there's no detectable /macro/banding, ie. no 30 pixel wide bands - all inaccuracies are limited to 1 pixel streaks here and there.

I understand now why you can't expose the pipeline using DirectShow. Next best option would be exposing a GPU pipeline API I suppose but that would require significant effort. I don't think it should be a priority.

cyberbeing

18th August 2011, 02:22

Hmmmm... I'll need to check what Reclock exactly looks for. Maybe I can talk to James (SlySoft) to find a solution for that. Shouldn't be too hard.

So you're saying CoreAVC isn't decoding, madVR is, but still Reclock picks up the frame rate? That sounds weird...

It turns out it works with FFDShow and probably other decoders as well when set to high merit (which you then need to override in MPC-HC). Reclock seems to open the video briefly with the default decoder. Since the madVR decoder is inaccessible, it uses the first thing available.

Edit: ReClock working also seems highly dependent on your splitter. If you install LAVFilters Splitter, for example, ReClock's DirectShow detection breaks, and you need to re-install Haali Media Splitter (or another working splitter) for it to work again. Nevcairiel if you read this, you may want to look into why this is happening with your splitter. (Edit2: see below)

Edit2: The problem with LAVSplitter appears to be that it doesn't register its source filter with the necessary file extensions under HKEY_CLASSES_ROOT\Media Type\Extensions and actually deletes whatever entries are there on installation, which is why ReClock breaks.

madshi, you'll still need to deal with the problem that ReClock can't access your decoder though. As it stands now, everybody needs a VC-1, MPEG2, and H264 decoder with a merit > 0x800000 and and a working splitter like Haali for ReClock DirectShow framerate detection to work, even if they are using madVR's VC-1/MPEG2/H264 decoders...

True. I could probably change that. However, it might cost performance, i'm not sure. How important is that? I mean we color correct video to make it look "correct". How can it matter for subtitles whether they're color corrected or not? Sounds weird to me.
It's important for color matched typesetting and vector shapes, since a 3DLUT correction applied to the video but not subs could result in two starkly different colors when they were meant to be identical. I would say this is important to implement at least as a toggle-able option for correctness.

I just tested the latest threaded VSfilter with Hi10P material,

It “converts” 10 bit material to 8 bit by rounding down. Causes insane, immediately noticeable banding.
As ForceX said, any converting from 10 bit to 8 bit happens before VSFilter gets it's hands on the video. The only way it would create more banding is if you were using its sub-optimal colorspace conversions to output a different color space.

Screenshots of the aforementioned banding with 4:4:4 video
That test video should probably be deleted... I realized later that I accidentally created banding when encoding since I forgot to specify the fullrange flag. It does show how good madVR is at smoothing out sources which contain banding though... :(

nand chan

18th August 2011, 02:43

As ForceX said, any converting from 10 bit to 8 bit happens before VSFilter gets it's hands on the video. The only way it would create more banding is if you were using its sub-optimal colorspace conversions to output a different color space.

Fair enough, we should probably get our hands on a decent 4:2:0 10-bit video for proper testing.

That test video should probably be deleted... I realized later that I accidentally created banding when encoding since I forgot to specify the fullrange flag. It does show how good madVR is at smoothing out sources which contain banding though... :(

That might explain it. Any chance you could re-encode it properly, using 4:2:0 Hi10P?

cyberbeing

18th August 2011, 02:54

I'm tempted to wait until a patch which fixes the x264 swscale levels bug lands, before making another attempt.

nand chan

18th August 2011, 03:32

I'm tempted to wait until a patch which fixes the x264 swscale levels bug lands, before making another attempt.

I just realized that the madTestPatternSource generator actually works quite well for this.

Any chance of updating it to output a 10 bit 4:2:0 signal?

liason

18th August 2011, 05:57

Problem: partially black picture (see http://img842.imageshack.us/img842/2446/madvrrgb32outputfromffd.th.jpg (http://imageshack.us/photo/my-images/842/madvrrgb32outputfromffd.jpg/)).
Conditions: when watching videos with ffdshow and in the Output tab only RGB32 is selected as supported colorspaces. Using MPC-HC 1.5.2.3456, madVR 0.73, ffdshow 3964.
Comments: this is observed with every video & only with madVR selected as renderer (with Haali renderer or VMR9 everything is fine).

e-t172

18th August 2011, 08:19

ReClock working also seems highly dependent on your splitter. If you install LAVFilters Splitter, for example, ReClock's DirectShow detection breaks, and you need to re-install Haali Media Splitter (or another working splitter) for it to work again. Nevcairiel if you read this, you may want to look into why this is happening with your splitter.

ReClock works perfectly fine with LAV Splitter on my system.

THEAST

18th August 2011, 08:45

Ah thanks, the autodetection fails there, need to check why. Do the colors look correct if you manually change the decoding matrix from BT.709 to BT.601?
How can I manually change the decoding matrix?

ForceX

18th August 2011, 08:53

How can I manually change the decoding matrix?
Ctrl+Alt+Shift+M

nevcairiel

18th August 2011, 09:12

It turns out it works with FFDShow and probably other decoders as well when set to high merit (which you then need to override in MPC-HC). Reclock seems to open the video briefly with the default decoder. Since the madVR decoder is inaccessible, it uses the first thing available.

Edit: ReClock working also seems highly dependent on your splitter. If you install LAVFilters Splitter, for example, ReClock's DirectShow detection breaks, and you need to re-install Haali Media Splitter (or another working splitter) for it to work again. Nevcairiel if you read this, you may want to look into why this is happening with your splitter. (Edit2: see below)

Edit2: The problem with LAVSplitter appears to be that it doesn't register its source filter with the necessary file extensions under HKEY_CLASSES_ROOT\Media Type\Extensions and actually deletes whatever entries are there on installation, which is why ReClock breaks.

madshi, you'll still need to deal with the problem that ReClock can't access your decoder though. As it stands now, everybody needs a VC-1, MPEG2, and H264 decoder with a merit > 0x800000 and and a working splitter like Haali for ReClock DirectShow framerate detection to work, even if they are using madVR's VC-1/MPEG2/H264 decoders...

Somehow i really doubt that ReClock depends on the way your filters are registered in the registry to figure out the FPS, rather then the filter graph itself.
Maybe ReClock has some fallback mode that it builds its own filtergraph of the file if it cannot detect it from the current one? Of course this would break if madVRs decoder has the highest merit. :p

Anyhow, i also have absolutely no issues with ReClocks FPS detection with LAV Splitter.

cyberbeing

18th August 2011, 11:01

Somehow i really doubt that ReClock depends on the way your filters are registered in the registry to figure out the FPS, rather then the filter graph itself.
Maybe ReClock has some fallback mode that it builds its own filtergraph of the file if it cannot detect it from the current one? Of course this would break if madVRs decoder has the highest merit. :p
Why do you doubt it? That is exactly what ReClockHelper.dll (ReClock's DirectShow frame rate detection) does. You can see this happen with Process Explorer in the lower pane (ReClockHelper.dll loads the default decoder and splitter) and Process Monitor (ReClockHelper.dll reads a small part of the file).

It's a bit hard to catch since it only runs for a split-second when you open a video, but it does appear to be building its own graph with the default decoder (by merit) and splitter source filter (if there is an entry under HKEY_CLASSES_ROOT\Media Type\Extensions for the file type).

madVR's merit is very low at 0x200000, so it will never get used by ReClock. This isn't a problem, since currently the madVR decoder can't be used by anything other than madVR...

The DirectShow estimator is instant framerate detection on video open.

The Built-in estimator is the fall-back solution when the DirectShow estimator fails. It is always watching the active graph (I assume), looking for frame-rate changes on things like VFR content or DVDs. It's completely broken with the madVR decoder though (which is a problem that madshi will need to deal with eventually...), but with other decoders it will eventually determine the current framerate.

Anyhow, i also have absolutely no issues with ReClocks FPS detection with LAV Splitter.

So it works for you when you uncheck "Use built-in estimator", leaving only "Use DirectShow" checked?

It doesn't for me, with any decoder when LAV Splitter is installed. Manually adding the correct registry entries fixes it though. This has the positive side-effect of LAV Splitter being used by default, even if Haali is installed.

All you need to do is add entries like the following, for every format LAV Splitter can act as a source filter for:
[HKEY_CLASSES_ROOT\Media Type\Extensions\.mkv]
"Source Filter"="{B98D13E7-55DB-4385-A33D-09FD1BA26338}"
If there is no protocol in the file name, the Filter Graph Manager looks in the registry for entries with the key HKEY_CLASSES_ROOT\Media Type\Extensions\.ext\, where .ext is the file extension. If this key exists, the value Source Filter contains the CLSID of the source filter, in string form.

nevcairiel

18th August 2011, 11:16

All you need to do is add entries like the following, for every format LAV Splitter can act as a source filter for:
[HKEY_CLASSES_ROOT\Media Type\Extensions\.mkv]
"Source Filter"="{B98D13E7-55DB-4385-A33D-09FD1BA26338}"

I know how those keys work, and if you use the installer for LAV Splitter, it will also write them for you (the install.bat will however not).
But its not the only way to open files - there is a second option to identify them by a bytemarker at the start of the file, and associate that to a source filter. Why ReClock does not use that, i don't know.

cyberbeing

18th August 2011, 11:24

I've always tested LAV Filters using the .bat files, so I guess that explains it. I'll make sure to use the installer the next time around. I assume ReClock uses only that method to shave a few milliseconds off the time it takes to build a filter graph, but it does seem like it should be falling back to the bytemarker method if it fails.

Pointing people to the LAV Filters installer if they have ReClock troubles, would be a good thing to add to your FAQ in the LAV Filters thread.

nand chan

18th August 2011, 12:22

I know how those keys work, and if you use the installer for LAV Splitter, it will also write them for you (the install.bat will however not).
But its not the only way to open files - there is a second option to identify them by a bytemarker at the start of the file, and associate that to a source filter. Why ReClock does not use that, i don't know.

Any reason the install.bat will not add them?

nlnl

18th August 2011, 13:18

madshi

Thanks for .73!

Could you tell, does embedded in Madvr Mpeg2, H264 decoders support correct inverse telecine transformation for PAL films (1080/576i50--> 1080/576p25)?

And do you have any plans for NTSC (1080/480i59-->1080/480p23) ?

nevcairiel

18th August 2011, 13:21

madVRs internal decoder does not do any form of deinterlacing, IVTC or any other magic to combine fields properly, as far as i can tell it just weaves them together.

Any reason the install.bat will not add them?

The bat file just registers the filter, its in no way suitable for a full blown installation process.
I could as well just remove the bat files, i dont think they really have an advantage over the installer. They are harder to use (need to manually run as admin), have less functionality .. meh.

nlnl

18th August 2011, 13:33

madVRs internal decoder does not do any form of deinterlacing, IVTC or any other magic to combine fields properly, as far as i can tell it just weaves them together.

And for PAL films is just weaves them together equivalent to correct IVTC 50i-->25p ?

nand chan

18th August 2011, 13:35

The bat file just registers the filter, its in no way suitable for a full blown installation process.
I could as well just remove the bat files, i dont think they really have an advantage over the installer. They are harder to use (need to manually run as admin), have less functionality .. meh.

Probably should. Either that, or just add the keys to the .bat file as well (reg add hklm\etc\bla). I've always run the .bat, never figured there was a difference from the installer.

nand chan

18th August 2011, 16:00

In the effort of finding out the difference between upscalers, I took a screenshot of a bunch of them on a 720p Hi10P release (as well as one of a 1080p release) using different upscaler settings.

I was surprised to find out that, well, there really isn't a damn difference because apart from nearest neighbour, they all looked the same to the eye (on this particular image, at least).

In effort to observe the difference beyond the limits of the human eye, I compared both lanczos 3 taps with lanczos 8 taps and lanczos 3 taps with spline 4 taps using a pixel-wise image difference calculation program, which basically subtracts the value in the latter image from the value in the former image, amplifies the change by a factor of 50.0, then normalizes it to 127/127/127 (so that gray = no difference, white = image 2 brighter than image 1, black = image 2 darker than image 1).

Some interesting comparisons have been made, especially between lanczos 3 taps and 8 taps.

An archive of all test images + the difference images can be found here: http://www.mediafire.com/?zcadn71upr1e597

In addition to this, I have (experimentally) separated the positive and negative offset components between lanczos3 and lanczos8, amplified them by 20.0 and brightened / darkened the original image using those images as masks. The resulting image is essentially a composite of lanczos8 as difference of lanczos3, except amplified significantly enough to observe a visible difference: http://img834.imageshack.us/img834/5237/lanczos3v8color.png

The image is interesting because it immediately reveals the ringing: what was originally one line now has one darker line beside it, and one brighter line on the other side (harder to see), which signify very faint ghost line in those places on the lanczos 8 tap image.

Razoola

18th August 2011, 17:21

nand chan

18th August 2011, 17:30

All is quiet from me while I await my gtx295 being RMA'd. I have to ask the question I've been dying to ask madshi for a while.....

Is the fact you added you own video decoders to madVR a step towards madVR supporting 3D content?

It's not his own video decoder, it's just a packaged form of libavcodec from the ffmpeg project.

Isn't stereoscopic content just encoded as a very wide or very high regular frame that is then simply split in the middle by the display device? Or is it a separate video stream?

Either way, actually rendering it would depend strongly on the type of stereoscopic device you are intending to use - the simplest option would be rendering each image to a separate display device (equal in size), which would allow you to use polarized or interferential projector setups (or just cross your eyes); the option of shutter glasses would probably be strongly dependent on some proprietary device-specific implementations.

ForceX

18th August 2011, 17:38

I was surprised to find out that, well, there really isn't a damn difference because apart from nearest neighbour, they all looked the same to the eye (on this particular image, at least).
Well of course. Unless you're doing huge upscaling/downscaling (like using 480p source to display in 1080p or vice-versa) the scaler does not create enough new pixels to make a big difference, and you'll never notice those differences unless you pause the video and compare individual frames. Bicubic 60 works more than fine enough for me, Lanczos is just a waste of GPU cycles and I could live without the ringing. Aliasing is more of a problem with anime than ringing, though.

It will be hard to notice differences with clean HD source. Maybe you can try this again with a low-res DVD source and see if existing noise gets amplified or not.

In effort to observe the difference beyond the limits of the human eye...
Now that's just placebo =P

e-t172

18th August 2011, 17:53

And for PAL films is it just weaves them together equivalent to correct IVTC 50i-->25p ?

Yes. Weave is enough to IVTC PAL 50i content back to 25p. To get 24p you'll need to slow down the audio (assuming the broadcaster did not use timestretching, which unfortunately is impossible to tell for sure).

Razoola

18th August 2011, 17:54

It's not his own video decoder, it's just a packaged form of libavcodec from the ffmpeg project.

Isn't stereoscopic content just encoded as a very wide or very high regular frame that is then simply split in the middle by the display device? Or is it a separate video stream?

Yes its just like you say, normally side by side which is why I wonder having the decoders with madvr is a step towards this.

nevcairiel

18th August 2011, 17:59

Isn't stereoscopic content just encoded as a very wide or very high regular frame that is then simply split in the middle by the display device? Or is it a separate video stream?

Both types exist. There are files around which are just encoded as SBS (Side-by-Side, basically giving you a 3840x1080 image). You'll find those if someone rips a 3D BD to MKV.

But there are also the original 3D Blu-rays, which use the MVC codec. Using that, one eye is encoded as normal H264/AVC1, and the other is using the special MVC coding. You can decode the one H264 eye alone just fine (backwards compat), but for the second eye you need a special MVC decoder (the MVC encoded stream is not a complete stream, its dependant on the H264 stream to be decoded).

As of now, the only MVC decoder in DirectShow is CoreMVC (except of course the integrated decoders in TMT/PowerDVD/etc), and thats not officially released yet. Its also quite uncertain how the frames from CoreMVC will be output and send to the renderer - well at least to me, maybe madshi has talked with someone at CoreCodec about that.

Now that i think about it, i'm not sure how they expect to sell CoreMVC, unless they also ship a renderer with it. ;)

nand chan

18th August 2011, 18:22

Both types exist. There are files around which are just encoded as SBS (Side-by-Side, basically giving you a 3840x1080 image). You'll find those if someone rips a 3D BD to MKV.

But there are also the original 3D Blu-rays, which use the MVC codec. Using that, one eye is encoded as normal H264/AVC1, and the other is using the special MVC coding. You can decode the one H264 eye alone just fine (backwards compat), but for the second eye you need a special MVC decoder (the MVC encoded stream is not a complete stream, its dependant on the H264 stream to be decoded).

As of now, the only MVC decoder in DirectShow is CoreMVC (except of course the integrated decoders in TMT/PowerDVD/etc), and thats not officially released yet. Its also quite uncertain how the frames from CoreMVC will be output and send to the renderer - well at least to me, maybe madshi has talked with someone at CoreCodec about that.

Now that i think about it, i'm not sure how they expect to sell CoreMVC, unless they also ship a renderer with it. ;)

Well, you could create a filter based on CoreMVC that generates an SBS image then just implement SBS support in madVR.

Or, heck, you could just display it normally and cross your eyes.

THEAST

18th August 2011, 19:24

Ctrl+Alt+Shift+M
Hmm, I thought that just showed the auto-detected matrix. ;p
Anyway, yeah, the colors are fine when I force the decoding matrix to BT.601.

mzso

18th August 2011, 21:03

Your graphics card doesn't seem to be powerful enough. Try using Bicubic resizers, and maybe increase the "present frames in advance" number.

Then why does it work sometimes. I watched a movie for like 10 minutes without a frame drop that usually stutters. And then a while later it stuttered again. Plus noee also with a 2600xt said it works well for him.
Something's definitely not right though. I just can't figure out what.

nand chan

18th August 2011, 21:33

Then why does it work sometimes. I watched a movie for like 10 minutes without a frame drop that usually stutters. And then a while later it stuttered again. Plus noee also with a 2600xt said it works well for him.
Something's definitely not right though. I just can't figure out what.

I hope you realize scenes are not consistent in bitrate and complexity, right?

Just because it works for half the scenes doesn't mean the other half are broken - it means your setup isn't powerful enough.

Unless of course you want to watch exciting documentaries about still frame art gallery exhibitions.

Like he said, enable all of the performance hacks, set the resizer to something lightweight.

mzso

18th August 2011, 21:50

I hope you realize scenes are not consistent in bitrate and complexity, right?

Just because it works for half the scenes doesn't mean the other half are broken - it means your setup isn't powerful enough.

Unless of course you want to watch exciting documentaries about still frame art gallery exhibitions.

Like he said, enable all of the performance hacks, set the resizer to something lightweight.
But that would affect the decoder, wouldn't it? Anyway I watched the more eventful parts and it was working fine then. Also sometimes just plain switchin from full screen to windowed a couple of times makes the stuttering go away.

nand chan

18th August 2011, 22:10

But that would affect the decoder, wouldn't it? Anyway I watched the more eventful parts and it was working fine then. Also sometimes just plain switchin from full screen to windowed a couple of times makes the stuttering go away.

No, the madVR performance hacks don't touch the decoder - they only apply to the renderer, which is solely on the GPU.

That might be due to varying levels of GPU load due to Aero in the background or something.

mzso

18th August 2011, 22:16

No, the madVR performance hacks don't touch the decoder - they only apply to the renderer, which is solely on the GPU.

That might be due to varying levels of GPU load due to Aero in the background or something.
I meant the complexity/bitrate of the scene.

nand chan

18th August 2011, 22:50

I meant the complexity/bitrate of the scene.

Oh, yeah, but the performance issues might be caused by a bottleneck anywhere being strained. You never know. (I also didn't read your problem fully so forgive me if I made a mistake in judgement, been awake for over 24 hours now)

ajp_anton

19th August 2011, 00:20

Bug:
10-bit video (http://www.multiupload.com/C70Z98HGY6) looks weird with madVR's decoders, but works fine with ffdshow.

nand chan

19th August 2011, 12:02

Bug:
10-bit video (http://www.multiupload.com/C70Z98HGY6) looks weird with madVR's decoders, but works fine with ffdshow.

Works fine using LAV Video as well for the reference.

nand chan

19th August 2011, 16:26

I wrote a small utility to parse an ArgyllCMS .ti3 file, average the white point, properly format it, and output the values so that it can be directly copy/pasted to madVR's yCMS window.

Executable can be found here, requires .NET 2.0 or above: http://www.mediafire.com/?ay5ituooueoit9m
Source code: http://www.mediafire.com/?4ozau7fa66nd402

This will come in handy for easily creating .3dluts from profile data. I recommend using Argyll's Curves + Matrix profile type to generate the .ti3. All code in public domain.

Edit 1: The same program as a command line utility for those who dislike GUIs:

Executable: http://www.mediafire.com/?8uk0u1ay9m18c9k
Source code: http://www.mediafire.com/?hkyg6z85kmnp4hp

Usage: TI3Console.exe <path to .ti3 file>

(If you're going to edit the source code, ignore this version - I'll properly outsource the logic into a library used by both programs soon)

Edit 2: Going to extend it to automatically create a .3dlut as well (by generating a yCMS script like madVR does) if anybody thinks this would be useful.

Edit 3: Started work on it already, I'm thinking of making it automatize ArgyllCMS (in addition to yCMS) so all you need to do is plug in your device and hit OK and it will generate a madVR friendly .3dLUT. Thoughts?

nand chan

19th August 2011, 21:11

One thing that has never worked properly with subtitle rendering resolution being different from the video resolution: 3D rotations. Any sign involving these (like the one THEAST posted) will get incorrectly aligned. The worse thing is that there's really no way to properly fix it either since MPC-HC doesn't have a "render subtitles at video resolution"-option.

Correct rendering of typesetting is the main reason why I keep using external VSFilter, since typesetting is always made with the video resolution (and subtitles being rendered in video resolution) in mind.

Also, another thing in regards to typesetting. The internal renderer, for whatever, doesn't seem to support the \fax and \fay tags properly. At least this is the case in the version I have right now (1.5.3.3514).

For comparison:
Correct external VSFilter rendering (http://i.imgur.com/G59Zz.jpg)
Incorrect internal VSFilter rendering (http://i.imgur.com/iaaey.jpg)

This was fixed,

http://img30.imageshack.us/img30/7683/isrscreen.jpg

^ using ISR

@madshi:

Would it be possible to implement an option like “[ ] delay playback start until subtitle render queue full” or something? Eg. the start of idolmaster 6 which has about a thousand softsubs in the first few seconds plays back horribly unless I pause it right away and wait for a few seconds so the buffer has time to fill. (With buffer size set to 60 to cover the whole animation, any less (eg. 15) starts to work but immediately cuts off because the buffer runs out)

Edit: Nevermind about this, setting the buffer size that high creates occasional stuttering and slowdown either way.

True. I could probably change that. However, it might cost performance, I'm not sure. How important is that? I mean we color correct video to make it look "correct". How can it matter for subtitles whether they're color corrected or not? Sounds weird to me.

Pic related? (http://img803.imageshack.us/img803/3100/cmsa.jpg)

hdboy

20th August 2011, 15:29

with system in sig, I get severe glitches if I check (ON) "use a separate device for presentation". Looks like my TV is having a seizure. It happens at random times and only with playback on second display (my TV). (Took me a while to nail down the problem. Hope this hepls someone)

spoke too soon, I still get occasional glitches at intervals with "use a separate device for presentation" off. By glitches I mean the video jerk back and forth like Max Headroom. To get glitch free playback, I have to use either exclusive mode or disable Aero. The path to true video never did run smooth :)

Xaurus

20th August 2011, 16:59

I am lately having a slight "problem" with what I assume to be banding; for example it's sometimes visible in the faces of people (the skin). Is there something that could be done about this?
My chain is LAV Splitter -> LAV CUVID -> MadVR. I am using Softcubic 70, Spline 4 and Spline 4. Makes no difference if I enable or disable dithering in MadVR.

Any help would be appreciated!

nand chan

20th August 2011, 17:10

Xaurus

20th August 2011, 18:02

Are you sure it's not in the source? Do you get it using any other renderers or decoders?
I have not yet tried other renderers or decoders (no time) but by briefly comparing the images on my computer LCD and my main plasma, the issue seems to be only on the plasma. Which basically leaves me as a big question mark...

edit: The source is 720p 4:2:0 23.976 material (normal mkv) and I cannot recall having seen this with 1080p material.
I tried briefly to add the deband filter via ffdshow raw video, but it didn't make a difference either (I confirmed that it was in use).

6233638

20th August 2011, 19:17

Not strictly related to madVR, but I'm hoping someone here will have an answer, as it seems related. (people using madVR care about quality)

Has anyone got IVTC working for NTSC DVDs with madVR?

I'm not in an NTSC region, so I haven't had to deal with 29.97Hz films yet, but I have been sent the wrong region DVD and can't get it working right at all.

In an old changelog for madVR, it was said to detect DVDs as 23.97 if DScaler5 IVTC Mod was being used but it still says 29.97 here, the film is full of combing, stuttering movement, and MPC-HC crashes when trying to bring up the DScaler preferences.

I can get rid of the combing with ffdshow's deinterlacing filter, but I still can't get smooth playback.