ffdshow tryouts project: Discussion & Development

[mark0077]

Guys, when using ffdshow + an avisynth script that changes the frame rate, do you ever have Reclock say

"Video No video stream found."

"Frame rate is unknown"

I am using avisynth and SVP to interpolate all content to 59.94fps, and using reclock to then help me with the difference beteen that and my display at 60.002 hz. This is fine for some content like blu-rays, but some other files that I'm trying like mp4's, with avisynth enabled I get those messages from Reclock and it won't kick in as it should to help me with the frame rate differences.

I'm just wondering does anyone know what causes this, is ffdshow possibly at fault?

Unticking avisynth during playback, immediately allows reclock to then find the Video frame rate of eg 24fps. Ticking avisynth on again shows the problem again but only after the media gets restarted.

Should ffdshow do something to change the output details / pin details after an avisynth script changes the playback rate. madVR for example still picks the rate up as 24fps in its pin input, even though it has been changed to 59.94fps.

Video: YV12 2048x1080 (16:9) 24fps

[nevcairiel]

If you always interpolate to 59.94, you could just set ReClock to this, and don't rely on its auto-detect.

[mark0077]

Hi Nev. Yeah I actually tried locking reclock to output at "refresh rate" bit it seems when it can't figure out the playback rate it won't make any changes ie it won't do anything to make the 59.94 -> 60.001 smoother due to the lack of being able to get the rate.

For this reason I think ill need to figure out a way around this I suppose or not use reclock and interpolate to 60.001 which I was doing originally but it uses more resources than interpolating to a nice multiple like 59.94

[nevcairiel]

You can configure a fallback rate in its settings (not during playback, but the actual settings) which it uses as an assumption if the fps probing fails. This should work for what you need it to do.

[mark0077]

Thanks nev. Ill use that as a workaround. Cheers.

[Blight]

I'm getting reports of crashes in Zoom Player's subtitle cycling feature from build 4496 (4494 doesn't suffer from this).
Was something in ffdshow's subtitle code changed recently to possibly account account for this issue?

[foxyshadis]

Originally Posted by Blight

I'm getting reports of crashes in Zoom Player's subtitle cycling feature from build 4496 (4494 doesn't suffer from this).
Was something in ffdshow's subtitle code changed recently to possibly account account for this issue?

Those two revisions in between were entirely subtitle code. The most likely reason is that embedded subtitles were enabled by default in 4496; if you can get in contact with anyone who had a crash, have them test whether disabling that fixes it. (The "Embedded Subtitles" box on the Subtitles config screen.) It can also be disabled in the registry; subTextpin (DWORD) 0. It'd be handy to have a crash dump or way to reproduce it, naturally.

There's a new release today but with no subtitle changes since then.

[clsid]

Those two changes also made ffdshow stricter with accepting subtitle pin connection. So perhaps you are not handling the situation correctly where ffdshow refuses connection.

[romulous]

@foxyshadis: I asked the person who reported this error to test again with your suggestion. He reported that the error disappears when embedded subtitles are disabled, but that also disables the Zoom Player subtitle cycling feature entirely. These are the reported reproduction steps - you will just need ffdshow 4496 or later, and a copy of Zoom Player (the just released 8.6.1 should be fine):

1. Install ffdshow rev 4496 or later.
2. Ensure Zoom is using ffdshow as the subtitle renderer. It uses VobSub by default - this can be set in Advanced Options-->Playback-->Video-->Subtitles.
3. Open the following test file in Zoom:
https://dl.dropbox.com/u/105555957/Bug%20%23371.mkv (50MB)
4. Ensure 'disabled' is selected in the Stream Selection menu (you see this when you right click the Zoom Player window), then use Ctrl+B to cycle through the subtitles.
5. Cycling to the first subtitle in the index should show the error.

This is what happened with the person who reported it disabled embedded subtitles in ffdshow:
"True, the "crash" does not happen any more, but also cycling to "disabled" does not work now, and it stays at the first subtitle in the Stream selection index. Using the keyboard to cycle stopped working as well. I can change the subtitle using the menu and mouse, but they do not display, obviously because embedded subs are not enabled I guess."

@clsid: Thanks, I have sent your comment to Blight.

[Vincent Vega]

like the title says..

if I uninstall ffdshow x64, dxdiag x64 runs normally

[clsid]

Uninstall it and install the latest version.

[kc7bfi]

I am having an issue where it is taking about 1.3 seconds on Windows 7 64bit to setup the filtergraph. I've set ffdshow to output the debug lines to the windows debugger. It seems that the time is being taken calling

ffdshowDecVideo::CheckTransform. Here is a portion of the debug log
00000061 0.48507661 [8296] TffdshowDecVideoOutputPin::CompleteConnect
00000062 0.48508927 [8296] TffdshowDecVideo::CompleteConnect output
00000063 0.48510885 [8296] TffdshowDecVideo:ecideBufferSize
00000064 0.48514879 [8296] TffdshowDecVideo::CheckTransform
00000065 0.72769743 [8296] TffdshowDecVideo::CheckTransform
00000066 0.95664924 [8296] TffdshowDecVideo::CheckTransform
00000067 1.19082177 [8296] TffdshowDecVideo::CheckTransform
00000068 1.40908623 [8296] TffdshowDecVideo::CheckTransform
00000069 1.59862018 [8296] TffdshowDecVideo::CheckTransform
00000070 1.59915233 [8296] TffOutputQueue::Constructor
00000071 1.59917843 [8296] Connection succeeded

Notice that the CheckTransform calls take over 1 second to finish. Any idea why it is taking so long? Is there something else happening that is taking the time? Any idea how to proceed to debug the delay?

Thanks, David

[clsid]

No idea. It is probably related to the other filters that are tried when creating the graph. GraphStudioNext can generate a graph construction report.

[LigH]

Reported in this forum for years and forgotten, therefore in the bugtracker now:

Noise exaggerates in the right third in reduced window

When the noise generate feature is enabled, and the video is displayed in a window smaller than the video size (e.g. a FullHD video in windowed mode fitting from inside, or in 50% Zoom), the noise is extremely exaggerated in the right third of the window.

[kc7bfi]

I have a problem with a particular camera where the "top" field of a H.264 video frame is being processed twice. Here is the output from the debugger log

00001478 27.11919022 [8892] slice:1 T mb:0 P fix pps:0 frame:5 poc:65546/2147483647 ref:1/1 qp:26 loop:1:0:0 weight:0
00001479 27.12047768 [8892] slice:1 B mb:0 P fix pps:0 frame:5 poc:65546/65547 ref:1/1 qp:25 loop:1:0:0 weight:0
00001480 27.16013908 [8892] slice:1 T mb:0 P fix pps:0 frame:6 poc:65548/2147483647 ref:1/1 qp:25 loop:1:0:0 weight:0
00001481 27.16434860 [8892] slice:1 T mb:0 P fix pps:0 frame:6 poc:65548/2147483647 ref:1/1 qp:25 loop:1:0:0 weight:0
00001482 27.16610336 [8892] mmco: unref short failure
00001483 27.16611290 [8892] mmco: unref short failure
00001484 27.16614342 [8892] illegal short term buffer state detected
00001485 27.16657066 [8892] slice:1 B mb:0 P fix pps:0 frame:6 poc:65548/65549 ref:1/1 qp:26 loop:1:0:0 weight:0
00001486 27.18620110 [8892] slice:1 T mb:0 P fix pps:0 frame:7 poc:65550/2147483647 ref:1/1 qp:25 loop:1:0:0 weight:0
00001487 27.18804932 [8892] slice:1 B mb:0 P fix pps:0 frame:7 poc:65550/65551 ref:1/1 qp:25 loop:1:0:0 weight:0
00001488 27.22807312 [8892] slice:1 T mb:0 P fix pps:0 frame:8 poc:65552/2147483647 ref:1/1 qp:26 loop:1:0:0 weight:0
00001489 27.23173714 [8892] slice:1 T mb:0 P fix pps:0 frame:8 poc:65552/2147483647 ref:1/1 qp:26 loop:1:0:0 weight:0
00001490 27.23297882 [8892] mmco: unref short failure
00001491 27.23299026 [8892] mmco: unref short failure
00001492 27.23301888 [8892] illegal short term buffer state detected
00001493 27.23345947 [8892] slice:1 B mb:0 P fix pps:0 frame:8 poc:65552/65553 ref:1/1 qp:25 loop:1:0:0 weight:0

Notice that the "top" field of frames 6 and 8 are processed twice. Is there some way to detect this and to ignore the duplicate field? I've looked at the h264.c source code but cannot seem to find an easy way to do this.

Thanks, David

[Blight]

Ok, I narrowed down why recent ffdshow subtitle renderer builds broke ZP's subtitle functions.
There's a bug in ffdshow where it doesn't update it's IAMStreamSelect interface dynamically if a selection was made on one of the filter ffdshow's sub renderer is aggregating IAMStreamSelect information from.

So if LAV is updated with a different subtitle stream through its IAMStreamSelect, the next time FFDShow's IAMStreamSelect is queried, it is still showing the old stream as enabled...

If my description isn't clear enough, let me know.

[clsid]

There haven't been any changes related to IAMStreamSelect. You can see recent changes at
http://ffdshow-tryout.svn.sourceforge.net/viewvc/ffdshow-tryout?revision=4505&view=revision

[Blight]

Originally Posted by clsid

There haven't been any changes related to IAMStreamSelect. You can see recent changes at
http://ffdshow-tryout.svn.sourceforge.net/viewvc/ffdshow-tryout?revision=4505&view=revision

And still, this bug exists and didn't previously...
And you can recreate it without ZP specific code.

[clsid]

Please understand that ffdshow development is inactive. So if you want this to be fixed, you will have to submit a patch.

I did a quick test in MPC-HC with the raw subtitle filter and it seemed to work just fine. Correct subtitle is selected in ffdshow after changing sub in player or splitter. Tested an mkv with two embedded srt subs.

[vedmysh]

Hello. I am wondering is the project rather dead than alive?
I am developing a software and use ffdshow-tryouts as a primary decoding facility. From time to time i encounter a random crash when i try to counstruct the video decoder:

 	ntdll.dll!_ZwRaiseException@12()  + 0x12 bytes	
 	ntdll.dll!_ZwRaiseException@12()  + 0x12 bytes	
 	ntdll.dll!_NtQueryValueKey@24()  + 0x12 bytes	
 	kernel32.dll!_LocalBaseRegQueryValue@24()  + 0x208 bytes	
 	ntdll.dll!@RtlpAllocateHeap@24()  + 0x95 bytes	
 	ntdll.dll!_RtlAllocateHeap@12()  + 0x5ca8 bytes	
 	ffdshow.ax!5136e66a() 	
 	[Frames below may be incorrect and/or missing, no symbols loaded for ffdshow.ax]	
 	ffdshow.ax!5136adf4() 	
 	ffdshow.ax!511928fe() 	
 	ffdshow.ax!5139dedc() 	
 	ffdshow.ax!511790bf() 	
 	ffdshow.ax!5139cb26() 	
 	ffdshow.ax!513624cf() 	
 	ole32.dll!CServerContextActivator::CreateInstance(IUnknown * pUnkOuter, IActivationPropertiesIn * pInActProperties, IActivationPropertiesOut * * ppOutActProperties)  Line 1000	C++
 	ole32.dll!ActivationPropertiesIn::DelegateCreateInstance(IUnknown * pUnkOuter, IActivationPropertiesOut * * ppActPropsOut)  Line 1917	C++
 	ole32.dll!CApartmentActivator::CreateInstance(IUnknown * pUnkOuter, IActivationPropertiesIn * pInActProperties, IActivationPropertiesOut * * ppOutActProperties)  Line 2268 + 0x12 bytes	C++
 	ole32.dll!CProcessActivator::CCICallback(unsigned long dwContext, IUnknown * pUnkOuter, ActivationPropertiesIn * pActIn, IActivationPropertiesIn * pInActProperties, IActivationPropertiesOut * * ppOutActProperties)  Line 1737 + 0x12 bytes	C++
 	ole32.dll!CProcessActivator::AttemptActivation(ActivationPropertiesIn * pActIn, IUnknown * pUnkOuter, IActivationPropertiesIn * pInActProperties, IActivationPropertiesOut * * ppOutActProperties, HRESULT (unsigned long, IUnknown *, ActivationPropertiesIn *, IActivationPropertiesIn *, IActivationPropertiesOut * *)* pfnCtxActCallback, unsigned long dwContext)  Line 1630	C++
 	ole32.dll!CProcessActivator::ActivateByContext(ActivationPropertiesIn * pActIn, IUnknown * pUnkOuter, IActivationPropertiesIn * pInActProperties, IActivationPropertiesOut * * ppOutActProperties, HRESULT (unsigned long, IUnknown *, ActivationPropertiesIn *, IActivationPropertiesIn *, IActivationPropertiesOut * *)* pfnCtxActCallback)  Line 1487 + 0x12 bytes	C++
 	ole32.dll!CProcessActivator::CreateInstance(IUnknown * pUnkOuter, IActivationPropertiesIn * pInActProperties, IActivationPropertiesOut * * ppOutActProperties)  Line 1377 + 0x18 bytes	C++
 	ole32.dll!ActivationPropertiesIn::DelegateCreateInstance(IUnknown * pUnkOuter, IActivationPropertiesOut * * ppActPropsOut)  Line 1917	C++
 	ole32.dll!CClientContextActivator::CreateInstance(IUnknown * pUnkOuter, IActivationPropertiesIn * pInActProperties, IActivationPropertiesOut * * ppOutActProperties)  Line 685	C++
 	ole32.dll!ActivationPropertiesIn::DelegateCreateInstance(IUnknown * pUnkOuter, IActivationPropertiesOut * * ppActPropsOut)  Line 1917	C++
 	ole32.dll!ICoCreateInstanceEx(const _GUID & Clsid, IUnknown * punkOuter, unsigned long dwClsCtx, _COSERVERINFO * pServerInfo, unsigned long dwCount, unsigned long dwActvFlags, tagMULTI_QI * pResults, ActivationPropertiesIn * pActIn)  Line 1334 + 0x19 bytes	C++
 	ole32.dll!CComActivator::DoCreateInstance(const _GUID & Clsid, IUnknown * punkOuter, unsigned long dwClsCtx, _COSERVERINFO * pServerInfo, unsigned long dwCount, tagMULTI_QI * pResults, ActivationPropertiesIn * pActIn)  Line 343	C++
 	ole32.dll!CoCreateInstanceEx(const _GUID & Clsid, IUnknown * punkOuter, unsigned long dwClsCtx, _COSERVERINFO * pServerInfo, unsigned long dwCount, tagMULTI_QI * pResults)  Line 157	C++
 	ole32.dll!CoCreateInstance(const _GUID & rclsid, IUnknown * pUnkOuter, unsigned long dwContext, const _GUID & riid, void * * ppv)  Line 110	C++
>	playerng.dll!RenderManager::video_decoder_create(IBaseFilter * * ppdecoder)  Line 538 + 0x20 bytes	C++

Is it possible to recieve the help here about this problem?
Thanks.