Media Player Classic Home Cinema (MPC-HC) - DXVA!

[Aleksoid1978]

i found a problem with my gt220 and videos who are decoded in vc-1. if i want to try vc-1 video i get a green screen and can only listen to the sound. this happens only if i try to start the video with dxva and not with ffdshow. so mayby is the dxva decoder for vc-1 videos broken? i am using win7 x64.

This happen on interlace VC-1. Try test on progressive.

[Zachs]

Quick question: why's MPC-HC's subtitle menu item (right-click pop-up context menu) disabled when EVR/EVR-sync/EVR-custom-preset is used in XP?

[clsid]

Quick question: why's MPC-HC's subtitle menu item (right-click pop-up context menu) disabled when EVR/EVR-sync/EVR-custom-preset is used in XP?
Either because the internal subtitle renderer is disabled or because no subtitles were present.

[iwod]

Haven't been following MPC-HC, just wondering why newer builds are out on the official Sourceforce page? Official builds haven't been updated for a long while..

[XhmikosR]

Because only stable builds are hosted on the SF.net page.

[Steveo08]

This happen on interlace VC-1. Try test on progressive.


i have only interlaced vc1, or is it possible to change to progressive without recoding?

[Px]

Win 7 x64, GTX 260 + GT 220, EVR Custom, VMR9 Renderles - all work fine at all video content with or whithout DXVA
So, the problem may be related to heterogeneous multi-adapter (NV+AMD), don't have 2 cards from same vendor, to test now, but this doesn't explain problems on second display on same card.
Any additional info needed, to debug this and related issues?

[Aleksoid1978]

i have only interlaced vc1, or is it possible to change to progressive without recoding?

Use other external decoder, but only not ffdshow(ffmpeg bases decoder can't decode VC-1 interlace)

[XhmikosR]

No, that's not how you do it. https://sourceforge.net/apps/trac/mpc-hc/wiki/Manual

[Razoola]

So, the problem may be related to heterogeneous multi-adapter (NV+AMD), don't have 2 cards from same vendor, to test now, but this doesn't explain problems on second display on same card.
Any additional info needed, to debug this and related issues?

I'm trying to understand the function of this new option to choose the D3D rendering device. I sort of understand its purpose but it seems not to work at all.... Maybe I am wrong on what its soppost to do though. I assume it its for choosing the card that going to do DXVA decoding etc?

I have a gtx295 and gt240 and a display attached to each (TV attached to 240 via hdmi and monitor to gtx295 via dvi).

Here the new option does not seem to be doing anything different from what was happening before.

For example, if I choose the gtx295 in the new option and then play a video on the TV attached to the gt240 the MPC video render still reports work its being done via the gt240 in the DXVA settings part at the bottom. This happens if I play the video on the card I have selected first and then expand to fullscreen on the other card. If I try to expand to fullscreen right away to the opposite card selected in this new setting (without using the other card display out at all) MPC-HC crashes.

I'm on Win7 x64.

Am I missing something here?

[Aleksoid1978]

Am I missing something here?

To see what video adapter you - use GPU-Z, see engine, memory load
In the MpcVideoDecoder setting you can see wrong name;
Also you can try this - run mpc on GTX 295(monitor), select GT 240 and open VC-1 video - you see that use VC-1 bitsream decoder - GTX 295 can't do this :)

[Zachs]

Either because the internal subtitle renderer is disabled or because no subtitles were present.

Strange. With all other settings untouched, switching from VMR9 to EVR-sync disables the subtitle menu item when the same file is reopened. Does the EVR-sync / custom preset renderer disable the internal subtitle renderer when it's being used in winXP? Otherwise, this would be a bug then?

This is with x86 r1731.

[edigee]

Strange. With all other settings untouched, switching from VMR9 to EVR-sync disables the subtitle menu item when the same file is reopened. Does the EVR-sync / custom preset renderer disable the internal subtitle renderer when it's being used in winXP? Otherwise, this would be a bug then?

This is with x86 r1731.
Yes it does! It is (EVR_Sync) based on EVR CP so its behaviour is the same. You can't use EVR CP in XP with internal MPC HC subtitles (at least with MPC internal H264 decoder). I guess the same thing happens with ffdshow(DXVA)_I don't believe you can use ffdshow subtitle renderer either with EVR CP(or EVR Sync) in XP.
Edit:
Conclusion: You can 't have both DXVA and subtitles with EVR CP (Sync) in XP.

[Steveo08]

Use other external decoder, but only not ffdshow(ffmpeg bases decoder can't decode VC-1 interlace)



which other external decoder? i only have coreavc as external decoder and this won't work with vc1.

[Px]

which other external decoder?
Built-in Microsoft decoder

[Aleksoid1978]

which other external decoder? i only have coreavc as external decoder and this won't work with vc1.

ArcSoft, Cyberlink, Nero ...

[Zachs]

Yes it does! It is (EVR_Sync) based on EVR CP so its behaviour is the same. You can't use EVR CP in XP with internal MPC HC subtitles (at least with MPC internal H264 decoder). I guess the same thing happens with ffdshow(DXVA)_I don't believe you can use ffdshow subtitle renderer either with EVR CP(or EVR Sync) in XP.
Edit:
Conclusion: You can 't have both DXVA and subtitles with EVR CP (Sync) in XP.

What's the reason behind that? What's the difference technically between EVR CP in XP vs EVR CP in Vista/Win7? What happens if it's forced? It just wouldn't show?

[Razoola]

To see what video adapter you - use GPU-Z, see engine, memory load
In the MpcVideoDecoder setting you can see wrong name;
Also you can try this - run mpc on GTX 295(monitor), select GT 240 and open VC-1 video - you see that use VC-1 bitsream decoder - GTX 295 can't do this :)

Ok, understand a little more now, I see you added info to stats about the device used which makes it more clear, I will test more later when home.

I have also noticed that the option to choose the render device is not available (greyed out) unless both cards have an active display. Is it not possible to use a card even though no display is used on that device? So in my case for example use the gt240 to render while the only active monitor is on the gtx295?

[Virtual_ManPL]

MPC-HC bug on three lined subtitles



MPC-HC
http://img186.imageshack.us/img186/6159/mpchc.th.png (http://img186.imageshack.us/i/mpchc.png/)

SMPlayer
http://img704.imageshack.us/img704/5865/smplayer.th.png (http://img704.imageshack.us/i/smplayer.png/)

subtitles download
http://www.mediafire.com/?xtygkg22gnn



and also why "Position subtitles relative to the video frame" is not fully enabled by default ?

[Razoola]

To see what video adapter you - use GPU-Z, see engine, memory load
In the MpcVideoDecoder setting you can see wrong name;
Also you can try this - run mpc on GTX 295(monitor), select GT 240 and open VC-1 video - you see that use VC-1 bitsream decoder - GTX 295 can't do this :)

I have had a play with build 1735 but I just cannot get anything to play right when I set the d3d render device to the card that is not displaying the media. If I try VMR9 renderless mpc freezes. With EVR custom the media tries to play but its constant stutter and MPC reports the media framerate as 0.

http://img176.imageshack.us/img176/1139/16755581.th.jpg (http://img176.imageshack.us/i/16755581.jpg/)

[PetitDragon]

Problem (or Bug) Report:

I'm getting the following error when I turn on "10 Bit RGB" and "D3D Full Screen Mode" at the same time for 10bit display:

EVR: CreateDevice failed, DX9AllocatorPresenter failed, D3DERR_NOTAVAILABLE

My config: Vista 32bit, NV 8800GTS 512

Please help! THNX in advance.

[Superb]

Quick technical question...
The position subtitles relative to the video frame has 3 states, but the code for getting the value is:
DDX_Check(pDX, IDC_CHECK_RELATIVETO, m_relativeTo);
Where m_relativeTo is a BOOL (=2 states).

What am I missing? how is the third state (interim) saved?
What is the meaning of all 3 values?

EDIT... wow... typedef int BOOL; never mind then. :/

EDIT2: interesting blog post about it... http://blogs.msdn.com/oldnewthing/archive/2004/12/22/329884.aspx

[thuan]

MPC-HC bug on three lined subtitles



MPC-HC
http://img186.imageshack.us/img186/6159/mpchc.th.png (http://img186.imageshack.us/i/mpchc.png/)

SMPlayer
http://img704.imageshack.us/img704/5865/smplayer.th.png (http://img704.imageshack.us/i/smplayer.png/)

subtitles download
http://www.mediafire.com/?xtygkg22gnn



and also why "Position subtitles relative to the video frame" is not fully enabled by default ?

Regression confirmed, I have the same issue over here. But I was lazy to report it :D. You might want to do this (http://sourceforge.net/apps/trac/mpc-hc/wiki/How_to_Report_Issues), too.

[rumpumpel1]

can someone please confirm, that the DXVAChecker reports ModeVC1_VLD on the following platform: Windows7 x64 + Mobile Intel 4 Series Express (GMA 4500MHD) ?

[rica]

Hi guys,

you may wanna have a look at this:

http://forum.doom9.org/showthread.php?p=1370134#post1370134


_ _ _ _ _

[cca]

MPC-HC bug on three lined subtitles



MPC-HC
http://img186.imageshack.us/img186/6159/mpchc.th.png (http://img186.imageshack.us/i/mpchc.png/)

SMPlayer
http://img704.imageshack.us/img704/5865/smplayer.th.png (http://img704.imageshack.us/i/smplayer.png/)

subtitles download
http://www.mediafire.com/?xtygkg22gnn



and also why "Position subtitles relative to the video frame" is not fully enabled by default ?

Problem confirmed, first appeared in revision 1690 (fom xvidvideo.ru) which makes commit 1688 rather suspicious. Revision 1687 from the same site displays the proper subtitles.

Created Ticket #345 for this issue in the bug tracker.

[THX-UltraII]

general MPC-HC question:

what exactly does the option 'reinitialize when display changing' does when you enable or disable it?

[betaking]

can not compiled last mpc-hc 1744 download last gcc form xvidvideo.ru!
Build Log Build started: Project: libavcodec_gcc, Configuration: Release|Win32
Command Lines Creating temporary file "C:\DOCUME~1\xxxx\LOCALS~1\Temp\BAT00015728482460.bat" with contents
[
@echo off

gccbuild.bat

if errorlevel 1 goto VCReportError

goto VCEnd

:VCReportError

echo Project : error PRJ0019: A tool returned an error code from "Performing Makefile project actions"

exit 1

:VCEnd
]
Creating command line "C:\DOCUME~1\david\LOCALS~1\Temp\BAT00015728482460.bat"
Output Window Performing Makefile project actions
gcc.exe -c -I. -I.. -Ilibavcodec -Ilibavutil -Ilibswscale -I../../../../zlib -I../../../../libpng -DHAVE_AV_CONFIG_H -D_ISOC99_SOURCE -D_POSIX_C_SOURCE=200112 -std=gnu99 -DARCH_X86_32 -DWIN32 -D_WIN32 -mno-cygwin -mdll -mthreads -pipe -DNDEBUG -UDEBUG -O3 -march=i686 -msse -mfpmath=sse -fno-common -fno-tree-vectorize -fomit-frame-pointer -fno-math-errno -MMD -o libavcodec/h264_loopfilter.o libavcodec/h264_loopfilter.c
In file included from libavcodec/h264_loopfilter.c:32:0:
libavcodec/h264.h: In function 'fill_filter_caches':
libavcodec/h264.h:1277:84: warning: integer overflow in expression
libavcodec/h264.h:1283:38: warning: initialization from incompatible pointer type
libavcodec/h264.h:1360:38: warning: initialization from incompatible pointer type
libavcodec/h264.h:1368:92: warning: integer overflow in expression
libavcodec/h264.h:1375:42: warning: initialization from incompatible pointer type
gcc.exe -c -I. -I.. -Ilibavcodec -Ilibavutil -Ilibswscale -I../../../../zlib -I../../../../libpng -DHAVE_AV_CONFIG_H -D_ISOC99_SOURCE -D_POSIX_C_SOURCE=200112 -std=gnu99 -DARCH_X86_32 -DWIN32 -D_WIN32 -mno-cygwin -mdll -mthreads -pipe -DNDEBUG -UDEBUG -O3 -march=i686 -msse -mfpmath=sse -fno-common -fno-tree-vectorize -fomit-frame-pointer -fno-math-errno -MMD -o libavcodec/mpegvideo.o libavcodec/mpegvideo.c
gcc.exe -c -I. -I.. -Ilibavcodec -Ilibavutil -Ilibswscale -I../../../../zlib -I../../../../libpng -DHAVE_AV_CONFIG_H -D_ISOC99_SOURCE -D_POSIX_C_SOURCE=200112 -std=gnu99 -DARCH_X86_32 -DWIN32 -D_WIN32 -mno-cygwin -mdll -mthreads -pipe -DNDEBUG -UDEBUG -O3 -march=i686 -msse -mfpmath=sse -fno-common -fno-tree-vectorize -fomit-frame-pointer -fno-math-errno -MMD -o libavcodec/msmpeg4.o libavcodec/msmpeg4.c
gcc.exe -c -I. -I.. -Ilibavcodec -Ilibavutil -Ilibswscale -I../../../../zlib -I../../../../libpng -DHAVE_AV_CONFIG_H -D_ISOC99_SOURCE -D_POSIX_C_SOURCE=200112 -std=gnu99 -DARCH_X86_32 -DWIN32 -D_WIN32 -mno-cygwin -mdll -mthreads -pipe -DNDEBUG -UDEBUG -O3 -march=i686 -msse -mfpmath=sse -fno-common -fno-tree-vectorize -fomit-frame-pointer -fno-math-errno -MMD -o libavcodec/msmpeg4data.o libavcodec/msmpeg4data.c
libavcodec/h264_loopfilter.c: In function 'ff_h264_filter_mb_fast':
libavcodec/h264_loopfilter.c:308:6: internal compiler error: Segmentation fault
Please submit a full bug report,
with preprocessed source if appropriate.
See for instructions.
make.exe: *** [libavcodec/h264_loopfilter.o] Error 1
make.exe: *** Waiting for unfinished jobs....
Results Build log was saved at "file://e:\mpc-hc\src\filters\transform\MPCVideoDec\ffmpeg\BuildLog.htm"
libavcodec_gcc - 1 error(s), 5 warning(s)
MPC-HC x86: (Microsoft Visual C++ 2008 SP1, Microsoft Windows SDK for Windows 7, DirectX SDK v.9.28.1886, GCC 4.5.0 (experimental, svn rev. 157199, 2010.03.03) x86_32)
but download GCC form xvidvideo.ru still is mingw_gcc_450_v20100225_0545_x86-x64-experimental_release_(pthreads_2.9.0.0_shared).7z

[hoborg]

Hi.
Is there plan to update MP4 splitter to support this sample (http://hobring.ic.cz/samples/mp4.AAC+AVC1_(1280x720@50fps).rar) ?

It is working with Haali, but only black screen using "MP4Splitter.ax".
:thanks:

[namaiki]

Hi.
Is there plan to update MP4 splitter to support this sample (http://hobring.ic.cz/samples/mp4.AAC+AVC1_(1280x720@50fps).rar) ?

It is working with Haali, but only black screen using "MP4Splitter.ax".
:thanks:

Wow. That's so weird. I can play this either if I dont use the internal mp4 splitter, or if I don't use mpc-hc's internal h.264 filters (both dxva and ffmpeg).

[JonasNo]

can not compiled last mpc-hc 1744 download last gcc form xvidvideo.ru!
Build Log Build started: Project: libavcodec_gcc, Configuration: Release|Win32
Command Lines Creating temporary file "C:\DOCUME~1\xxxx\LOCALS~1\Temp\BAT00015728482460.bat" with contents
[
@echo off

gccbuild.bat

if errorlevel 1 goto VCReportError

goto VCEnd

:VCReportError

echo Project : error PRJ0019: A tool returned an error code from "Performing Makefile project actions"

exit 1

:VCEnd
]
Creating command line "C:\DOCUME~1\david\LOCALS~1\Temp\BAT00015728482460.bat"
Output Window Performing Makefile project actions
gcc.exe -c -I. -I.. -Ilibavcodec -Ilibavutil -Ilibswscale -I../../../../zlib -I../../../../libpng -DHAVE_AV_CONFIG_H -D_ISOC99_SOURCE -D_POSIX_C_SOURCE=200112 -std=gnu99 -DARCH_X86_32 -DWIN32 -D_WIN32 -mno-cygwin -mdll -mthreads -pipe -DNDEBUG -UDEBUG -O3 -march=i686 -msse -mfpmath=sse -fno-common -fno-tree-vectorize -fomit-frame-pointer -fno-math-errno -MMD -o libavcodec/h264_loopfilter.o libavcodec/h264_loopfilter.c
In file included from libavcodec/h264_loopfilter.c:32:0:
libavcodec/h264.h: In function 'fill_filter_caches':
libavcodec/h264.h:1277:84: warning: integer overflow in expression
libavcodec/h264.h:1283:38: warning: initialization from incompatible pointer type
libavcodec/h264.h:1360:38: warning: initialization from incompatible pointer type
libavcodec/h264.h:1368:92: warning: integer overflow in expression
libavcodec/h264.h:1375:42: warning: initialization from incompatible pointer type
gcc.exe -c -I. -I.. -Ilibavcodec -Ilibavutil -Ilibswscale -I../../../../zlib -I../../../../libpng -DHAVE_AV_CONFIG_H -D_ISOC99_SOURCE -D_POSIX_C_SOURCE=200112 -std=gnu99 -DARCH_X86_32 -DWIN32 -D_WIN32 -mno-cygwin -mdll -mthreads -pipe -DNDEBUG -UDEBUG -O3 -march=i686 -msse -mfpmath=sse -fno-common -fno-tree-vectorize -fomit-frame-pointer -fno-math-errno -MMD -o libavcodec/mpegvideo.o libavcodec/mpegvideo.c
gcc.exe -c -I. -I.. -Ilibavcodec -Ilibavutil -Ilibswscale -I../../../../zlib -I../../../../libpng -DHAVE_AV_CONFIG_H -D_ISOC99_SOURCE -D_POSIX_C_SOURCE=200112 -std=gnu99 -DARCH_X86_32 -DWIN32 -D_WIN32 -mno-cygwin -mdll -mthreads -pipe -DNDEBUG -UDEBUG -O3 -march=i686 -msse -mfpmath=sse -fno-common -fno-tree-vectorize -fomit-frame-pointer -fno-math-errno -MMD -o libavcodec/msmpeg4.o libavcodec/msmpeg4.c
gcc.exe -c -I. -I.. -Ilibavcodec -Ilibavutil -Ilibswscale -I../../../../zlib -I../../../../libpng -DHAVE_AV_CONFIG_H -D_ISOC99_SOURCE -D_POSIX_C_SOURCE=200112 -std=gnu99 -DARCH_X86_32 -DWIN32 -D_WIN32 -mno-cygwin -mdll -mthreads -pipe -DNDEBUG -UDEBUG -O3 -march=i686 -msse -mfpmath=sse -fno-common -fno-tree-vectorize -fomit-frame-pointer -fno-math-errno -MMD -o libavcodec/msmpeg4data.o libavcodec/msmpeg4data.c
libavcodec/h264_loopfilter.c: In function 'ff_h264_filter_mb_fast':
libavcodec/h264_loopfilter.c:308:6: internal compiler error: Segmentation fault
Please submit a full bug report,
with preprocessed source if appropriate.
See for instructions.
make.exe: *** [libavcodec/h264_loopfilter.o] Error 1
make.exe: *** Waiting for unfinished jobs....
Results Build log was saved at "file://e:\mpc-hc\src\filters\transform\MPCVideoDec\ffmpeg\BuildLog.htm"
libavcodec_gcc - 1 error(s), 5 warning(s)
MPC-HC x86: (Microsoft Visual C++ 2008 SP1, Microsoft Windows SDK for Windows 7, DirectX SDK v.9.28.1886, GCC 4.5.0 (experimental, svn rev. 157199, 2010.03.03) x86_32)
but download GCC form xvidvideo.ru still is mingw_gcc_450_v20100225_0545_x86-x64-experimental_release_(pthreads_2.9.0.0_shared).7z

GCC 4.5 is a experimental release which means it can be unstable, crash, generate bad code (slow performance), etc.. so use GCC 4.4 instead, it is stable.
Link: http://www.xvidvideo.ru/2009-10-22-10-49-14/cat_view/28-cross-mingwgcc-x86x64/159-cross-mingwgcc-44-x86x64.html

[fastplayer]

GCC 4.5 is a experimental release which means it can be unstable, crash, generate bad code (slow performance), etc.. so use GCC 4.4 instead, it is stable.
Link: http://www.xvidvideo.ru/2009-10-22-10-49-14/cat_view/28-cross-mingwgcc-x86x64/159-cross-mingwgcc-44-x86x64.html
Then why is ffmpeg in all MPC-HC builds on xvidvideo.ru build with 4.5.0?

[p0w3rh0u5e]

Just two things i would love to see in future-releases of MPC-HC.

- support for R3D-files, with the upcoming release of (cheaper) scarlett-cameras this will become more important
there is a sdk available, you can request it here http://www.red.com/developer
some samples are here: http://www.reduser.net/forum/showthread.php?t=42130
there is a virtualdub-plugin too: http://arenafilm.hu/alsog/vdr3d/

- support for stereoscopic 3d
first thing would be support for stereo-mkv files (mkvmerge can already mux two videostreams into a stereoscopic mkv) and seperate files with some output options like nvidias vision 3d and so on... well, the final goal would be playback of 3d-blurays via hdmi-output on 3d-ready tvs, but yeah, for that, some pieces are missing :)

What do you guys think?

[clumba]

Hello.

I have problem, using last build MPC-HC with EVR Custom (Windows 7 x64 + Radeon HD 5670):

http://i2.fastpic.ru/thumb/2010/0309/bb/6e92a10210008b3e2427c386fe5c7ebb.jpeg (http://fastpic.ru/view/2/2010/0309/6e92a10210008b3e2427c386fe5c7ebb.jpg.html)

Other files & renders works fine.

File can be downloaded here (http://anonym.to/?http://vidding.ru/Video/Full003/01999.[AKROSS_Con_2009]_S_studio_-_D-ON_alt.mkv).

MediaInfo (http://i2.fastpic.ru/big/2010/0309/37/0c6455242b6b4f3ae3b8ad9848c03037.png)

What's wrong?

[Razoola]

Just two things i would love to see in future-releases of MPC-HC.

- support for R3D-files, with the upcoming release of (cheaper) scarlett-cameras this will become more important
there is a sdk available, you can request it here http://www.red.com/developer
some samples are here: http://www.reduser.net/forum/showthread.php?t=42130
there is a virtualdub-plugin too: http://arenafilm.hu/alsog/vdr3d/

- support for stereoscopic 3d
first thing would be support for stereo-mkv files (mkvmerge can already mux two videostreams into a stereoscopic mkv) and seperate files with some output options like nvidias vision 3d and so on... well, the final goal would be playback of 3d-blurays via hdmi-output on 3d-ready tvs, but yeah, for that, some pieces are missing :)

What do you guys think?

I think before 3D can be added the sync render will need to be complete. You can't really be having dropped frames with 3D. I would like to see support in the future though.

[Glaucous]

Hello.

I'm using Windows 7 x64 together with svn 1732 x86 MPC-HC and ReClock, working great, the main feature I need here is WASAPI - mainly so that I won't hear hissing in my back speakers when viewing 2.0 material.

And apparently MPC-HC does support WASAPI natively, by using the MPC-HC Audio Renderer, however I can't get this to work. I tried svn 1732 of both x86 and x64, I tried the stable versions as well.

The setup is the following:
Win7 x64 -> MPC HC -> SPDIF or HDMI -> 5.1 Receiver.
On the receiver I can check if the input signal is 5.1 or 2.0, as well as the sampling rate.
When using MPC HC without ReClock the signal is always as the Windows Audio Mixer settings.

[rica]

Hello.

I'm using Windows 7 x64 together with svn 1732 x86 MPC-HC and ReClock, working great, the main feature I need here is WASAPI - mainly so that I won't hear hissing in my back speakers when viewing 2.0 material.

And apparently MPC-HC does support WASAPI natively, by using the MPC-HC Audio Renderer, however I can't get this to work. I tried svn 1732 of both x86 and x64, I tried the stable versions as well.

The setup is the following:
Win7 x64 -> MPC HC -> SPDIF or HDMI -> 5.1 Receiver.
On the receiver I can check if the input signal is 5.1 or 2.0, as well as the sampling rate.
When using MPC HC without ReClock the signal is always as the Windows Audio Mixer settings.

Post 12201.

[bobdynlan]

I think before 3D can be added the sync render will need to be complete. You can't really be having dropped frames with 3D. I would like to see support in the future though.Before 3D support can be added, a 3D SOFTWARE DECODER, SPLITTER etc needs to be made available (Cyberlink first?!). "Native" support for such thing is far fetched (Nokia helped a little). The thing is that you cannot put such high hopes, when interlaced video is so hard to deal with, even at this time. Mastering interlaced content is the starting point. And to make MPC-HC a true stereo player, it would need a new implementation and integration of the shader engine. With the current one, you can only do anaglyph manipulation for progressive 3d video content. Stuff like converting on-the-fly between color pairs, swap them, and with some work converting up/down or left/right to anaglyph. I should check the new H.264 MVC standard, could have built-in anaglyph conversion, or at least easier to be done than MPEG-2 interlaced 3D content.


Sync renderer is a priority, but it should be coupled with improved file reading.
- Opening time and seeking time are just as important factors in the viewing experience compared to 15 dropped frames, 3 glitches and 33ms audio drift at the end, at least from the POV of an angry german kid (http://www.youtube.com/watch?v=PbcctWbC8Q0). Not even Microsoft could do this right, but there are ways to improve the current situation. Asynchronous I/O, device cache management, device lock, huge buffers as free RAM is wasted RAM.
- But the real issue at least with seeking comes from the fact that pairing a progress bar with a mouse it is very bad for the hardware that makes it happen. A tick marked progress bar that by default jumps to predetermined intervals would make things much easier. And you could also keep the ticks hidden :D. Let's say 5% interval ticks. You click once over what would be 32.84% - would get you to 30%. You click twice there or alternative click, it gets you to 32.84%. And in the background, all those ticks could be cached, and served on-demand almost instantly without trashing the hardware for a seamless jump.

I believe that Media Player Classic - Home Cinema, as a flagship of GREEN software media player, with it's reduced development team (main reason), reluctant to GUI, OSD shiny mumbo-jumbo pollution, reluctant to rememberable pan&scan screen-space wasting, responsible for the DXVA Global Warming-aware push without sacrificing the foreign art subtitle revealer, responsible for the endangered frame saving Renderer and so forth, should also consider implementing a green progress/seek bar as described above.
According to IPCC quoting a thoroughly study made by the world's most talented scientists over the last few minutes, such implementation would save 12.1g of fossil fuel per video played per normal user, or 22.4Kg of fossil fuel per video played per angry german kid (http://www.youtube.com/watch?v=PbcctWbC8Q0).

[X-Dron]

Of course I tried it, otherwise I wouldn't be writing this. That's the exact thing I'm opposing to. I do not want that. If I open a movie that's is in, let's say, a 10 movie folder, and it doesn't have chapters I don't want those buttons active so I can accidentally press them (like I've done today) while I'm in the middle of Iron Man's best scene action to see Transformers' beginning.

I understand this may be useful for folders containing audio files, but for those containing video files is awful IMO. If you guys plan to keep it this way, an option to disable it would be great.

Done in Changeset [1748]. See "Tweaks" settings.

[STaRGaZeR]

Done in Changeset [1748]. See "Tweaks" settings.

Thanks!

[clsid]

A random idea I came up with:
Create a small DLL with functions to detect whether a user has a DXVA capable graphics card. This DLL could be use by the MPC installer to determine the optimal default configuration. For example, disabling DXVA by default on non-compatible systems could speed up video loading time by a few milliseconds.
The DLL could also be used to detect the brand and model of a card, and perhaps even the installed version of the driver. With such information we could inform users about known bad drivers (or hardware) that has known problems with DXVA or any of the video renderers in MPC.

[Razoola]

A random idea I came up with:
Create a small DLL with functions to detect whether a user has a DXVA capable graphics card. This DLL could be use by the MPC installer to determine the optimal default configuration. For example, disabling DXVA by default on non-compatible systems could speed up video loading time by a few milliseconds.
The DLL could also be used to detect the brand and model of a card, and perhaps even the installed version of the driver. With such information we could inform users about known bad drivers (or hardware) that has known problems with DXVA or any of the video renderers in MPC.

It sounds like a good idea but maybe it would be even better to have it as an option within MPC-HC. Say something like 'DXVA hardware test'.

First though I think ticket 344 needs a fixing as that will have a direct relation on this for those of us with multiple GPU's. :)

[clsid]

The advantage of doing it in the installer is that it will be automated. Any unnecessary user interaction is undesirable, certainly in mass-deployment.

But having both would be nice of course.

[tetsuo55]

please open a ticket on the tracker, some devs have expressed interest in seeing this happen

[JonasNo]

Then why is ffmpeg in all MPC-HC builds on xvidvideo.ru build with 4.5.0?

I don't know, ask him, maybe he knows how to avoid the bugs or maybe he go a computer that this experimental build happens to work on.

For example, on my system gcc v4.5 just crashes, hangs and/or outputs bad errors and refuses to compile.

[p0w3rh0u5e]

Before 3D support can be added, a 3D SOFTWARE DECODER, SPLITTER etc needs to be made available (Cyberlink first?!). "Native" support for such thing is far fetched (Nokia helped a little).

Mmmh... from my understanding, to support 3d via mkv, we "only" need to enhance the mkv-splitter, run the two seperate videostreams through two instances of a decoder and deliver the two decoded streams to a renderer that uses nvidia's or ati's stereoscopic-processing. Later, for 3d-BluRay and MVC, only the m2ts-splitter and AVC-decoder needs to be enhanced. There is no need to build a own postprocessing-engine for 3d.


The thing is that you cannot put such high hopes, when interlaced video is so hard to deal with, even at this time. Mastering interlaced content is the starting point. And to make MPC-HC a true stereo player, it would need a new implementation and integration of the shader engine. With the current one, you can only do anaglyph manipulation for progressive 3d video content. Stuff like converting on-the-fly between color pairs, swap them, and with some work converting up/down or left/right to anaglyph. I should check the new H.264 MVC standard, could have built-in anaglyph conversion, or at least easier to be done than MPEG-2 interlaced 3D content.

You're right, supporting those old methods would need much more work, but to be honest, i think it would be a waste of time. Interlaced and over/under etc. are the past, nobody will use it when there is the possibility to use two seperate streams in a container or encode your material into such a new h264-stream. Old content will be re-released or can be converted to new standards. So, my opinion, forget those old methods. ;)

[tetsuo55]

I don't know, ask him, maybe he knows how to avoid the bugs or maybe he go a computer that this experimental build happens to work on.

For example, on my system gcc v4.5 just crashes, hangs and/or outputs bad errors and refuses to compile.alexins seems to prefer bleeding edge versions of all his software.

[ar-jar]

general MPC-HC question:

what exactly does the option 'reinitialize when display changing' does when you enable or disable it?

This means that you create a new DirectX device for the new screen. It *could* mean that you get better performance but I'm not sure. I know that a window spanning over two screen makes a dent in performance. There is a bug somewhere causing the screen go black when playing DXVA with the internal decoder and having checked that option. -A

[ar-jar]

A debug exception was introduced between builds 1738 and 1739 I believe. This makes it hard to run debug builds. I have traced it down to CVMROSD DisplayMessage, line m_MainFont.CreatePointFont(m_FontSize*10, m_OSD_Font) if my debugger is to trust. Haven't been able to pin down the exact cause since the code looks ok to me. The regular build seems to ignore whatever it is that goes wrong. Same behavior on two different computers. -A

[ar-jar]

A debug exception was introduced between builds 1738 and 1739 I believe. This makes it hard to run debug builds. I have traced it down to CVMROSD DisplayMessage, line m_MainFont.CreatePointFont(m_FontSize*10, m_OSD_Font) if my debugger is to trust. Haven't been able to pin down the exact cause since the code looks ok to me. The regular build seems to ignore whatever it is that goes wrong. Same behavior on two different computers. -A

Further evidence: the player works if you deselect OSD.