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

[v0lt]

MF MPEG-4 Source works better with sound. For example WMP12 without any problems playing MOVs with "16-bit little-endian PCM audio" ('sowt'), and MPC-HC this can not.

QTLite solves the problem, but adds the brakes.

[Sebastiii]

You can load the MPLS for the main movie and that'll work just fine. I use eac3to to give me the most likely candidates. If you do an open DVD on the BD, MPC-HC will make a best guess, but, as you've seen it doesn't always get it right.

Hi SamuriHL,

I'm try to see why on certain BD movies and .mpls the movie timeline is not correct, i have open a ticket for that too.
Did you see this pb too ? and maybe hard to fix it, because i can't upload a sample because we need entire movie.

http://sourceforge.net/apps/trac/mpc-hc/ticket/601

Thanks,
Seb.

[tetsuo55]

Should not be too hard to find all code locations where a thread is created (e.g. search for CreateThread, _beginthread(ex) and AfxBeginThread) and then to make sure that all of these secondary threads are properly run through before allowing the process to close down. Of course it needs a bit of programming time and maybe one of the devs needs to get his hands dirty a bit. But it's not really rocket science for any programmer with experience in multi threaded programming.Thanks, using this information i was able to track down that we are using a dangerous call, namely TerminateThread, but not all the time.

After digging deeper and deeper into google i found the page below:
http://members.cox.net/doug_web/threads.htm

This pageexplains how MFC/Win32 applications should deal with threading and exit calls.

We need a dev to use that document and update the existing thread handling.
Since this is almost exactly the problem ar-jar ran into with renderer destruction it might also fix more bugs than just the crash on exit.

[SamuriHL]

Hi SamuriHL,

I'm try to see why on certain BD movies and .mpls the movie timeline is not correct, i have open a ticket for that too.
Did you see this pb too ? and maybe hard to fix it, because i can't upload a sample because we need entire movie.

http://sourceforge.net/apps/trac/mpc-hc/ticket/601

Thanks,
Seb.

I hadn't noticed that bug but I don't always pay attention to the timeline, either. Interesting. Next time I'm in MPC-HC I'll see if I can repro it just to give you confirmation. My list of tasks is growing, however. :)

[SamuriHL]

Thanks, using this information i was able to track down that we are using a dangerous call, namely TerminateThread, but not all the time.

After digging deeper and deeper into google i found the page below:
http://members.cox.net/doug_web/threads.htm

This pageexplains how MFC/Win32 applications should deal with threading and exit calls.

We need a dev to use that document and update the existing thread handling.
Since this is almost exactly the problem ar-jar ran into with renderer destruction it might also fix more bugs than just the crash on exit.

Nice work! Hopefully someone has time to look at that soon.

[Virtual_ManPL]

If you guys working on terminating threads, don't forget about bug in MPC-HC causing player to works in background, even if you close it before ;p

[madshi]

Thanks, using this information i was able to track down that we are using a dangerous call, namely TerminateThread, but not all the time.
Yes, TerminateThread should be avoided, if possible in any way. If you use TerminateThread, there could be synchronization objects which will stay locked forever, resulting in deadlocks immediately, or some time later...

After digging deeper and deeper into google i found the page below:
http://members.cox.net/doug_web/threads.htm

This pageexplains how MFC/Win32 applications should deal with threading and exit calls.
Yep, it also confirms what I said earlier:

"In general, it's a bad idea to allow secondary threads to continue executing as the application is shutting down."

[tetsuo55]

@madshi Yes you where right.
Could you take a look ? http://sourceforge.net/apps/trac/mpc-hc/browser/trunk/src/apps/mplayerc/MainFrm.cpp

If you guys working on terminating threads, don't forget about bug in MPC-HC causing player to works in background, even if you close it before ;pthis behaviour is caused by the same problem

[madshi]

Could you take a look ? http://sourceforge.net/apps/trac/mpc-hc/browser/trunk/src/apps/mplayerc/MainFrm.cpp
It seems that the thread responsible for the DirectShow graph is "violently" terminated in 2 different situations, if it doesn't react as expected in a timeframe of 5 seconds. I don't know the MPC HC code well enough to say anything about why the thread could possibly fail to reply quickly enough. But since the shutdown problem seems to be reproduceable with MSVC++ 2010, it might be worth a try to (in a test version) simply replace those two TerminateThread calls with a simple MessageBox call, just to check whether the MSVC++ 2010 shutdown problem is related to those two TerminateThread calls at all.

[Sebastiii]

I hadn't noticed that bug but I don't always pay attention to the timeline, either. Interesting. Next time I'm in MPC-HC I'll see if I can repro it just to give you confirmation. My list of tasks is growing, however. :)

Great :) if you can test :)
If it was only a display wrong lenght is not a big deal but it make an issue for seeking :(

[SamuriHL]

I'll see what I can do. I may not get to it today though.

[oddball]

If anyone needs me to test a build let me know. ;)

[Sebastiii]

I'll see what I can do. I may not get to it today though.

:p not a pb :)
Cheers,
Seb.

[mariush]

While you're there, can you also please check the file handle leak? I've reported it a while ago.

double click video, it plays... drag another file while the first video still plays, it goes playing... but now the first file is locked (handle is still open, can't delete file for example) until player is closed. Basically the file before the last one open is always locked.

[xiulet]

hi , all ;)

i have a simple question:
what are the benefits of compilation (or coding) MPC-HC on MSVC2010 vs. MSVC2008? if there are any.


thank you, so much.

adéu.

[Octo-puss]

Tetsuo, have you considered ninjaing Madshi into the development team? :)

[XhmikosR]

hi , all ;)

i have a simple question:
what are the benefits of compilation (or coding) MPC-HC on MSVC2010 vs. MSVC2008? if there are any.


thank you, so much.

adéu.

Just use MSVC 2008. MSVC 2010 binaries feel faster, at least for me. On the other hand size increases and a few bugs are more visible in MSVC 2010 builds.

[xiulet]

Just use MSVC 2008. MSVC 2010 binaries feel faster, at least for me. On the other hand size increases and a few bugs are more visible in MSVC 2010 builds.

ok , thanks :)


see you.

adéu.

[madshi]

Tetsuo, have you considered ninjaing Madshi into the development team? :)
Thanks, but I'm much too busy with my own projects...

[tetsuo55]

we tried that ;)

[Aleksoid1978]

I make little test - VS 2010 version MPC-HC does not fall because "TerminateThread".
Need more testing and debug.

[SamuriHL]

I make little test - VS 2010 version MPC-HC does not fall because "TerminateThread".
Need more testing and debug.

If you're able to debug it, set a breakpoint and step through it. If you're not, just add messages so you can trace it. If you can narrow down exactly where it's crashing, someone will figure out a way to fix it.

[Aleksoid1978]

I think that crash on exit in VS2010 is linked with sizecbar and that it's not worked in MPC-HC + VS2010.

[Aleksoid1978]

I spent the extra tests and was proved right - in 2010 build errors due to sizecbar component.
I disable code with create control based on sizecbar ... and MPC-HC not crash on exit;

Here the listing of control:


CPlayerSubresyncBar m_wndSubresyncBar;
CPlayerPlaylistBar m_wndPlaylistBar;
CPlayerCaptureBar m_wndCaptureBar;
CPlayerNavigationBar m_wndNavigationBar;
CPlayerShaderEditorBar m_wndShaderEditorBar;
CEditListEditor m_wndEditListEditor;


This means that the error is not in the code of MPC-HC

P.S. sizecbar don't work in VS2010 build, need rewrite or change to another component.

[vBm]

Good thing that at least we now know the cause for the crashes :)
Maybe contact Cristi with the problem ... might be an easy fix for him.

[fastplayer]

I spent the extra tests and was proved right - in 2010 build errors due to sizecbar component.
Can you try with the latest update? Right now 2.43 (from August 2000) is in trunk, 2.45 (April 2010) is the latest:
http://www.datamekanix.com/sizecbar/
The homepage says that VS2010 is supported...

Changes since 2.43:

CSizingControlBar 2.45 / Apr 16, 2010

* Maintenance release, fixing some incompatibilities with the newer versions of Visual C++ and MFC. While the sample projects are still VC6, they should now convert and build with VC 2003, 2005, 2008, and 2010.
* Changed the return value of OnNcHitTest() from UINT to LRESULT.
* Removed the use of afxChNil.

CSizingControlBar 2.44 / Mar 31, 2002

* This is a maintenance release. A couple of bug fixes:
* Fixed the bug that caused the bars using CSCBMiniDockFrameWnd under Windows XP to be shrinked when floating. Thanks to DJ for first pointing this out and to everyone that suggested fixes.
* Removed an unnecessary call to ReleaseDC in CSizingControlBar::OnNCPaint(), because the CWindowDC destructor releases the device context handle. Thanks to Pat Kusbel, and then Aleksey Malyshev, who both described the bug and sent the fix.

[roytam1]

latest version of CSizingControlBar is 2.4.5
http://www.datamekanix.com/sizecbar/

[Aleksoid1978]

Can you try with the latest update? Right now 2.43 (from August 2000) is in trunk, 2.45 (April 2010) is the latest:
http://www.datamekanix.com/sizecbar/
The homepage says that VS2010 is supported...

Changes since 2.43:

Try - nothing don't change, don't work with VS2010 build.

[Sebastiii]

Hi :)

Aleksoid has made a very good fix for the BD seamless branching movie length problem in the MPC MPEG splitter
But there now is a problem with seeking, i have made an ugly hack by commenting out some lines of code, someone will have to look at it because in its current state it probably cannot be committed with this patch however, seeking works for both branching and non branching.
You can find the link for Filter splitter MPEG in Debug version and Release version and the patch file.

Source .diff file (http://www.mediafire.com/?k5xz4joblc4c2pu)
Release Filter (http://www.mediafire.com/?xcwpir0ng82h5r9)
Debug Filter (http://www.mediafire.com/?cn23ued7i0ao2iq)
MPC-HC Build x86 (http://www.mediafire.com/?66ry0pr719w4ubc)

Can you test on your side ?
Cheers,
Seb.

[burfadel]

Just use MSVC 2008. MSVC 2010 binaries feel faster, at least for me. On the other hand size increases and a few bugs are more visible in MSVC 2010 builds.

The pixel shader editor doesn't work with the MSVC 2010 build (does in the 2008 build).

I've probably missed it if someone has mentioned that already!...

What difference does it make anyways, if you select PS 3.0 over PS 2.0 for example? (for filters that are currently PS 2.0).

Also, if there is a difference since the latest Directx library is required, why not PS 4.0 and higher?

[XhmikosR]

http://xhmikosr.1f0.de/mpc-hc_msvc2010/BIG_FAT_WARNING.txt

[fastplayer]

http://xhmikosr.1f0.de/mpc-hc_msvc2010/BIG_FAT_WARNING.txt
Do the playlist and shader editors use the sizecbar library (see couple of posts above)? If so, does updating it to version 2.45 fix the issues?

[XhmikosR]

Yes and no. I've already sent an email to the author of sizecbar asking for his help.

[fastplayer]

Yes and no. I've already sent an email to the author of sizecbar asking for his help.
Good idea!
Although, I'm kinda skeptical that UI components are the root cause of this nasty VS2010 issue...

[XhmikosR]

Aleksoid confirmed it too. So the only thing we can do right now is wait unless you or someone else can find another way to fix these issues.

[romanesque]

Is there any place where I can download the latest builds? Thank you. :)

[Octo-puss]

One post above yours... Even I can see it.

[Aleksoid1978]

Is there any place where I can download the latest builds? Thank you. :)

http://xhmikosr.1f0.de/

[romanesque]

Thanks, Aleksoid ... really appreciate. :D

[XhmikosR]

What version is the last that didn't crash? Obviously I cannot reproduce it, so provide more info, like filters used and a sample.

EDIT: Alright, I managed to reproduce it, I'll take a look.

[sneaker_ger]

I have the same problem using 2200 (tried both 32 and 64 bit).

/edit: VS2010 builds do not crash on loading a file. (but on exit)

[XhmikosR]

It's r2200.

[Aleksoid1978]

MPC-HC SVN 2200 crashes when trying to play anything, MKV, MP4, AVI etc. Tested on Intel Core 2, Core iX and Athlon XP, they all crash when trying to play any videos.

http://img231.imageshack.us/img231/8466/crashd.png

http://img8.imageshack.us/img8/7365/crash2.png

I think is't last ffmpeg update bug.

Revert in 2201 :)

[sneaker_ger]

Problem with x64 builds:
When I click "modify" (Windows 7) under "options">"formats" a second "options" window opens. Changing the settings in there works so it's more of a cosmetic problem.

[sneaker_ger]

I think is't last ffmpeg update bug.

In case you didn't see my edit: the VS2010 builds do not crash. (except on exit, but that seems to be known already)

[BibiCode]

Last try before giving up....

I am having jitter playback using mpc HC with DXVA and EVR Custom.

Whenever the two lines are jaggy I have smooth playback, whenever there is a spike, the video halts for a split second, enough to see jitter. The audio stays good.

This is specialy visible on panoramic images. I tried ALL MPC Settings and combinations with nvidia panel and the mpc own settings to no effect.

I uploaded two screenshots for examples

Please HELP !

I have WIN7 core i5

[Aleksoid1978]

Last try before giving up....

I am having jitter playback using mpc HC with DXVA and EVR Custom.

Whenever the two lines are jaggy I have smooth playback, whenever there is a spike, the video halts for a split second, enough to see jitter. The audio stays good.

This is specialy visible on panoramic images. I tried ALL MPC Settings and combinations with nvidia panel and the mpc own settings to no effect.

I uploaded two screenshots for examples

Please HELP !

I have WIN7 core i5

Maybe you use VSYNC - disable all VSYNC and try again.

[Momber]

Sigh... I would so like to be able to watch my BBC HD caps in style... that is, without corruption or jerkyness.
Why do the MPC developers hate the BBC so much?

[BibiCode]

Maybe you use VSYNC - disable all VSYNC and try again.

Hey man, thanks for replying. I have tried EVERYTHING in those settings dialogs.

Same thing happens.

This happens quite randomly, and even with madVR renderer Also.

Sometimes it happens after a heavy scene and stays like that the hole movie.

The weird thing about it, is that when it starts, it happens every Second. It's like i know it's gonna happen. You just have to look at your watch, each second, you have a spike

[Aleksoid1978]

Hey man, thanks for replying. I have tried EVERYTHING in those settings dialogs.

Same thing happens.

This happens quite randomly, and even with madVR renderer Also.

Sometimes it happens after a heavy scene and stays like that the hole movie.

The weird thing about it, is that when it starts, it happens every Second. It's like i know it's gonna happen. You just have to look at your watch, each second, you have a spike

I say about Vsync setting in popup menu:

http://s53.radikal.ru/i139/1008/36/5da3428b0497t.jpg (http://radikal.ru/F/s53.radikal.ru/i139/1008/36/5da3428b0497.png.html)