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

[SamuriHL]

SIGH. Nope. I'm still getting this kind of stuff:


60>MpaDecFilter.cpp(1568): error C2065: 'm_ps2_state' : undeclared identifier
60>MpaDecFilter.cpp(1568): error C2228: left of '.sync' must have class/struct/u
nion
60> type is ''unknown-type''
60>MpaDecFilter.cpp(1570): error C2065: 'm_ps2_state' : undeclared identifier
60>MpaDecFilter.cpp(1570): error C2228: left of '.sync' must have class/struct/u
nion
60> type is ''unknown-type''
60>MpaDecFilter.cpp(1575): error C2065: 'm_ps2_state' : undeclared identifier
60>MpaDecFilter.cpp(1575): fatal error C1003: error count exceeds 100; stopping
compilation
60>
60>Build FAILED.


Tons of errors above that saying pretty much the same thing.

[clsid]

Fixed that too.

[SamuriHL]

Thanks. Will try again. :)

[SamuriHL]

P.S. I haven't pulled yet, but, have you run into this yet?


67>c:\dev\projects\mpc-hc\src\filters\./transform/MpaDecFilter/MpaDecFilter.h(13
0): error C2143: syntax error : missing ';' before '*'
67>c:\dev\projects\mpc-hc\src\filters\./transform/MpaDecFilter/MpaDecFilter.h(13
0): error C4430: missing type specifier - int assumed. Note: C++ does not suppor
t default-int
67>c:\dev\projects\mpc-hc\src\filters\./transform/MpaDecFilter/MpaDecFilter.h(13
0): error C4430: missing type specifier - int assumed. Note: C++ does not suppor
t default-int
67>c:\dev\projects\mpc-hc\src\filters\./transform/MpaDecFilter/MpaDecFilter.h(13
1): error C2143: syntax error : missing ';' before '*'
67>c:\dev\projects\mpc-hc\src\filters\./transform/MpaDecFilter/MpaDecFilter.h(13
1): error C4430: missing type specifier - int assumed. Note: C++ does not suppor
t default-int
67>c:\dev\projects\mpc-hc\src\filters\./transform/MpaDecFilter/MpaDecFilter.h(13
1): error C4430: missing type specifier - int assumed. Note: C++ does not suppor
t default-int
67>c:\dev\projects\mpc-hc\src\filters\./transform/MpaDecFilter/MpaDecFilter.h(13
2): error C2143: syntax error : missing ';' before '*'
67>c:\dev\projects\mpc-hc\src\filters\./transform/MpaDecFilter/MpaDecFilter.h(13
2): error C4430: missing type specifier - int assumed. Note: C++ does not suppor
t default-int
67>c:\dev\projects\mpc-hc\src\filters\./transform/MpaDecFilter/MpaDecFilter.h(13
2): error C4430: missing type specifier - int assumed. Note: C++ does not suppor
t default-int

[SamuriHL]

Brilliant. The last update did it. Thank you so much for fixing this!

[SamuriHL]

WOOPS. It wasn't done. :(


67>..\..\thirdparty\mfc\afxglobals.cpp(44): error C2039: 'm_hinstD2DDLL' : is no
t a member of 'AFX_GLOBAL_DATA'
67> C:\Program Files (x86)\Microsoft Visual Studio 10.0\VC\atlmfc\inclu
de\afxglobals.h(111) : see declaration of 'AFX_GLOBAL_DATA'
67>..\..\thirdparty\mfc\afxglobals.cpp(45): error C2039: 'm_hinstDWriteDLL' : is
not a member of 'AFX_GLOBAL_DATA'
67> C:\Program Files (x86)\Microsoft Visual Studio 10.0\VC\atlmfc\inclu
de\afxglobals.h(111) : see declaration of 'AFX_GLOBAL_DATA'
67>..\..\thirdparty\mfc\afxglobals.cpp(47): error C2143: syntax error : missing
';' before '*'
67>..\..\thirdparty\mfc\afxglobals.cpp(47): error C4430: missing type specifier
- int assumed. Note: C++ does not support default-int
67>..\..\thirdparty\mfc\afxglobals.cpp(47): error C2039: 'm_pDirect2dFactory' :
is not a member of 'AFX_GLOBAL_DATA'
67> C:\Program Files (x86)\Microsoft Visual Studio 10.0\VC\atlmfc\inclu
de\afxglobals.h(111) : see declaration of 'AFX_GLOBAL_DATA'
67>..\..\thirdparty\mfc\afxglobals.cpp(47): error C4430: missing type specifier
- int assumed. Note: C++ does not support default-int
67>..\..\thirdparty\mfc\afxglobals.cpp(48): error C2143: syntax error : missing
';' before '*'
67>..\..\thirdparty\mfc\afxglobals.cpp(48): error C4430: missing type specifier
- int assumed. Note: C++ does not support default-int
67>..\..\thirdparty\mfc\afxglobals.cpp(48): error C2039: 'm_pWriteFactory' : is
not a member of 'AFX_GLOBAL_DATA'
67> C:\Program Files (x86)\Microsoft Visual Studio 10.0\VC\atlmfc\inclu
de\afxglobals.h(111) : see declaration of 'AFX_GLOBAL_DATA'
67>..\..\thirdparty\mfc\afxglobals.cpp(48): error C4430: missing type specifier
- int assumed. Note: C++ does not support default-int
67>..\..\thirdparty\mfc\afxglobals.cpp(49): error C2143: syntax error : missing
';' before '*'
67>..\..\thirdparty\mfc\afxglobals.cpp(49): error C4430: missing type specifier
- int assumed. Note: C++ does not support default-int
67>..\..\thirdparty\mfc\afxglobals.cpp(49): error C2039: 'm_pWicFactory' : is no
t a member of 'AFX_GLOBAL_DATA'
67> C:\Program Files (x86)\Microsoft Visual Studio 10.0\VC\atlmfc\inclu
de\afxglobals.h(111) : see declaration of 'AFX_GLOBAL_DATA'
67>..\..\thirdparty\mfc\afxglobals.cpp(49): error C4430: missing type specifier
- int assumed. Note: C++ does not support default-int
67>..\..\thirdparty\mfc\afxglobals.cpp(51): error C2039: 'm_pfD2D1MakeRotateMatr
ix' : is not a member of 'AFX_GLOBAL_DATA'
67> C:\Program Files (x86)\Microsoft Visual Studio 10.0\VC\atlmfc\inclu
de\afxglobals.h(111) : see declaration of 'AFX_GLOBAL_DATA'
67>..\..\thirdparty\mfc\afxglobals.cpp(51): error C2146: syntax error : missing
';' before identifier 'm_pfD2D1MakeRotateMatrix'
67>..\..\thirdparty\mfc\afxglobals.cpp(51): error C4430: missing type specifier
- int assumed. Note: C++ does not support default-int
67>..\..\thirdparty\mfc\afxglobals.cpp(51): error C4430: missing type specifier
- int assumed. Note: C++ does not support default-int
67>..\..\thirdparty\mfc\afxglobals.cpp(53): error C2039: 'm_bD2DInitialized' : i
s not a member of 'AFX_GLOBAL_DATA'
67> C:\Program Files (x86)\Microsoft Visual Studio 10.0\VC\atlmfc\inclu
de\afxglobals.h(111) : see declaration of 'AFX_GLOBAL_DATA'
67>..\..\thirdparty\mfc\afxglobals.cpp(915): error C3861: 'ReleaseD2DRefs': iden
tifier not found
67>..\..\thirdparty\mfc\afxglobals.cpp(1463): error C2039: 'InitD2D' : is not a
member of 'AFX_GLOBAL_DATA'
67> C:\Program Files (x86)\Microsoft Visual Studio 10.0\VC\atlmfc\inclu
de\afxglobals.h(111) : see declaration of 'AFX_GLOBAL_DATA'
67>..\..\thirdparty\mfc\afxglobals.cpp(1463): error C2065: 'D2D1_FACTORY_TYPE' :
undeclared identifier
67>..\..\thirdparty\mfc\afxglobals.cpp(1463): error C2146: syntax error : missin
g ')' before identifier 'd2dFactoryType'
67>..\..\thirdparty\mfc\afxglobals.cpp(1463): error C2059: syntax error : ')'
67>..\..\thirdparty\mfc\afxglobals.cpp(1464): error C2143: syntax error : missin
g ';' before '{'
67>..\..\thirdparty\mfc\afxglobals.cpp(1464): error C2447: '{' : missing functio
n header (old-style formal list?)
67>..\..\thirdparty\mfc\afxglobals.cpp(1521): error C2039: 'ReleaseD2DRefs' : is
not a member of 'AFX_GLOBAL_DATA'
67> C:\Program Files (x86)\Microsoft Visual Studio 10.0\VC\atlmfc\inclu
de\afxglobals.h(111) : see declaration of 'AFX_GLOBAL_DATA'
67>..\..\thirdparty\mfc\afxglobals.cpp(1530): error C2227: left of '->Release' m
ust point to class/struct/union/generic type
67> type is 'int *'
67>..\..\thirdparty\mfc\afxglobals.cpp(1536): error C2227: left of '->Release' m
ust point to class/struct/union/generic type
67> type is 'int *'
67>..\..\thirdparty\mfc\afxglobals.cpp(1542): error C2227: left of '->Release' m
ust point to class/struct/union/generic type
67> type is 'int *'
67> winmdi.cpp
67> AppSettings.cpp
67> AuthDlg.cpp
67>c:\dev\projects\mpc-hc\src\apps\mplayerc\AppSettings.h(442): error C2229: cla
ss 'CAppSettings' has an illegal zero-sized array
67>c:\dev\projects\mpc-hc\src\apps\mplayerc\AppSettings.h(555): error C2229: cla
ss 'CAppSettings' has an illegal zero-sized array
67> BaseGraph.cpp


I guess I didn't have enough patience. :D

[SamuriHL]

Yea, this isn't good. Even reverting the header to include all the filters, the build is still broken on a clean full build.

[clsid]

Perhaps your visual studio isn't up-to-date? You need SP1.

[SamuriHL]

Oh son of a! Sometimes I don't like this laptop. I watched WU update it to SP1 a few weeks ago. What I DIDN'T apparently notice is that the damn update failed. Ok, off to fix VC2010 and get it up to SP1. This blows!

[molitar]

@Molitar: reset the registry settings: http://sourceforge.net/apps/trac/mpc-hc/wiki/Troubleshooting_guidelines

Thanks did that.. and someone stated them are CCCP defaults but them are the same defaults in the original MPC non cinema version because I tested dropped in another folder and tested and the defaults were just about the same as CCCP pack except for two more items checked in that list. But like I stated dropping the new version in rewrote all the defaults to being all checked. Good to know how to clear registry of what it changes when first ran..

[betaking]

CLSID.thansk you work but when i install last MSVC2008 sp1 after latest windows updates i can not compiler last svn!
输出窗口 正在编译...
OggFile.cpp
stdafx.cpp
OggSplitter.cpp
.\OggSplitter.cpp(697) : error C2668: “abs”: 对重载函数的调用不明确
C:\Program Files\Microsoft Visual Studio 9.0\VC\include\math.h(539): 可能是“long double abs(long double)”
C:\Program Files\Microsoft Visual Studio 9.0\VC\include\math.h(491): 或 “float abs(float)”
C:\Program Files\Microsoft Visual Studio 9.0\VC\include\math.h(487): 或 “double abs(double)”
C:\Program Files\Microsoft Visual Studio 9.0\VC\include\math.h(485): 或 “long abs(long)”
C:\Program Files\Microsoft Visual Studio 9.0\VC\include\stdlib.h(380): 或 “int abs(int)”
试图匹配参数列表“(REFERENCE_TIME)”时
结果 生成日志保存在“file://C:\mpc-hc\bin\obj\Release_Win32\OggSplitter\BuildLog.htm”
OggSplitter - 1 个错误，0 个警告
输出窗口 正在编译...
stdafx.cpp
正在编译...
MatroskaFile.cpp
MatroskaMuxer.cpp
.\MatroskaMuxer.cpp(667) : error C2668: “abs”: 对重载函数的调用不明确
C:\Program Files\Microsoft Visual Studio 9.0\VC\include\math.h(539): 可能是“long double abs(long double)”
C:\Program Files\Microsoft Visual Studio 9.0\VC\include\math.h(491): 或 “float abs(float)”
C:\Program Files\Microsoft Visual Studio 9.0\VC\include\math.h(487): 或 “double abs(double)”
C:\Program Files\Microsoft Visual Studio 9.0\VC\include\math.h(485): 或 “long abs(long)”
C:\Program Files\Microsoft Visual Studio 9.0\VC\include\stdlib.h(380): 或 “int abs(int)”
试图匹配参数列表“(REFERENCE_TIME)”时
结果 生成日志保存在“file://C:\mpc-hc\bin\obj\Release_Win32\MatroskaMuxer\BuildLog.htm”
MatroskaMuxer - 1 个错误，0 个警告
PPageFileInfoDetails.cpp
.\PPageFileInfoDetails.cpp(158) : error C3861: “StrFormatByteSizeW”: 找不到标识符
Playlist.cpp
.\Playlist.cpp(271) : error C3861: “StrCmpLogicalW”: 找不到标识符
.\MainFrm.cpp(4972) : error C3861: “StrFormatByteSizeW”: 找不到标识符
.\MainFrm.cpp(11702) : error C3861: “StrCmpLogicalW”: 找不到标识符
结果 生成日志保存在“file://C:\mpc-hc\bin\obj\Release_Win32\mpc-hc\BuildLog.htm”
mpc-hc - 4 个错误，18 个警告
PS: vs2010sp1 compiler is ok!

[SamuriHL]

My MSVC 2010 install is borked. :( I'm gonna have to completely uninstall and reinstall it. I got the SP1 DVD image from MSDN and it also failed to install and is in the process of rolling back. SIGH. I guess I'll get back to building this at some point in the near future. cough cough. :(

[SamuriHL]

WOO HOO!

MPC-HC's compilation started on Wed 05/04/2011-22:28:00.20
and completed on Wed 05/04/2011-22:48:12.47

bin10\MPC-HomeCinema.1.5.2.3080.x86.MSVC2010.exe

Hurray. That's with only 1 filter selected. Thank you, clsid! Looks like I'm back in business.

[clsid]

This patch might fix the MSVC2008 build:

Index: include/stdafx_common_afx.h
===================================================================
--- include/stdafx_common_afx.h (revision 3078)
+++ include/stdafx_common_afx.h (working copy)
@@ -1,19 +1,15 @@
// turns off MFC's hiding of some common and often safely ignored warning messages
#define _AFX_ALL_WARNINGS

-#include <afxwin.h> // MFC core and standard components
-#include <afxext.h> // MFC extensions
-
-#include <afxdisp.h> // MFC Automation classes
-
-#ifndef _AFX_NO_OLE_SUPPORT
-#include <afxdtctl.h> // MFC support for Internet Explorer 4 Common Controls
+#if (_MSC_VER < 1600)
+#include <afx.h>
#endif
-#ifndef _AFX_NO_AFXCMN_SUPPORT
-#include <afxcmn.h> // MFC support for Windows Common Controls
-#endif // _AFX_NO_AFXCMN_SUPPORT

+#include <afxwin.h> // MFC core and standard components
+
+#if (_MSC_VER >= 1600)
#include "afxpriv.h"
#include "afxole.h"
#include "..\src\mfc\oleimpl2.h"
#include "..\src\mfc\afximpl.h"
+#endif
Index: include/stdafx_common_afx2.h
===================================================================
--- include/stdafx_common_afx2.h (revision 0)
+++ include/stdafx_common_afx2.h (revision 0)
@@ -0,0 +1,22 @@
+// turns off MFC's hiding of some common and often safely ignored warning messages
+#define _AFX_ALL_WARNINGS
+
+#include <afxwin.h> // MFC core and standard components
+#include <afxext.h> // MFC extensions
+#include <afxdisp.h> // MFC Automation classes
+
+#ifndef _AFX_NO_OLE_SUPPORT
+#include <afxdtctl.h> // MFC support for Internet Explorer 4 Common Controls
+#endif
+
+#ifndef _AFX_NO_AFXCMN_SUPPORT
+#include <afxcmn.h> // MFC support for Windows Common Controls
+#endif
+
+#include "afxole.h"
+
+#if (_MSC_VER >= 1600)
+#include "afxpriv.h"
+#include "..\src\mfc\oleimpl2.h"
+#include "..\src\mfc\afximpl.h"
+#endif
Index: src/apps/mplayerc/stdafx.h
===================================================================
--- src/apps/mplayerc/stdafx.h (revision 3078)
+++ src/apps/mplayerc/stdafx.h (working copy)
@@ -26,7 +26,7 @@
#include "../../../DSUtil/SharedInclude.h"
#include "../../../../include/stdafx_common.h"
#undef _ATL_CSTRING_EXPLICIT_CONSTRUCTORS
-#include "../../../../include/stdafx_common_afx.h"
+#include "../../../../include/stdafx_common_afx2.h"
#include "../../../../include/stdafx_common_dshow.h"

#include <afxdlgs.h>
Index: src/filters/transform/VSFilter/stdafx.h
===================================================================
--- src/filters/transform/VSFilter/stdafx.h (revision 3078)
+++ src/filters/transform/VSFilter/stdafx.h (working copy)
@@ -22,6 +22,6 @@

#include "../../../DSUtil/SharedInclude.h"
#include "../../../../include/stdafx_common.h"
-#include "../../../../include/stdafx_common_afx.h"
+#include "../../../../include/stdafx_common_afx2.h"
#include "../../../../include/stdafx_common_dshow.h"
#include "../../../DSUtil/DSUtil.h"

Please also test MSVC2010 build after this patch. I haven't got the time right now to do full builds to test.

[betaking]

thanks can you put this patch to svn? thanks!

[SamuriHL]

It needs to be fully tested with both a 2008 and 2010 compile first. That's what clsid was asking people to do. I'm off to bed so I can't test the 2010 part of that right now.

[SamuriHL]

Ok, clsid, I was able to build it again with no filters now. That's so awesome! Thanks for doing this for us. madshi's likely to be quite happy, as well. I wasn't able to test your patch to fix the 2008 build with 2010 but if no one's done it by the morning I'll take a look.

[Aleksoid1978]

http://aleksoid.tosei.ru/Dist/mpc-hc.exe.3083.7z - Only .exe
http://aleksoid.tosei.ru/Dist/mpc-hc.3083.7z - Complete set.

[betaking]

test clsid patch still have
PPageFileInfoDetails.cpp
.\PPageFileInfoDetails.cpp(158) : error C3861: “StrFormatByteSizeW”: 找不到标识符
Playlist.cpp
.\Playlist.cpp(271) : error C3861: “StrCmpLogicalW”: 找不到标识符
.\MainFrm.cpp(4972) : error C3861: “StrFormatByteSizeW”: 找不到标识符
.\MainFrm.cpp(11702) : error C3861: “StrCmpLogicalW”: 找不到标识符
结果 生成日志保存在“file://C:\mpc-hc\bin\obj\Release_Win32\mpc-hc\BuildLog.htm”
mpc-hc - 4 个错误，18 个警告

[nevcairiel]

Perhaps your visual studio isn't up-to-date? You need SP1.

While installing SP1 is certainly a good idea for everyone, breaking compilation for people that don't have it is just horrible.

I should really switch players, MPC-HC is a lost cause.. no developers, and enough people that claim to be developers breaking things left and right and arguing pointless semantics..

LAV Player, anyone?

[ryrynz]

While installing SP1 is certainly a good idea for everyone, breaking compilation for people that don't have it is just horrible.

I should really switch players, MPC-HC is a lost cause.. no developers, and enough people that claim to be developers breaking things left and right and arguing pointless semantics..

LAV Player, anyone?

Personally I have no problems with MPC-HC although I do think it's interface is severely outdated.
I'm in agreement with you regarding MPC-HC, If you'd like to develop for a media player probably best making your own.
Potplayer seems to be the best choice for a free media player right now.

[mindbomb]

isn't making a player too large an undertaking for a single person?

that being said, i would probably use lav player

[nevcairiel]

isn't making a player too large an undertaking for a single person?

Would probably take a while, but other devs would be welcome to join/help. :rolleyes:

BTW, i'm not saying i will do it. Maybe if some other active developers express some interest, there could be formed some kind of plan, but for now its just a pipe-dream. :p

[madshi]

At the speed you're currently developing you'll soon run out of work, so maybe a new player is not a bad idea... ;) Personally, I don't have enough time for developing or helping to develop a player, too much other work to do.

[ForceX]

isn't making a player too large an undertaking for a single person?

that being said, i would probably use lav player
Not really, AFAIK KMPlayer and later PotPlayer was developed by a single person and you can see that PotPlayer is being feverishly updated now, you just have to dedicate your complete life to that, that's all. Admittedly, a lot of the base for both KMP and PotP seems to be based on the work of MPC(-HC) and I doubt these players would have been this far if not for the existence of MPC. While MPC development does seem to have stagnated a bit, it is still pretty advanced in terms of good video playback capability and still setting standards for others to follow.

As a non-funded open source program, it is as good as it gets.

[nevcairiel]

I'm not talking about the player itself, i still use it and i think its fine. I'm mostly talking about decisions and actions by the people currently trying to maintain it. I, however, have no ambitions to try to maintain and enhance it myself, because the code is just horribly ugly.

Anyhow, this really doesn't belong here.

[neoufo51]

While installing SP1 is certainly a good idea for everyone, breaking compilation for people that don't have it is just horrible.

I should really switch players, MPC-HC is a lost cause.. no developers, and enough people that claim to be developers breaking things left and right and arguing pointless semantics..

LAV Player, anyone?
If you make a LAV Player, I will drop MPC-HC in an instant, and so will the majority of people who use MPC-HC. You'd also win over the people who use Zoomplayer and PotPlayer because they want something high quality that isn't MPC-HC, and your player would be the way to go, without question.

All you need to do is create a player that is simple, minimalistic, high quality, built from scratch and MPC-HC would just die in a matter of months. Sorry to anybody who loves MPC-HC but that's just the truth. It will be a lot of work, but I don't think anybody here wants something fancy-looking.

[tetsuo55]

We don't "need" a new player. Most of the bugs that exist are in filters. The GUI is really the only thing most people want changed.

The reason the project is moving so slow is because we lack a motivated, dedicated developer. Someone who is willing and able to spend 20-30 hours a week on MPC-HC.(all the dev activity you do see is all filter based)

Nevcairiel is right about the source being difficult to maintain. A dedicated dev would focus mostly on rewriting the actual player into a new clean codebase as mpc-hc2.0.

[nevcairiel]

A dedicated dev would focus mostly on rewriting the actual player into a new clean codebase as mpc-hc2.0.

Which is, essentially, a whole new player. :cool:
Loads of other players already use the same filters as MPC-HC (which i'm in the process to replace, anyhow :p)

Here is basically a concise list how i see this would work, if ever.
- A new base player (GUI, configuration, filter management, DirectShow graph building)
- A new, easier and universal way to render subtitles - preferably as a DirectShow filter, not directly hacked into the client.
- MASSIVE cleanup in the renderers (i mean actual cleanup, maybe even rewriting of huge parts, not hacking new "features" in there)
- Evaluation of filters. LAV can mostly replace internal source filters, it can fully replace the internal audio decoder, video is planned - DXVA1/2 decoders need evaluation.

The first step is the most important, just get some new player going that works with stand-alone filters, with standalone renderers, like default VMR9/EVR and madVR. Then cleanup the custom renderers from MPC-HC and try to reuse EVR-CP and VMR9(renderless). This would already give you proper playback of everything.
Next up would be subtitles. Thats probably a more complicated task and will require cooperation with the renderers.

And as the final touch, select open-source filters that should be bundled with the new player.

[tetsuo55]

Yeah, for maintainability we need a new player.
Will that be mpc-hc2.0, lavplayer, or something else, i don't know

[CruNcher]

Potplayer is in those regards another MPC-HC shares the same codebase though with a lot of additions (especially the skinning engine and the enhanced filter management), only still unique is Zoomplayer, but many already prefer Potplayer also Gom Player and i dunno but i think their are more MPC-HC based Player out their especially in the Asian direction. :)
And that is really sad the Potplayer guy could at least contribute back and normally he would have to under the GPL, testuo55 did you ever talked with him about this he seems to also have problems with ffmpeg and comply with the GPL their ?

However, the The KMPlayer authors have published an official comment on their own forum that denies the alleged GPL violation by Gabest.
Does somebody has a link to the written statement ?

[Daiz]

- A new, easier and universal way to render subtitles - preferably as a DirectShow filter, not directly hacked into the client.

Ever heard of VSFilter? :rolleyes:

[tetsuo55]

- A new, easier and universal way to render subtitles - preferably as a DirectShow filter, not directly hacked into the client.
- MASSIVE cleanup in the renderers (i mean actual cleanup, maybe even rewriting of huge parts, not hacking new "features" in there)
Neither of those are part of the core player imho, but filters. Granted these also need a rewrite and dedicated devs/

@Cruncher, afaik all the forks except splayer violate GPL. And to my knowledge none of them send us patches.

[nevcairiel]

Ever heard of VSFilter? :rolleyes:

VSFilter is crap. Its super slow, a simple text sub can make your video playback crawl - and it renders subs on the image before scaling and color space expansion.

Not to mention its buggy as hell, and gets sub positions wrong half the time.

Neither of those are part of the core player imho, but filters. Granted these also need a rewrite and dedicated devs

Indeed they are not, but for a full playback experience, they are needed. As you see above, only the first point is "core" player, everything else is stuff needed for a full playback package.

[Daiz]

VSFilter is crap. Its super slow, a simple text sub can make your video playback crawl - and it renders subs on the image before scaling and color space expansion.

Not to mention its buggy as hell, and gets sub positions wrong half the time.

While I agree that VSFilter is slow due to being poorly optimized and only single-threaded, it's not so slow that "a simple text sub" could make your video playback crawl unless you were using a Pentium II or something for playback. Neither does it get sub positions wrong half the time. You need to do quite heavy transform effects with ASS to make VSFilter slow down into unplayable territory, and those aren't something you could even do with simpler/unstyled subtitle formats.

But there isn't really much choice in terms of subtitle renderers. It's either VSFilter or libass. And libass doesn't have a DirectShow filter. That's certainly something I'd like to see, though. But while libass is a lot faster than VSFilter, it has also has some quirky rendering issues not present in VSFilter...

[tetsuo55]

While I agree that VSFilter is slow due to being poorly optimized and only single-threaded, it's not so slow that "a simple text sub" could make your video playback crawl unless you were using a Pentium II or something for playback. Neither does it get sub positions wrong half the time. You need to do quite heavy transform effects with ASS to make VSFilter slow down into unplayable territory, and those aren't something you could even do with simpler/unstyled subtitle formats.

But there isn't really much choice in terms of subtitle renderers. It's either VSFilter or libass. And libass doesn't have a DirectShow filter. That's certainly something I'd like to see, though. But while libass is a lot faster than VSFilter, it has also has some quirky rendering issues not present in VSFilter...VSfilter/internal renderer can cause framedrops from even the simplest subtitles. Its a major design flaw in the core that is not related to CPU time.

[nevcairiel]

While I agree that VSFilter is slow due to being poorly optimized and only single-threaded, it's not so slow that "a simple text sub" could make your video playback crawl unless you were using a Pentium II or something for playback. Neither does it get sub positions wrong half the time.

I recently fixed two bugs in VSFilter, and while testing i noticed all this.

I have a 1080p MKV with a very simple SRT subtitle, and displaying it just kills playback - and this is on a i7. Maybe its SRT thats so horribly inefficient, i don't have any real content with ASS subs, except samples for testing.

The bug i fixed was actually a positioning problem of PGS subs, half of them just showed up in the top left corner. Due to its weird code i was unable to fix the underlying issue, so i just fixed it in PGS - i don't doubt other subs still show up in the corner.

Not to mention that VSFilter suffers from the same code quality problems MPC-HC does (hey, its the same project), so recycling it is not really so easy.
If one were to write a new subtitle filter, there would at least be the possibility to look into re-using some of the subtitle rendering logic (after cleanups), as i realize that most fan-sub content is actually designed for VSFilter.

[zhfi]

test clsid patch still have
PPageFileInfoDetails.cpp
.\PPageFileInfoDetails.cpp(158) : error C3861: “StrFormatByteSizeW”: 找不到标识符
Playlist.cpp
.\Playlist.cpp(271) : error C3861: “StrCmpLogicalW”: 找不到标识符
.\MainFrm.cpp(4972) : error C3861: “StrFormatByteSizeW”: 找不到标识符
.\MainFrm.cpp(11702) : error C3861: “StrCmpLogicalW”: 找不到标识符
结果 生成日志保存在“file://C:\mpc-hc\bin\obj\Release_Win32\mpc-hc\BuildLog.htm”
mpc-hc - 4 个错误，18 个警告


try this patch while compile with VC2008:

Index: include/stdafx_common.h
===================================================================
--- include/stdafx_common.h (revision 3084)
+++ include/stdafx_common.h (working copy)
@@ -8,6 +8,10 @@
#define VC_EXTRALEAN // Exclude rarely-used stuff from Windows headers
#endif

+#ifndef NO_DSHOW_STRSAFE
+#define NO_DSHOW_STRSAFE
+#endif
+
#include <SDKDDKVer.h>

#define _ATL_CSTRING_EXPLICIT_CONSTRUCTORS // some CString constructors will be explicit

[Daiz]

I have a 1080p MVK with a very simple SRT subtitle, and displaying it just kills playback

Interesting. Was this with the internal VSFilter of MPC-HC or an external VSFilter? Also, can you cut a sample of the SRT track that I could try on a random 1080p video?

The bug i fixed was actually a positioning problem of PGS subs, half of them just showed up in the top left corner. Due to its weird code i was unable to fix the underlying issue, so i just fixed it in PGS - i don't doubt other subs still show up in the corner.

Admittedly I haven't tried to render any PGS subtitles with VSFilter (I don't think the external one even supports them?), but that's mainly because I OCR/convert all of my subtitles into ASS for superior subtitle quality.

If you were to write a new subtitle renderer, it'd definitely need to support ASS - if it didn't, it'd be a major step back from VSFilter/libass, even if it had a lot less bugs and was more stable. I for one would not be willing to give up all the features that ASS has to offer and neither would many others.

[nevcairiel]

Interesting. Was this with the internal VSFilter of MPC-HC or an external VSFilter? Also, can you cut a sample of the SRT track that I could try on a random 1080p video?

External, internal didn't seem to suffer from the problem.

Here is the sample i tested on, it has both PGS and SRT
http://www.mediafire.com/?ucid8fzk282jm8b


Admittedly I haven't tried to render any PGS subtitles with VSFilter (I don't think the external one even supports them?), but that's mainly because I OCR/convert all of my subtitles into ASS for superior subtitle quality.

It does. Make sure you use a build from the mpc-hc standalone filters package, not the original vsfilter from gabest's old project (last updated years ago)

Oh btw, If i look at the bitmap subs of the forced subs in Avatar, i don't think you could get the same quality, or anything close, without the original Font. Blu-ray PGS are already pre-rendered at 1080p, they look pretty good to me. The "normal" subs (not forced) look a bit worse though, like some standard font - but i only really want forced subs.

OCR'ing DVD subs can only improve their quality, i guess.


If you were to write a new subtitle renderer, it'd definitely need to support ASS - if it didn't, it'd be a major step back from VSFilter/libass, even if it had a lot less bugs and was more stable. I for one would not be willing to give up all the features that ASS has to offer and neither would many others.

Of course it would aim to support as many subtitle formats as possible. Possibly even allowing both libass and vsfilter backends for rendering.

Reminder: This is just thinking here, there are no concrete plans for coding anything.

[Daiz]

Oh btw, If i look at the bitmap subs of the forced subs in Avatar, i don't think you could get the same quality, or anything close, without the original Font.

Well, considering that they're using Papyrus (http://en.wikipedia.org/wiki/Papyrus_%28typeface%29) for it, it wouldn't be particularly hard to use the exact same font... and personally, a big reason to OCR it would be to change that travesty of a font to something actually pleasant to look at. And from what I've seen of BD subtitles in general, they seem to either use quite basic fonts or alternatively cock it up completely by doing something as stupid as using a serif font for the main dialogue. Or having no subtitle borders, only a dropshadow. Or both. And the list goes on...

[clsid]

While installing SP1 is certainly a good idea for everyone, breaking compilation for people that don't have it is just horrible.
Compiling should now work again with 2010 RTM.

Do you have any idea which code (or project settings) in the standalone filters might trigger the MFC bloat inclusion? Your filters don't suffer from it, which means it should be avoidable. A clean solution is of course preferred over the current workaround.

[nevcairiel]

Well, considering that they're using Papyrus (http://en.wikipedia.org/wiki/Papyrus_%28typeface%29) for it, it wouldn't be particularly hard to use the exact same font... and personally, a big reason to OCR it would be to change that travesty of a font to something actually pleasant to look at. And from what I've seen of BD subtitles in general, they seem to either use quite basic fonts or alternatively cock it up completely by doing something as stupid as using a serif font for the main dialogue. Or having no subtitle borders, only a dropshadow. Or both. And the list goes on...


I don't usually look at subs, i really hate them. I only watch content if i understand it - alien languages are of course the exception to this.
The forced subs in Avatar were actually "ok" to read during watching - IMHO of course - the normal subs for the english parts are as ugly as ever.


Do you have any idea which code (or project settings) in the standalone filters might trigger the MFC bloat inclusion? Your filters don't suffer from it, which means it should be avoidable. A clean solution is of course preferred over the current workaround.

My filters don't statically link MFC, i rather require the VS2010 runtime to be installed.

[Daiz]

I don't usually look at subs, i really hate them.

Well, if you're a German I can see why you would have an attitude like that :) In comparison to your experience with the alien language subtitles in Avatar, my thoughts during the first lines of alien dialogue were "Are they seriously using Papyrus as the font? That's goddamn cheap!" Personally I prefer to watch all media that I consume in its original language with subtitles. I also work with subtitles and subtitling on a weekly basis, so subtitle quality, formats, their features and rendering are quite important things for me and something I pay close attention to.

[clsid]

I recently fixed two bugs in VSFilter, and while testing i noticed all this.

I have a 1080p MKV with a very simple SRT subtitle, and displaying it just kills playback - and this is on a i7. Maybe its SRT thats so horribly inefficient, i don't have any real content with ASS subs, except samples for testing.

The bug i fixed was actually a positioning problem of PGS subs, half of them just showed up in the top left corner. Due to its weird code i was unable to fix the underlying issue, so i just fixed it in PGS - i don't doubt other subs still show up in the corner.

Not to mention that VSFilter suffers from the same code quality problems MPC-HC does (hey, its the same project), so recycling it is not really so easy.
If one were to write a new subtitle filter, there would at least be the possibility to look into re-using some of the subtitle rendering logic (after cleanups), as i realize that most fan-sub content is actually designed for VSFilter.
MPC-HC by default uses insane subtitle settings on quadcores (and higher). It enables "allow animation while buffering" which pretty much just eats CPU cycles. It is supposed to help with tearing issues, but its just a terrible code hack. I think it should be disabled or just removed entirely.
To get animation, use subpicture buffer of zero. For best performance (without correct animation) use non-zero buffer.

The biggest flaw in the subtitle renderer is that is uses the animation code path for every type of subtitle, even static ones like srt and vobsub. Fixing that would give a nice performance boost.

[clsid]

My filters don't statically link MFC, i rather require the VS2010 runtime to be installed.I forgot to mention that I modified them to be static to check out the size increase. The added size was in the order of 100KB, so pretty small.

[nevcairiel]

Well, if you're a German I can see why you would have an attitude like that :)

How is that?

I also don't watch content in german, unless its a german production. I watch it in english, as thats the language most the content i watch was produced in, come to think of it, nearly all of it, i guess.

PS:
Sometimes ignorance is bliss - i just enjoy a movie, and don't wonder about which font was used for the forced subs. :D

[cca]

MPC-HC by default uses insane subtitle settings on quadcores (and higher). It enables "allow animation while buffering" which pretty much just eats CPU cycles. It is supposed to help with tearing issues, but its just a terrible code hack. I think it should be disabled or just removed entirely.
To get animation, use subpicture buffer of zero. For best performance (without correct animation) use non-zero buffer.

The biggest flaw in the subtitle renderer is that is uses the animation code path for every type of subtitle, even static ones like srt and vobsub. Fixing that would give a nice performance boost.

Hack or not, with a buffer length of zero it's almost impossible to get subtitles without introducing Tearing. That's why this was introduced and it does the job, I watch fansubed anime for years now and the majority of them use ASS animated subtitles, I have tested those settings over time.

I do agree that it shouldn't be used for SRT and VobSubs though, no reason to.

[madshi]

Where does tearing come from in a subtitle renderer? Normally it's a video renderer problem...

[nevcairiel]

I forgot to mention that I modified them to be static to check out the size increase. The added size was in the order of 100KB, so pretty small.

It happens only if you use some function that actually causes those objects to be pulled in. Apparently i don't. My way of creating the properties pages is quite different to how MPC-HC filters do it, which may explain it (only place with alot of MFC usage).

Not sure what function exactly triggers it though, may be something in the MPC-HC base classes, DSUtil or whatever.

If its only 100Kb, i might even consider turning that on by default, and make those people shut up that don't have the runtime installed. :p