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

[betaking]

It's not really high priority but it will probably be done.



We usually don't translate things until they get into the trunk I think.
@Underground78:last Chinese (Simplified) mplayerc.sc.rc
http://hotfile.com/dl/126125207/7558fc9/mplayerc.sc.rar.html

[JarrettH]

Which DVD is this happening? Take a screenshot of this subtitles.

What video decoder do you use? Take a screenshot of the settings decoder.

Actually I just investigated using an ffdshow setup to decode DVDs instead, it only happened using the Microsoft DTV decoder :thanks:

[Underground78]

@Underground78:last Chinese (Simplified) mplayerc.sc.rc
http://hotfile.com/dl/126125207/7558fc9/mplayerc.sc.rar.html

Thanks, I have created a new ticket on Trac (https://sourceforge.net/apps/trac/mpc-hc/ticket/1643) asking for the previous translators' approval before commiting. :)

[JohnLai]

So...any chance of my double click as command for exiting from D3D mode to be fixed?

[Underground78]

So...any chance of my double click as command for exiting from D3D mode to be fixed?

Exiting the whole application right ? It must be a race condition or something like that but I'm still not totally sure. I will try to have a look before the end of this week.

[Jtacdf]

Hi, the latest rev. 3592 seems to break some function of vsfilter on ass subs.
http://i197.photobucket.com/albums/aa315/Jtacdf/MPC-HC%20samples/Wrong.jpg
Last known working on rev. 3557.
http://i197.photobucket.com/albums/aa315/Jtacdf/MPC-HC%20samples/Right.jpg
Forgotten to link a sample.
http://www.mediafire.com/file/6b0fgpbvet7bbfe/%5BDoki%5D%20Suzumiya%20Haruhi%20no%20Yuuutsu%20%282009%29%20-%2008%20%281920x1080%20h264%20BD%20FLAC%29%20%5B79EC4B68%5D%20Sample.7z

[JanWillem32]

That looks like the /p drawing mode. Can you demux the ASS or SSA text subtitle file and look inside if it's indeed that?

edit: Verified, the sample indeed has colored rectangles using the /p drawing mode.

[Jtacdf]

I'm not sure. Opening up the ass in aegisub, the code goes something like "{\p1\c&H6C44D7&\pos(1107,1008)}m 0 0 l 233 0 233 750 0 750{\p0}" for that specific sample in the screenshot.

[clsid]

That bug is still there in 3605.

[Underground78]

This bug was introduced somewhere between r3557 and r3577.

Can you please try this build with r3600 reverted: http://www.mediafire.com/file/8dr5satajdzzqbp/mpc-hc.exe.7z ?

Edit: It seems to fix the problem, can you open a ticket on Trac please ?

[clsid]

The first part of this change is the cause:
http://mpc-hc.svn.sourceforge.net/viewvc/mpc-hc/trunk/src/Subtitles/RTS.cpp?r1=3560&r2=3559&pathrev=3560

[Underground78]

I reopened ticket #1563 (https://sourceforge.net/apps/trac/mpc-hc/ticket/1563) as r3560 was a fix for it.

[clsid]

The sample file from that ticket does not crash with LAV Splitter after reverting the patch. However, I also see no subtitles.

[Underground78]

Without reverting r3560 and using the internal splitter, I see two subtitles tracks and one does have subs. (Obviously, when r3560 is reverting, MPC-HC just crashes.)

[clsid]

LAV Splitter sees only one subtitle track. MediaInfo sees only one as well.

[betaking]

@Underground78:last Chinese (Simplified) mplayerc.sc.rc update again
http://hotfile.com/dl/126189202/ef394e6/mplayerc.sc.rc.txt.html
fix Compile error!

[Aleksoid1978]

Without reverting r3560 and using the internal splitter, I see two subtitles tracks and one does have subs. (Obviously, when r3560 is reverting, MPC-HC just crashes.)

MP4Splitter for some subtitle stream - add another, as plain text.

[CruNcher]

Hi could somebody work on that the Graph gets fully out of memory when the video is closed so that if you register a new external filter the same time you used a older one @ playback and close the video the new one gets utilized instead of the old instance (overwrite the same file and register while MPC-HC is open), so you don't need to restart MPC-HC anymore just for the new filter to become active :(

[Underground78]

@Underground78:last Chinese (Simplified) mplayerc.sc.rc update again
http://hotfile.com/dl/126189202/ef394e6/mplayerc.sc.rc.txt.html
fix Compile error!

Commited as r3609 ! :thanks:

[betaking]

Commited as r3609 ! :thanks:

To:Underground78,My friend zhfi be a new patch and a function for Play next in the folde to optimize this feature patch! If you are interested I will send you to see!:p

[Underground78]

To:Underground78,My friend zhfi be a new patch and a function for Play next in the folde to optimize this feature patch! If you are interested I will send you to see!:p

Please open a ticket on Trac. :)

[betaking]

Please open a ticket on Trac. :)

done!
https://sourceforge.net/apps/trac/mpc-hc/ticket/1646

[betaking]

If your folder contains audio and video files, in accordance with the practice before the check Play next in the folder, then finished playing the audio file will automatically play video files, and so was the entire folder where the files are played one time! But to apply this patch only after the file under the file name and extension automatically recognize the player! For example your file name contains 01, then play the end of this file will automatically play the file contains 02 instead of playing back the file 01 file!for example 01.MKV 02.MKV CD1.AVI CD2.AVI

[JanWillem32]

You mean that the file sorting method for when adding a folder to the playlist, should sort by extension first, and then by file name?

[betaking]

You mean that the file sorting method for when adding a folder to the playlist, should sort by extension first, and then by file name?

Bingo! you are right!:)

[betaking]

can not compiled last mpc-hc svn MPCVideoDec.ax for standalone filters,but compiled mpc-hc.exe is ok!
正在创建库 C:\mpc-hc\bin10\Filters_x86\MPCVideoDec.lib 和对象 C:\mpc-hc\bin10\Filters_x86\MPCVideoDec.exp
1>libavcodec.lib(log.o) : warning LNK4217: 本地定义的符号 __iob 在函数 _av_log_default_callback 中导入
1>libavcodec.lib(interf_dec.o) : warning LNK4049: 已导入本地定义的符号 __iob
1>libavcodec.lib(sp_dec.o) : warning LNK4049: 已导入本地定义的符号 __iob
1>libmingwex.a(wcrtomb.o) : warning LNK4217: 本地定义的符号 ___mb_cur_max 在函数 _wcrtomb 中导入
1>libmingwex.a(mbrtowc.o) : warning LNK4049: 已导入本地定义的符号 ___mb_cur_max
1>MPCVideoDecFilter.obj : error LNK2019: 无法解析的外部符号 "unsigned int __cdecl GetAdapter(struct IDirect3D9 *,struct HWND__ *)" (?GetAdapter@@YAIPAUIDirect3D9@@PAUHWND__@@@Z)，该符号在函数 "protected: void __thiscall CMPCVideoDecFilter::detectVideoCard(struct HWND__ *)" (?DetectVideoCard@CMPCVideoDecFilter@@IAEXPAUHWND__@@@Z) 中被引用
1>C:\mpc-hc\bin10\Filters_x86\MPCVideoDec.ax:fatal error LNK1120:1 个无法解析的外部命令
1>已完成生成项目“C:\mpc-hc\src\filters\transform\MPCVideoDec\MPCVideoDec.vcxproj”(build 个目标)的操作 - 失败。

生成失败。

已用时间 00:00:39.04

[Underground78]

I know what the problem is, I will try to fix that as soon as possible.

[betaking]

I know what the problem is, I will try to fix that as soon as possible.

thanks!

[Underground78]

thanks!

Fixed at r3622 (http://sourceforge.net/apps/trac/mpc-hc/changeset/3622/trunk).

[GTPVHD]

Microsoft has released a security update for Visual Studio 2010 SP1. All MPC-HC devs should install this update.

http://www.microsoft.com/download/en/details.aspx?displaylang=en&id=27049

http://support.microsoft.com/kb/2565057

[betaking]

Microsoft has released a security update for Visual Studio 2010 SP1. All MPC-HC devs should install this update.

http://www.microsoft.com/download/en/details.aspx?displaylang=en&id=27049

http://support.microsoft.com/kb/2565057

after install this update compiled last mpc-hc svn no problem!:)

[v0lt]

Microsoft has released a security update for Visual Studio 2010 SP1.
Would again like last time? :mad:
Now everything works without problems. Should not rush.

[CruNcher]

Does someone work on a DXVA Mpeg-2 Studio Profile (4:2:2) fallback to the Software Decoder for the internal MPC-HC Decoder ?

[Underground78]

Does someone work on a DXVA Mpeg-2 Studio Profile (4:2:2) fallback to the Software Decoder for the internal MPC-HC Decoder ?

Hmm, currently we only check (s->chroma_format < 2), I'm not sure if it's sufficient.

[Aleksoid1978]

Does someone work on a DXVA Mpeg-2 Studio Profile (4:2:2) fallback to the Software Decoder for the internal MPC-HC Decoder ?

Give a sample.

[v0lt]

Give a sample.
maybe it is. (http://forum.doom9.org/showpost.php?p=1491609&postcount=185)

[Aleksoid1978]

Does someone work on a DXVA Mpeg-2 Studio Profile (4:2:2) fallback to the Software Decoder for the internal MPC-HC Decoder ?

It's already done in rev. 3123 (http://mpc-hc.svn.sourceforge.net/viewvc/mpc-hc?view=revision&revision=3123)

[betaking]

to all dev:
compiled last mpc-hc svn standalone filters and mpc-hc.exe give me some warning C4018 !
for example
1>DXVADecoderVC1.cpp(214): warning C4018: “<”: 有符号/无符号不匹配
1>DXVADecoderVC1.cpp(215): warning C4018: “>=”: 有符号/无符号不匹配
1>DXVADecoderVC1.cpp(225): warning C4018: “>=”: 有符号/无符号不匹配
Able to compile successfully, but this is normal does it?
and use reshack open mpc-hc.exe-String Table-2582-1033 give me a RichEdit line insertion error.

[Underground78]

to all dev:
compiled last mpc-hc svn standalone filters and mpc-hc.exe give me some warning C4018 !
for example
1>DXVADecoderVC1.cpp(214): warning C4018: “<”: 有符号/无符号不匹配
1>DXVADecoderVC1.cpp(215): warning C4018: “>=”: 有符号/无符号不匹配
1>DXVADecoderVC1.cpp(225): warning C4018: “>=”: 有符号/无符号不匹配
Able to compile successfully, but this is normal does it?

That's "normal", those warnings were previously disabled, we have re-enabled them so that they can be eventually fixed.

and use reshack open mpc-hc.exe-String Table-2582-1033 give me a RichEdit line insertion error.

No idea about that, why would you need to use ResHack ?

Edit: @CruNcher: I can confirm, it seems that it already fall-backs to software decoding for 4:2:2 MPEG2.

[CruNcher]

It's already done in rev. 3123 (http://mpc-hc.svn.sourceforge.net/viewvc/mpc-hc?view=revision&revision=3123)

Huh right it seems i get garbage even without any DXVA @ all :( ? Build 1.5.3.3627

http://img824.imageshack.us/img824/6174/garbagestudio.png

[Underground78]

I would guess that the internal MPEG2 decoder based on libmpeg2 doesn't support 4:2:2.

[CruNcher]

I could swear it worked maybe it was another build though not from xhmikosr or it was decoded with ffdshow that should do it to so im a little surprised same for dscaler

I mean falling back to the internal decoder makes no sense if it doesn't support it then it's better to block the connection and fallback into the chain search.

Strange ffdshow seems to also fail now so only dscaler works currently.

Dscaler works (libmpeg2) YUY2

http://img215.imageshack.us/img215/9661/dscalerworks.png

though i would be for a fallback solution so MPC-HCs internal Decoder as long as DXVA is needed and if 4:2:2 content is tried then fallback to the chain and let the user define his prefered 4:2:2 Decoder (Dscaler,Mainconcept,Nero,Bitcontrol,ffdshow ect) :)

PS: Seems to be a bug in ffdshow YV12 needs to be disabled if the interlace flag is on else it corrupts

http://img171.imageshack.us/img171/2259/ffdshowworks.png


I pinpointed the problem in ffdshow to be the interlace flag in combination with YV12 selected (enabled it causes problems for Studio Profile input)
The output gets NV12 now and correctly rendered (seems Nvidia related) so YV12 should not be used when Studio Profile gets in and interlace flag is enabled instead NV12, this was tested with VMR9 Mixer mode and only happens though with YUV Mixing disabled with Yuv Mixing enabled it works with the default setup.

[betaking]

to all dev:last Chinese (Simplified) mplayerc.sc.rc
http://hotfile.com/dl/126489476/d115ecd/mplayerc.sc.rc.rar.html

[lovelove]

Is there a way to prevent MPC-HC from pausing the video upon single mouse left-click (i.e. left-mouse click should not trigger any action)?

[TheElix]

Is there a way to prevent MPC-HC from pausing the video upon single mouse left-click (i.e. left-mouse click should not trigger any action)?
http://rghost.ru/17604401/thumb.png (http://rghost.ru/17604401.view)

[Underground78]

The screenshot looks strange, you should do that in the "Mouse" column ("Mouse Windowed"/"Mouse Fullscreen" columns in latest versions).

[clsid]

Can AYUV input for the EVR CP renderer be disabled? It doesn't work correctly, giving either awful performance or borked video. See LAV Filters topic.

[Underground78]

Can AYUV input for the EVR CP renderer be disabled? It doesn't work correctly, giving either awful performance or borked video. See LAV Filters topic.

You should probably create a ticket and put JanWillem (janwillem32) in Cc.

[v0lt]

I would guess that the internal MPEG2 decoder based on libmpeg2 doesn't support 4:2:2.
It does not, because based on the old and much modified libmpeg2.

[JanWillem32]

Can AYUV input for the EVR CP renderer be disabled? It doesn't work correctly, giving either awful performance or borked video. See LAV Filters topic.When there's no support for a AYUV surface format on the video card, software emulation is enabled by the EVR mixer. Vanilla EVR does the same thing as EVR CP and EVR Sync in this regard. (And it's indeed very slow.)
It only gets borked because the software emulation mixer can't handle anything but X8R8G8B8 output frames.
As AYUV input isn't accepted natively by any video card at the moment, I don't have any problem with disabling it as a valid format until a proper custom mixer is written for the renderer. (As usual, C++ developers are very welcome to help write a custom mixer. The one I tried to write only made garbage images out of every incoming frame...)