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

[Mercury_22]

http://http://img827.imageshack.us/img827/6574/capturestat.th.jpg (http://img827.imageshack.us/i/capturestat.jpg/)
Here's what I get:
Surface A2R10G10B10 Backbuffer A2R10G10B10 Display X8R8G8B8
Using what ? = drivers, filters, npc-hc version, dxva or nondxva ?

And no black screen or "burned" colors when enabling "Force 10-bit RGB Input"?

[edigee]

I edited with the screenshot there too
MPCHC rev 2305 ,MPC video decoder ,DXVA, 10.8

[Momber]

http://img827.imageshack.us/img827/6574/capturestat.jpg
Here's what I get:
Surface A2R10G10B10 Backbuffer A2R10G10B10 Display X8R8G8B8
Same here. So you don't have a 10bit "chain" after all, either. Thanks for your replies.
I'm glad we've cleared this up now.

[Hera]

Is it me or does FullFP make things 'lagg' a bit?

1.3.2371.0 (MSVC 2010, GCC 4.5.2)
EVR Custom - No VSync - 0-255 - Bilinear (PS 2.0) - Buffers: 4
1920x1080 60Hz Monitor via HDMI
Windows 7 x86_32
Intel Atom N270 @ 800Mhz*
Nvidia ION (DX10 258.96 WHQL)

Video,
1080p MP4 from YouTube

*Same behavior @1.6Ghz, just not needed for h264

[namaiki]

Try set the scaler to Bilinear. (not PS 2.0) Also, you should keep VSync on to prevent stuttering.

[Hera]

Try set the scaler to Bilinear. (not PS 2.0) Also, you should keep VSync on to prevent stuttering.

Disabling hardware assisted bilinear scaling didn't assist with fullFloatingPoint option stutter. Reverting back.

V-Sync sometimes causes stuttering to a point of lagging the audio as it sometimes maxes out the CPU...
Having only one V-Sync option enabled relieves this issues but, disabling V-Sync altogether and enabling "Wait for Flushes" actively fixes tearing without stuttering or CPU side effects.

I think the issues goes away when using D3D fullscreen, but I don't find that a good compromise as I visually cant distinguish between fullFP on and off.

Also keep up the good work, MPC uses about 10% CPU for video (@1.6Ghz) which is nice. The other 20-25% comes from Desktop Window Manager and driver overhead :(

[namaiki]

With VSync disabled, I would get stutter every so often for ~30fps video on ~60Hz screen. If you don't experience it.. I guess that's good.

I can only notice the PQ difference of fullFP + Force 10bit (not 10 bit display) with this certain test video which displays a gradient.

[tetsuo55]

We are almost ready to release a new stable build.

What we still need is updated translations, practically every language needs updating. Since the Sourceforge mailing system is broken i am unable to use email to request an update from every contributor.

So please look at the translation for your language and update it if you can using notepad++ (other apps may mess with the formatting), more information here https://sourceforge.net/apps/trac/mpc-hc/wiki/Translations

The files that need translating are the mplayer.rc for your language, and the lines in custom_messages.iss for your language
:thanks:

[XhmikosR]

I have have updated the ffmpeg API to the latest one, though the old one will be still used but I need people to test it before committing it. Here's (http://www.mediafire.com/file/zz99c2kvsz1yjes/mpc-hc.exe.7z) the build. Please test every file that works with the current svn build. I'm more interested in DXVA and especially in MPEG2 DXVA.

[Snowknight26]

MPEG-2 DXVA never kicks in on ATI cards still.

[khagaroth]

Well I have continuously updated Czech translation, so it's uptodate already, but there is still this:
// Pls, don't translate this while ->
CONTROL "Don't use 'search in folder' on commands 'Skip back/forward' when only one item in playlist",IDC_CHECK7,
"Button",BS_AUTOCHECKBOX | BS_MULTILINE | WS_TABSTOP,7,164,223,18
LTEXT "OSD font",IDC_STATIC,7,191,57,8
COMBOBOX IDC_COMBO1,65,189,117,64,CBS_DROPDOWNLIST | CBS_SORT | WS_VSCROLL | WS_TABSTOP
COMBOBOX IDC_COMBO2,187,189,30,64,CBS_DROPDOWNLIST | WS_VSCROLL | WS_TABSTOP
// <- Pls, don't translate this while
Are we supposed to still leave it alone, or someone just forgot that he put it there?

[vBm]

Well I have continuously updated Czech translation, so it's uptodate already, but there is still this:
// Pls, don't translate this while ->
CONTROL "Don't use 'search in folder' on commands 'Skip back/forward' when only one item in playlist",IDC_CHECK7,
"Button",BS_AUTOCHECKBOX | BS_MULTILINE | WS_TABSTOP,7,164,223,18
LTEXT "OSD font",IDC_STATIC,7,191,57,8
COMBOBOX IDC_COMBO1,65,189,117,64,CBS_DROPDOWNLIST | CBS_SORT | WS_VSCROLL | WS_TABSTOP
COMBOBOX IDC_COMBO2,187,189,30,64,CBS_DROPDOWNLIST | WS_VSCROLL | WS_TABSTOP
// <- Pls, don't translate this while
Are we supposed to still leave it alone, or someone just forgot that he put it there?

Feel free to translate that and remove that comment surrounding it. It has been translated in almost all localizations already. Only left are CZ, NL, PL, SV and TR

[XhmikosR]

MPEG-2 DXVA never kicks in on ATI cards still.

Never said it did. I want someone to try who already has a card which has this capability.

[ryrynz]

The build numbers always end in .0, since the builds are incremented without ever using the last number, for the sake of tidiness wouldn't it be a good idea to remove it?

[XhmikosR]

Nope.

[ryrynz]

Fair enough :) quick question, I have the fullscreen option ticked and any file I click in explorer launches fullscreen fine but I'd like to have MPC start playback fullscreen when I play a file that's in the playlist. Is there any way of doing this without double clicking on the screen or using a shortcut key after playback?

[madshi]

But not in a middle-of-the road home cinema with 8bit amateur equimpent and with light bouncing back from white walls, ceiling, etc.
There, a gamma of 2.35 or 2.4 will effectively swallow perfectly good information. As in, kill it.
I agree that 2.35 or 2.4 is not a good choice if you have white walls, ceiling etc. As I said, it's for a batcave, only. I do not think that 8bit has anything to do with it, though. Being limited to 8bit means that the difference between neighbor brightness shades is bigger than when using a higher bitdepth. As a result there is a certain danger to see banding. But this is mainly a source limitation, and shows with *any* gamma curve you use.

[Tenki]

MPC-HC x86 MSVC 2010 2371 will crash when open .mov video.

[XhmikosR]

MPC-HC x86 MSVC 2010 2371 will crash when open .mov video.

That is a very useful report. Thanks, rest assure that someone will fix it.
Use the trac to report tickets and do it the right way (provide a sample, sys info, settings used and so on), otherwise the report is useless.

[Tenki]

That is a very useful report. Thanks, rest assure that someone will fix it.
Use the trac to report tickets and do it the right way (provide a sample, sys info, settings used and so on), otherwise the report is useless.

I don't know where to do it.
You can download this video and try to run it.
http://cdn2.gran-turismo.com/jp/common/data2/movie/20100615_e3_jp_eu/gt5_e3_2010_b.mov

[6205]

I don't know where to do it.
You can download this video and try to run it.
http://cdn2.gran-turismo.com/jp/common/data2/movie/20100615_e3_jp_eu/gt5_e3_2010_b.mov

Nice video. It makes me wanna sell X360 and buy PS3.
BTW. it's playing fine :)
http://img830.imageshack.us/i/gt5e.png/

[XhmikosR]

https://sourceforge.net/apps/trac/mpc-hc/newticket

https://sourceforge.net/apps/trac/mpc-hc/wiki/Bugs_-_Reporting

[Momber]

As a result there is a certain danger to see banding.
That is one of the effects.
Another is reduced delineation, for example shadow delineation. As a result, subtle differences in dark picture areas merge, and get swallowed up by the high gamma.
In other words, ppl see an undifferentiated dark swamp where with 12- or 14bit and/or gamma 2.2 they may still be able to make out differences in tonality.
This is all the more important because it seems to be the latest "fashion" for studios to crush black levels on BluRay encodes, in order to give their poor ancient video transfers more "pop".

And the same principle of delineation, or lack thereof, applies to very bright areas, near the opposite end of the scale, obviously.

It's easy to reproduce and verify this effect visually, I named a few movies that are suitable, but there are literally 1000 others. And with the newer builds of MPC you can switch gamma "on the fly" during critical scenes.
Try it - it's very interesting.

Your mileage may vary though. The better your equipment and calibration (and HT with respect to dark walls and ceiling, screen masking etc.) is, the less dramatic the losses will be.

[6205]

I don't know how to update my "Simple Gradient" toolbar on wiki. Any help? https://sourceforge.net/apps/trac/mpc-hc/wiki/Toolbar_images

[Remicade]

One thing to resolve in the future:change audio renderer on the fly (for those with 2 soundcards) like in KMPlayer and PotPlayer. One smart thing in PotPlayer: click right on the video, you don't change anything, click again on the video and the movie don't stop like in MPHC.

[Snowknight26]

You can always hit escape to close menus...

[Remicade]

I sit on the couch and i have only the mouse around :).

[Fortatus]

Small Bugreport (I don't want to register at sourceforge):

/src/filters/parser/MpegSplitter/MpegSplitter.cpp

Lines 751 to 755:
if(m_rtPlaylistDuration)
{
m_rtNewStop = m_rtStop = m_rtDuration = m_rtPlaylistDuration;
}
else
breaks lenght detection of normal DVD ifo-files. I loaded a movie into the playlist using VTS_01_0.ifo (the movie ifo) while MPEG PS/TS/PVA was deactivated and it first detected a lenght of 4 minutes (movie lenght: 2h 10m 21s) and I wasn't able to change the audiostream/language (the language submenu was grayed out). Then I turned on the internal MPEG PS/TS/PVA filter. When I started the movie again it detected the correct lenght. But when I started the movie a third time it shows a playback length of 225h. A fourth time clicking the ifo-entry in the playlist and the length ist 0:00 (overflow?). When closing MPC and restarting it the lenght is detected as 2h 10m, the next time 225h and so on.

Commenting out the code I posted above fixes the growing lenght when restarting a movie by clicking the playlist entry again. Of course it isn't a fix but the bug has to do something with this lines.

[Tenki]

https://sourceforge.net/apps/trac/mpc-hc/newticket

https://sourceforge.net/apps/trac/mpc-hc/wiki/Bugs_-_Reporting

It's ok in 2397, so I think it's fixed, thanks.

[tetsuo55]

I don't know how to update my "Simple Gradient" toolbar on wiki. Any help? https://sourceforge.net/apps/trac/mpc-hc/wiki/Toolbar_imagesill update it once your attachment is ok'd but you could also upload it to something like mediafire.

[tetsuo55]

Small Bugreport (I don't want to register at sourceforge):

/src/filters/parser/MpegSplitter/MpegSplitter.cpp

Lines 751 to 755:
if(m_rtPlaylistDuration)
{
m_rtNewStop = m_rtStop = m_rtDuration = m_rtPlaylistDuration;
}
else
breaks lenght detection of normal DVD ifo-files. I loaded a movie into the playlist using VTS_01_0.ifo (the movie ifo) while MPEG PS/TS/PVA was deactivated and it first detected a lenght of 4 minutes (movie lenght: 2h 10m 21s) and I wasn't able to change the audiostream/language (the language submenu was grayed out). Then I turned on the internal MPEG PS/TS/PVA filter. When I started the movie again it detected the correct lenght. But when I started the movie a third time it shows a playback length of 225h. A fourth time clicking the ifo-entry in the playlist and the length ist 0:00 (overflow?). When closing MPC and restarting it the lenght is detected as 2h 10m, the next time 225h and so on.

Commenting out the code I posted above fixes the growing lenght when restarting a movie by clicking the playlist entry again. Of course it isn't a fix but the bug has to do something with this lines.I have asked one of the developers to look at this, but without a real ticket (through which we can contact you) chances are high that this will get lost in this thread.

[Snowknight26]

I have asked one of the developers to look at this, but without a real ticket (through which we can contact you) chances are high that this will get lost in this thread.

Cough.
http://sourceforge.net/apps/trac/mpc-hc/ticket/654

[tetsuo55]

thanks Snowknight26

[6205]

ill update it once your attachment is ok'd but you could also upload it to something like mediafire.

Preview is jpg, but dowloaded is proper bmp for toolbar.
http://www.mediafire.com/imageview.php?quickkey=bdavyncot8r25fj&thumb=4

[Vavtnen]

Anyone have problem using MpcApi in 64bit version? The 32bit version does work, but I can't get my program to work with 64bit

Thanks

[hissatsu]

Just out of curiosity, is there any possibility of adding the ability to associate subtitle files in playlists? Playlists are little more than plain text listing of the file names. Would it be very difficult to have mpc-hc parse something like c:\videos\test.mkv,c:\subs\testsub.ssa and load the associated subtitle file for each video in the playlist? I know about automatically playing subtitle files with the same name as the video as long as they're in the same folder, but that doesn't help out my situation. Even if I had to manually edit the playlist files in an editor and there was no interface in mpc-hc to do it, it would still be very useful to me.

[MPC-HTPC]

Would it be possible to add an option in MPC-HC to specify what folder to auto-load subtitles from? Or if I were to do it myself, in what part of the code would I have to look for the auto-load functionality?

Alright, I'll just use the latest SVN-build in that case. I found where in the code to edit the auto-load locations, gonna see if I can change the code to enable users to set the folders.

heksesang (named in this forum Dreamhacker), one of the devs of MPC-HC, worked on a feature to specify a certain folder where MPC-HC should load the video subtitle from (among other subtitle handling features like preferring external subtitles and disable internal subtitles).
He already finished his patch, if I remeber correctly, but it has not yet been submitted to SVN. Also I am not sure if heksesang's patch includes the load subtitle from a specified folder function.
Maybe you can please shed some light here, Dreamhacker. Or XhmikosR could you maybe please give us some updated information?

[hissatsu]

While I'm sure those features would be useful to many, it unfortunately wouldn't help my situation (playing files off BDs with external subtitle files). That or a fix for the timer falling behind where it should be by about 1 second every time there's a a switch between m2ts files during seamless mpls playback - I discussed this with Sebastiii a while back and he confirmed there's an issue with the timer. I figured adding this sort of functionality to mpc-hc playlists would be easier than tracking down the BD timer issue - which probably involves the mpeg splitter, and it's a decent workaround for me.

[Aleksoid1978]

Small Bugreport (I don't want to register at sourceforge):

/src/filters/parser/MpegSplitter/MpegSplitter.cpp

Lines 751 to 755:
if(m_rtPlaylistDuration)
{
m_rtNewStop = m_rtStop = m_rtDuration = m_rtPlaylistDuration;
}
else
breaks lenght detection of normal DVD ifo-files. I loaded a movie into the playlist using VTS_01_0.ifo (the movie ifo) while MPEG PS/TS/PVA was deactivated and it first detected a lenght of 4 minutes (movie lenght: 2h 10m 21s) and I wasn't able to change the audiostream/language (the language submenu was grayed out). Then I turned on the internal MPEG PS/TS/PVA filter. When I started the movie again it detected the correct lenght. But when I started the movie a third time it shows a playback length of 225h. A fourth time clicking the ifo-entry in the playlist and the length ist 0:00 (overflow?). When closing MPC and restarting it the lenght is detected as 2h 10m, the next time 225h and so on.

Commenting out the code I posted above fixes the growing lenght when restarting a movie by clicking the playlist entry again. Of course it isn't a fix but the bug has to do something with this lines.

Try rev. 2404

[Polcius]

CPU Usage Increase Problem persists...

But I noticed something strange...

If I load Windows' ressource monitor to check CPU usage BEFORE I open MPC-HC, CPU usage stays at normal levels (15-20%).

But, If I open MPC-HC and THEN check for CPU usage, it shows 50% usage.

Weird...

[Snowknight26]

Try rev. 2404

Seems to be fixed.

[Vavtnen]

CMD_GETAUDIOTRACKS doesn't work with vob files. How can I report this?

I always get one audio tracks even tho there are multiple tracks within the vob

[MPC-HTPC]

Please use the links XhmikosR provided on the previous page and make a proper bug report at the trac following the guidelines.

https://sourceforge.net/apps/trac/mpc-hc/newticket

https://sourceforge.net/apps/trac/mpc-hc/wiki/Bugs_-_Reporting

[Superb]

Guys... what's up w/ the patch that gave precedence to external subs (instead of embedded subs)?

[madshi]

That is one of the effects.
Another is reduced delineation, for example shadow delineation. As a result, subtle differences in dark picture areas merge, and get swallowed up by the high gamma.
In other words, ppl see an undifferentiated dark swamp where with 12- or 14bit and/or gamma 2.2 they may still be able to make out differences in tonality.
This is all the more important because it seems to be the latest "fashion" for studios to crush black levels on BluRay encodes, in order to give their poor ancient video transfers more "pop".
Yeah, but this again is a source limitation. If the source is encoded by the studio with bad shadow detail then there's nothing we can do about that, no matter which gamma curve/value we use. So the 8bit source limitation does not have much to do with whether we should use a 2.35 or 2.2 gamma curve. Or in other words: Using a 2.35 gamma curve does not reduce those 8bit to something less. There are no steps that are going away. E.g. when watching at an 8bit grayscale, using a 2.35 curve just modifies the brightness differences between the 220 (16-235) different steps, but doesn't eat any of those 220 steps. Of course that's true only if video processing is done correctly...

[edigee]

Swithing back to the driver 10.5 beta:
http://www.megaupload.com/?d=0UTSKFBK
the weird CPU ussage behaviour is gone. Have to make more tests though.

@Mercury_22 : Sorry for a offtopic question:
Is the Samsung 6 series 10bit capable?

[XhmikosR]

Guys... what's up w/ the patch that gave precedence to external subs (instead of embedded subs)?

It wasn't ready and it still isn't ready.

[Steveo08]

i have a question again :-)

i want to try hd audio (dts-hd, dolby-true hd) over my gt220 as lpcm and the others (ac3, dts) over ac3filter to my receiver is it possible to manage with mpc-hc?

[Polcius]

How do I completely delete MPC-HC? (preferences, registry, filters... complete uninstallation)

[nevcairiel]

i have a question again :-)

i want to try hd audio (dts-hd, dolby-true hd) over my gt220 as lpcm and the others (ac3, dts) over ac3filter to my receiver is it possible to manage with mpc-hc?

For LPCM Streaming of HD formats, you need a decoder for those. TrueHD is supported by ffmpeg, so that will work. DTS-HD is however not, so you'll need to install a decoder that can do that. They are not free however, and usually only come in a package with big playback software, like PowerDVD or TotalMediaTheatre