Media Player Classic - Home Cinema (MPC-HC) - v1.7.13

[cyberbeing]

http://pan.baidu.com/share/link?shareid=421917&uk=3558042035
mpc-hc can't handle the subtitle properly.

The first line under [Script Info] is GB2312 ASCII which is invalid because it cannot co-exist with the rest of the file which was saved as UTF-8 with signature. Essentially the file is corrupted. Remove or fix that first line to be valid UTF-8, and that script should work properly.

[Aleksoid1978]

The first line under [Script Info] is GB2312 ASCII which is invalid because it cannot co-exist with the rest of the file which was saved as UTF-8 with signature. Essentially the file is corrupted. Remove or fix that first line to be valid UTF-8, and that script should work properly.

MPC-BE handle the subtitle properly.

[cyberbeing]

xy-VSFilter used to handle that subtitle properly as well. At least on our end, this behavior is caused from merging in MPC-HC's UTF-8 without BOM auto-detection commits.

https://github.com/mpc-hc/mpc-hc/commit/7cb86ad153e6c6409a49cd1f9546aaa7411c514a
https://github.com/mpc-hc/mpc-hc/commit/fcfb1bac41f643fcbecfab81af5c45384819c8db
https://github.com/mpc-hc/mpc-hc/commit/6d1472b2f18266d92e5bc068667de348c0cd3b3b

@Aleksoid1978
Do you have a fixed version of those commits in MPC-BE, or do you just not include the UTF-8 auto-detection behavior at all?

[Aleksoid1978]

xy-VSFilter used to handle that subtitle properly as well. At least on our end, this behavior is caused from merging in MPC-HC's UTF-8 without BOM auto-detection commits.

https://github.com/mpc-hc/mpc-hc/commit/7cb86ad153e6c6409a49cd1f9546aaa7411c514a
https://github.com/mpc-hc/mpc-hc/commit/fcfb1bac41f643fcbecfab81af5c45384819c8db
https://github.com/mpc-hc/mpc-hc/commit/6d1472b2f18266d92e5bc068667de348c0cd3b3b

@Aleksoid1978
Do you have a fixed version of those commits in MPC-BE, or do you just not include the UTF-8 auto-detection behavior at all?

I do not use code from MPC-HC for UTF-8 detection.

[cyberbeing]

I do not use code from MPC-HC for UTF-8 detection.

Let me ask this a different way...

Has MPC-BE enhanced the original Guliverkli/Gabest VSFilter code to include support for UTF-8 w/o BOM auto-detection, or are you using the original Gabest code for opening subtitles? The original Gabest code works with wanezhiling's sample, but it does not do auto-detection.

In other words, does MPC-BE have source code which provides UTF-8 w/o BOM auto-detection, which I could merge into xy-VSFilter as a replacement for MPC-HC's commits?

[Aleksoid1978]

Let me ask this a different way...

Has MPC-BE enhanced the original Guliverkli/Gabest VSFilter code to include support for UTF-8 w/o BOM auto-detection, or are you using the original Gabest code for opening subtitles? The original Gabest code works with wanezhiling's sample, but it does not do auto-detection.

In other words, does MPC-BE have source code which provides UTF-8 w/o BOM auto-detection, which I could merge into xy-VSFilter as a replacement for MPC-HC's commits?

I use utf8 cpp library Release 2.3.2 for detect UTF-8.
https://sourceforge.net/projects/utfcpp/

[nevcairiel]

as a replacement for MPC-HC's commits?

Its not that MPC-HCs detection code is not working, its actually functioning perfectly - it does what its supposed to do (if thats the right thing is another question)
It tries to read the file as UTF-8, and if that fails, it by design goes back to Text/ANSI parsing. Considering this file is a mixture of both, it could easily go either way.

However, since this file actually has a UTF-8 BOM, this should probably overrule the fallback to ASCII/ANSI, and ignore the invalid line instead.
Sounds like an easy fix, i'll open a PR at Github.

For some reason people always seem to run to "revert" or "replace" instead of going with the solution of "fix". ;)

PS:
Changes here: https://github.com/mpc-hc/mpc-hc/pull/69

[cyberbeing]

Its not that MPC-HCs detection code is not working, its actually functioning perfectly - it does what its supposed to do (if thats the right thing is another question)
It tries to read the file as UTF-8, and if that fails, it by design goes back to Text/ANSI parsing. Considering this file is a mixture of both, it could easily go either way.

However, since this file actually has a UTF-8 BOM, this should probably overrule the fallback to ASCII/ANSI, and ignore the invalid line instead.

That's what I figured was going on, when I noticed ASCII on the first line of a script saved as UTF-8 w/ BOM.


For some reason people always seem to run to "revert" or "replace" instead of going with the solution of "fix". ;)
As a non-programmer myself, "fix" usually isn't a viable option. ;)

Sounds like an easy fix, i'll open a PR at Github.

PS:
Changes here: https://github.com/mpc-hc/mpc-hc/pull/69

Thanks nevcairiel, and Aleksoid1978 as well for pointing out that alternative UTF-8 detection library.

I had already finished merging what Aleksoid1978 pointed out before I noticed your post, so I guess I'll compare the two methods at some point.

[wanezhiling]

http://forum.doom9.org/showpost.php?p=1623115&postcount=983
MPC-HC can handle this smi subtitle, but can't handle like what PotPlayer did.

PotPlayer (http://i.imgur.com/NyBKy19.png) can choose KR+EN or KR or EN.:)

mpc-hc (http://i.imgur.com/MY0bLrw.png)

[vBm]

S.Chinese update
http://www.sendspace.com/file/vaoq67

That file is already in the HEAD. You either didn't save your changes or you uploaded wrong file.

Can you post new one ?

[betaking]

That file is already in the HEAD. You either didn't save your changes or you uploaded wrong file.

Can you post new one ?

OK!
http://www.sendspace.com/file/v1zpw1

[vBm]

OK!
http://www.sendspace.com/file/v1zpw1

Thanks. Merged as of 4868661a (https://github.com/mpc-hc/mpc-hc/commit/4868661a40733274c4d0bfcf405b2bed6a2e44dc)

[dbcooper]

Forgive me if this is a silly question, but is it possible that support for Media Foundation filters will be added?

[Armada]

Forgive me if this is a silly question, but is it possible that support for Media Foundation filters will be added?
We'd like to support it, but MPC-HC is in essence a DirectShow player. Supporting Media Foundation would require a lot of work. On top of that it would most likely mean dropping support for DirectShow, I can't see those two frameworks working nicely alongside each other in one player.

[JanWillem32]

EVR-CP already has some commented out sections for the Media Foundation parts. These were left by previous developers. However, there's no Media Foundation graph builder to use it in the main program. (I guess using the system default graph builder would suffice, though.) Because splitters, decoders, renderers and menu navigation filters for Media Foundation are rare right now, a lot of these would have to be converted from DirectShow parts.
I personally very much prefer Media Foundation over DirectShow. The Media Foundation systems are much more modern, and there are more features.
Because we will still support Windows XP SP3 for some time to come, we can't drop support for DirectShow just yet. I don't see Media Foundation and DirectShow inside the same executable happening, either.

[truexfan81]

EVR-CP already has some commented out sections for the Media Foundation parts. These were left by previous developers. However, there's no Media Foundation graph builder to use it in the main program. (I guess using the system default graph builder would suffice, though.) Because splitters, decoders, renderers and menu navigation filters for Media Foundation are rare right now, a lot of these would have to be converted from DirectShow parts.
I personally very much prefer Media Foundation over DirectShow. The Media Foundation systems are much more modern, and there are more features.
Because we will still support Windows XP SP3 for some time to come, we can't drop support for DirectShow just yet. I don't see Media Foundation and DirectShow inside the same executable happening, either.

would dropping DirectShow cause filters such as lav and madvr to stop working?

[JanWillem32]

In their current form, yes indeed.

[dukey]

There isn't even a dvd navigator for media foundation I believe.

[MokrySedeS]

Is there any progress with this (https://trac.mpc-hc.org/ticket/1295) feature request? Any plans?
MediaPortal already added support (http://wiki.team-mediaportal.com/1_MEDIAPORTAL_1/0_What's_New/1.3.x/Native_Blu-ray_Support) for this, but I like MPC-HC so much more...

[TheElix]

http://msdn.microsoft.com/en-us/library/windows/desktop/ms684247%28v=vs.85%29.aspx
Are developers aware of this Windows function? This is different than simply setting process priority in task manager. This feature is already used by foobar2000 and it helps to reduce drop-outs (and its consequences: pops, crackling, stuttering etc.). It could help video playback greatly.

[vBm]

http://gcc.gnu.org/gcc-4.7/

GCC 4.7.3 released.
Next nightly should be compiled using 4.7.3 toolchain.

http://msdn.microsoft.com/en-us/library/windows/desktop/ms684247%28v=vs.85%29.aspx
Are developers aware of this Windows function? This is different than simply setting process priority in task manager. This feature is already used by foobar2000 and it helps to reduce drop-outs (and its consequences: pops, crackling, stuttering etc.). It could help video playback greatly.

Underground78 replied to me on IRC and said that he is aware of it and that we already use that.

[21:07:04] <@vBm> unrelated ... when you have time ... check this out Underground78 -> http://forum.doom9.org/showpost.php?p=1624019&postcount=372
[21:07:55] <@Underground78> vBm: we already use that
[21:09:22] <@Underground78> vBm: yes I've found call to that function

[Mangix]

It could help video playback greatly.
no.

video is not audio. that being said, the audio renderer can use this if it does not already.

also, madVR does something similar already.

[TheElix]

Underground78 replied to me on IRC and said that he is aware of it and that we already use thatNice. Is it set by default to max priority (Pro Audio)?

[Armada]

Nice. Is it set by default to max priority (Pro Audio)?
No, EVR-CP is set to AVRT_PRIORITY_HIGH. The maximum priority would be AVRT_PRIORITY_CRITICAL, but I don't think that's a good idea, it probably wouldn't make that much of a difference anyway.

[TheElix]

What about audio renderer? I think video and audio renderer are separate processes, are they not?

[nevcairiel]

What about audio renderer? I think video and audio renderer are separate processes, are they not?

They are not, all in the same process.

[Aleksoid1978]

To developer - very bad playback on internal video decoder
http://aleksoid.voserver.net/Sample/H264/Interlace/422/Taylor.Swift.BRIT.Awards.2013.1080i50.H264.422.44Mbps.ts

[JanWillem32]

I just tried that sample. The sample is 50 fields per second interlaced, 8-bit, 4:2:2, AVC. The internal video decoder doesn't pass any interlaced 8-bit 4:2:2 format to the video renderer, while external decoders pass interlaced UYVY and YUY2.
For the internal decoder that means that the deinterlacer and converter to RGB are invoked. These are two really expensive conversions. Once the filter notices that many frames have been dropped and decoding is out of sync, it seeks to a reasonable point ahead to catch up. That's what's happening.
Even with the modifications I locally made to the video decoder parts, I could only achieve output of interlaced NV12. That does play decently, but I very much dislike the 4:2:2 to 4:2:0 chroma down-sampling in the process.
I could not get any 8-bit, 4:2:2 output from the decoder. Does any other developer have experience with the code for the internal AVC codec?

[Aleksoid1978]

This sample play fine with MPC-BE's internal decoder.

P.S.
Why do I post these examples, and files - I think that the developers will deal with the player, it froze the development of a couple of 2 years ago :)

Or opt out of the internal filters - why mislead users.

[jkauff]

Is Media Foundation fully supported on anything but Win 8? I know it's partially supported on Win 7 SP1. Might not be worth the effort for the partial support.

[clsid]

First of all, there is no need to drop DirectShow when adding Media Foundation support (which I don't see happening). MPC-HC already supports rendering through different frameworks (DS/QT/RM/Shockwave).
Secondly, Media Foundation has no benefits. It would be a total waste of efforts.
Thirdly, XP is still used by 40% of people worldwide. Supporting it is still essential and there is absolutely no need to drop support for it, nor would it bring any significant benefits atm.

[Reino]

DirectShow graph manager: Use the real filter name for source filters.
Previously the filename was used instead.Looks a lot better for DC-BSM this way. Thanks a lot!

[Aleksoid1978]

Hi dev. On ATI have image corruptions on many WEB-DL rip, short sample - http://aleksoid.voserver.net/Sample/H264/L3.1/Mentalist.s05e16.There.Will.Be.Blood.HD720p.WEB-DL.Rus.Eng.BaibaKo%20(1)-001.mkv

[LigH]

Forum rule #6 (http://forum.doom9.org/forum-rules.htm): Do not ask about downloaded copyrighted material (http://forum.doom9.org/announcement.php?a=36) (at least not so blatantly...)

[kasper93]

Hi dev. On ATI have image corruptions on many WEB-DL rip

This is known issue. AMD broken it some time ago in their drivers. I'm sure it can be fixed somehow, lav's dxva works fine. But really it's upstream bug and we should wait for the fix. That's not problem with web-dl but with h.264 L3.1 in general.

Fell free to help us and send patch with proper fix. :)

[judelaw]

Forum rule #6 (http://forum.doom9.org/forum-rules.htm): Do not ask about downloaded copyrighted material (http://forum.doom9.org/announcement.php?a=36) (at least not so blatantly...)
nothing venture nothing have (волков бояться - в лес не ходить)

[TheElix]

What MPC-HC needs, I think, is a nice fast-forward and rewind feature for all types of video.

[Aleksoid1978]

This is known issue. AMD broken it some time ago in their drivers. I'm sure it can be fixed somehow, lav's dxva works fine. But really it's upstream bug and we should wait for the fix. That's not problem with web-dl but with h.264 L3.1 in general.

Fell free to help us and send patch with proper fix. :)

Problem not in driver :) Problem in MPC-HC code - see

void CopyScalingMatrix(DXVA_Qmatrix_H264* pDest, PPS* pps, DWORD nPCIVendor)


and this code for ATI, but only for older UVD/UVD+ video card. For new UVD2+ - no need special code ...

If you want a fix - give me a access to MPC-HC repository. :)

[clsid]

This may be helpful for the DXVA issue:
http://git.1f0.de/gitweb?p=lavfsplitter.git;a=commitdiff;h=28ac28dacce9202b424a1d1bc7e06b5ee4b18eaa
http://git.1f0.de/gitweb?p=lavfsplitter.git;a=commitdiff;h=043ffa8d670e45dd8e87da1b5048a62f5e404115
http://git.1f0.de/gitweb?p=lavfsplitter.git;a=commitdiff;h=0e09dd31237a30cd3bbab4d126b95dbb127b07a5

[kasper93]

I'm sure you are right... I don't know if I should comment on this, but I think we don't care about fix for that. Changes are coming...

[clsid]

What do you mean?

[Orz]

What do you mean?

I guess mpc-hc will use LAVFilters as internal filter

[LigH]

@ TheElix:

Learn about I, IDR, P, and B frames in GOPs as long as 10 seconds. Then talk again about "fast rewind in any case"...

[Superb]

Guys, the download link on the main website should be:
http://sourceforge.net/projects/mpc-hc/files/MPC%20HomeCinema%20-%20Win32/MPC-HC_v1.6.7.7114_x86/MPC-HC.1.6.7.7114.x86.exe/download

And not:
http://sourceforge.net/projects/mpc-hc/files/MPC%20HomeCinema%20-%20Win32/MPC-HC_v1.6.7.7114_x86/MPC-HC.1.6.7.7114.x86.exe/download?accel_key=72%3A1366907644%3Ahttp%253A//mpc-hc.org/%3Aedb58f73%2483eac959afc05a4f975383f09e58b609e674d51f&click_id=f17130d8-adc5-11e2-89ad-0200ac1d1d8e&source=accel

A little bit clearer...
(Anyone can get the clean links by surfing here (http://sourceforge.net/projects/mpc-hc/files/MPC%20HomeCinema%20-%20Win32/))

EDIT: Btw, congrats on the new version!

[vBm]

After almost 1 and a half month since the last stable build, we decided to release 1.6.7.

This release is mainly a bugfix release. It also has a big performance improvement when using an INI file to store the settings.

Highlights of this release:

New INI parser: the accesses are now cached for faster settings loading/saving. Using the INI file should now be as fast as using the registry
DVB capture: Improve channel switching
Playlist improvements



You can download the new version here (http://mpc-hc.org/downloads/). For the complete changes see the changelog (http://mpc-hc.org/changelog/).

And if you're interested in looking at commits since the last stable you can see it here (https://github.com/mpc-hc/mpc-hc/compare/1.6.6...1.6.7). We've had 157 commits by 22 authors. :)

[minaust]

Playlist improvements


Speaking of which, where do I look for info on the .mpcpl playlist format for use in manual tweaking/editing?

[wanezhiling]

http://pan.baidu.com/share/link?shareid=451001&uk=3558042035
Can't display the sub in fullscreen, work fine in windowed mode.

[vivan]

Works fine here, in both fullscreen (http://3.firepic.org/3/images/2013-04/26/oqv5w0jkj2u3.png) and windowed modes...

[JanWillem32]

No, it actually is broken, as I get a debug warning: "Error in Rasterizer::ScanConvert: size (4579x35) is too big".
The rasterizer has a hard limit of 4000 pixels internally. Seeing that the full aperture and 4k video formats already go beyond that, I think this check should be removed. If we also count the various ASS/SSA special effects, 4000 was already rather small to begin with.
An easy fix, I'll see if I can get it integrated:https://github.com/JanWillem32/mpc-hc/commit/16c0994233c2e6325927316c94013835092da9d6

[JEEB]

No, it actually is broken, as I get a debug warning: "Error in Rasterizer::ScanConvert: size (4579x35) is too big".
The rasterizer has a hard limit of 4000 pixels internally. Seeing that the full aperture and 4k video formats already go beyond that, I think this check should be removed. If we also count the various ASS/SSA special effects, 4000 was already rather small to begin with.
This check was added lately to MPC-HC's VSFilter, and I have patched it into my xy-VSFilter fork because VSFilter-likes tend to take quite a bit of RAM for their rendering, and there are some samples that can just make it go OOAS (esp. on 32bit operating systems, where the single process's limit is 2GB). This limit was also discussed with xy-VSFilter's cyberbeing, and while it was found not to be perfect, it shouldn't even come up if the size of the rendered area is as big as the general 4K resolution (3840x<something>).

There is a similar sanity check in libass as well, but that doesn't use as much memory to render its surfaces to begin with, so the check is somewhat different.

You are welcome to comment on how to improve this check, or improve the memory usage of VSFilter's rendering, but I don't think removing this check is a good idea at this moment. Things like this really shouldn't OOM or OOAS in my honest opinion :) .

Edit: The check was at first an x * y kind of thing, but then it was changed to be an or check. Maybe some kind of such a check would be better in the end?