Log in

View Full Version : Media Player Classic Home Cinema (MPC-HC) - DXVA!


Pages : 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 [147] 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410

lych_necross
20th April 2009, 09:08
If I am using the vsync feature of mpc-hc, should I disable the vsync option in reclock? My instincts tell me that having both vsync tools enabled might cause problems.

Mark_A_W
20th April 2009, 09:36
If I am using the vsync feature of mpc-hc, should I disable the vsync option in reclock? My instincts tell me that having both vsync tools enabled might cause problems.

When using EVR Custom, with Alternate V-sync set, I have Reclock set to:

Slave reference clock to audio = TICKED

Media adaption speed set to ORIGINAL, and LOCKED is ticked.


This turns Reclock resampling off, leaving me with a WASAPI renderer.

Playback is smooth, really smooth, apart from a small glitch about once an hour.

Jong
20th April 2009, 09:39
If I am using the vsync feature of mpc-hc, should I disable the vsync option in reclock? My instincts tell me that having both vsync tools enabled might cause problems.You are right. If the MPC vsync correction is working for you then turn off Reclock's. Having both on at the same time causes periodic burst of judder every 20s to couple of minutes depending on hoe far away from the top of the screen the Reclock vsync target is set.

Peuj
20th April 2009, 10:10
The handiest way I know to apply changes is to reopen the same video with File / Recent Files (in Player "Keep history of recent opened files" must be checked).

If the options "Remember ..." in Miscellaneous are checked, the video restart from the same point and everything is quite fast.

Great, I'll use this for my test.

Thanks !!

Brazil2
20th April 2009, 16:56
Since the merge with Beliyaal branch, there's a slight delay after double click on a file (after the initial window is opened) and the moment it starts playing.
I've noticed the same issue but only when using EVR, it's OK with VMR7/9.

Since EVR is not working very fine on my XP system I'm usually not using it but for testing purposes, but I guess it must be annoying for people who are using EVR on a regular basis.

mark0077
20th April 2009, 18:56
Just an update on my vc-1 seek problems. I am now getting mpc-hc crashing SOMETIMES if I seek in a vc-1 .m2ts file. Very annoying today as a friend was over and it took about 5 attempts to open the file and seek to 10mins into the file.

Here is some output I get when the crash occurs if its useful?

Problem signature:
Problem Event Name: APPCRASH
Application Name: mplayerc.exe
Application Version: 1.2.1059.0
Application Timestamp: 49eba1a2
Fault Module Name: mplayerc.exe
Fault Module Version: 1.2.1059.0
Fault Module Timestamp: 49eba1a2
Exception Code: c0000005
Exception Offset: 003a46eb
OS Version: 6.0.6001.2.1.0.256.1
Locale ID: 6153
Additional Information 1: fd00
Additional Information 2: ea6f5fe8924aaa756324d57f87834160
Additional Information 3: fd00
Additional Information 4: ea6f5fe8924aaa756324d57f87834160

Just to note, my old problem of corruption and mpc-hc repeating 2-3 frames (one seemingly corrupt) over and over still happens 1 out of 5 times, with now 1 out of 5 times getting a crash also.... (sometimes more often)

ADude
20th April 2009, 19:22
Hi !

Can we get a post here in the MPC-HC thread with specific instructions for the vsync option changes added by the Beliyaal merge ?

Thanks !

saint-francis
20th April 2009, 21:13
Hi !

Can we get a post here in the MPC-HC thread with specific instructions for the vsync option changes added by the Beliyaal merge ?

Thanks !

I second this. I haven't had the time to keep up with the Beliyaal thread so a lot of the new functionality is unclear to me. I'm sure it is for others too; or at least everyone on the planet who didn't religiously keep up with that thread.

nijiko
20th April 2009, 21:45
I suggest to make a user manual with chm.

ADude
20th April 2009, 22:01
I suggest to make a user manual with chm.
That's a great idea for a long-term project, but first we were just looking for the sync instructions to be posted here in this thread.

ikarad
21st April 2009, 08:25
Hello everyone,

We have 2 batches left, most or all of these bugs should be fixed in 1053.

Subtitle problems(13):
https://sourceforge.net/tracker/?limit=100&func=&group_id=170561&atid=854651&assignee=&status=1&category=1165521&artgroup=&keyword=&submitter=&artifact_id=&limitz=100&assignee=&status=1&category=1165521&artgroup=&submitter=&keyword=&artifact_id=&submit=Filter&limitz=100&mass_category=&mass_priority=&mass_resolution=&mass_assignee=&mass_artgroup=&mass_status=&mass_cannedresponse=

Renderer problems(12):
https://sourceforge.net/tracker/?limit=100&func=&group_id=170561&atid=854651&assignee=&status=1&category=1165521&artgroup=&keyword=&submitter=&artifact_id=&limitz=100&assignee=&status=1&category=1165522&artgroup=&submitter=&keyword=&artifact_id=&submit=Filter&limitz=100&mass_category=&mass_priority=&mass_resolution=&mass_assignee=&mass_artgroup=&mass_status=&mass_cannedresponse=

Please check and confirm as many as possible.
Thank you!

2731336
http://sourceforge.net/tracker/?func=detail&aid=2731336&group_id=170561&atid=854651

and
2449648
http://sourceforge.net/tracker/?func=detail&aid=2449648&group_id=170561&atid=854651

the problems still exist with mpc homecinema 1059

nlnl
21st April 2009, 08:49
Hi !

Can we get a post here in the MPC-HC thread with specific instructions for the vsync option changes added by the Beliyaal merge ?

Thanks !

+1 :thanks:

tetsuo55
21st April 2009, 12:35
2731336
http://sourceforge.net/tracker/?func=detail&aid=2731336&group_id=170561&atid=854651

and
2449648
http://sourceforge.net/tracker/?func=detail&aid=2449648&group_id=170561&atid=854651

the problems still exist with mpc homecinema 1059

Thanks for testing

clsid
21st April 2009, 13:05
BUG: VSFilter fails to compile with current SVN

alexins
21st April 2009, 13:56
BUG: VSFilter fails to compile with current SVN

I confirm.

mark0077
21st April 2009, 14:25
Just an update on my vc-1 seek problems. I am now getting mpc-hc crashing SOMETIMES if I seek in a vc-1 .m2ts file. Very annoying today as a friend was over and it took about 5 attempts to open the file and seek to 10mins into the file.

Here is some output I get when the crash occurs if its useful?

Problem signature:
Problem Event Name: APPCRASH
Application Name: mplayerc.exe
Application Version: 1.2.1059.0
Application Timestamp: 49eba1a2
Fault Module Name: mplayerc.exe
Fault Module Version: 1.2.1059.0
Fault Module Timestamp: 49eba1a2
Exception Code: c0000005
Exception Offset: 003a46eb
OS Version: 6.0.6001.2.1.0.256.1
Locale ID: 6153
Additional Information 1: fd00
Additional Information 2: ea6f5fe8924aaa756324d57f87834160
Additional Information 3: fd00
Additional Information 4: ea6f5fe8924aaa756324d57f87834160

Just to note, my old problem of corruption and mpc-hc repeating 2-3 frames (one seemingly corrupt) over and over still happens 1 out of 5 times, with now 1 out of 5 times getting a crash also.... (sometimes more often)

Just looking through recent changes in svn and noticed this

Updated ffmeg
Fixed : crash on seek with True HD tracks

I only get the above quoted CRASH on seek with my True HD m2ts file. I get corruption / repeatedly repearing 2-3 frame bug on all vc-1 files after a couple of seeks.

Basically the fix for the True HD crash on seek in the svn above doesn't seem to work for me, and may have even made things worse as before these newer svn builds I didn't have any crashes seeking in True HD files. I will test later and try to figure out what exact version brought on the crashes, and see can I find a version that doesn't give me corruption / repeated frames on random seeks.

alexins
21st April 2009, 14:38
mark0077,

Can you try to start mpc in a command line with the option "mplayerc.exe /minidump" and upload the ".dmp" file generated after crash.

mark0077
21st April 2009, 14:42
mark0077,
Of course. I will try this when I am at my home machine later.

nijiko
21st April 2009, 18:23
That's a great idea for a long-term project, but first we were just looking for the sync instructions to be posted here in this thread.

Or make a wiki as ffdshow.

Casimir666
21st April 2009, 19:18
Or make a wiki as ffdshow.

I can open the wiki on SourceForge, but I haven't time to manage it. Do you want to take care of it?


PS : two wiki services are available : MediaWiki or Wikispaces (please tell me which one is better)

Casimir666
21st April 2009, 19:30
Or make a wiki as ffdshow.

BUG: VSFilter fails to compile with current SVN

Ok fixed

nijiko
21st April 2009, 19:37
I can open the wiki on SourceForge, but I haven't time to manage it. Do you want to take care of it?


PS : two wiki services are available : MediaWiki or Wikispaces (please tell me which one is better)


I consider MediaWiki is better.
And about the management, I suggest to open to everyone and check by moderators.

Egh
22nd April 2009, 00:42
In regards to EVR (not Custom) and MPC AVC decoder.

In my case it sends video in YUY2 format, and blank screen is shown. Is it correct to use YUY2 with EVR? As I found out it works here only with RGB (with ffdshow).

Also, with MPC decoder, when I want to view PIN [OUT], it scrolls down a lot of data in the text window, and only then displays data for [OUT].

What is this scrolling for?

hdboy
22nd April 2009, 01:21
Is there a way to play BD with MPC-HC when the movie is split into a lot of m2ts files? for most BDs, the movie is contained in one or two big m2ts files and I can play it fine with MPC. But with some BDs, like Pixar/Disney films, it's split into a 100 little m2ts files and they're not all contiguous, i.e., 101.m2ts and 102.m2ts may be completely different sections of the movie. I have to resort to powerdvd, but I prefer MPC for many different reasons. I'd love it if BD navigation (not decryption; that could be done elsewhere) were added to MPC.

Inspector.Gadget
22nd April 2009, 04:09
I still have the problem where the internal H.264 DXVA decoder stretches any content that isn't mod8 to fill out to mod8 dimensions, resulting in distortion of the bottom mod8-n pixels. I searched the thread but couldn't find a workaround for this problem. Has anyone discovered one?

Vista SP1 x64, NV 9650M GT w/ 179.48 drivers (most recent version that won't ever BSOD), C2D @ 2.53 gHz, 4GB Ram if any of this is relevant.

Mark_A_W
22nd April 2009, 06:12
Is there a way to play BD with MPC-HC when the movie is split into a lot of m2ts files? for most BDs, the movie is contained in one or two big m2ts files and I can play it fine with MPC. But with some BDs, like Pixar/Disney films, it's split into a 100 little m2ts files and they're not all contiguous, i.e., 101.m2ts and 102.m2ts may be completely different sections of the movie. I have to resort to powerdvd, but I prefer MPC for many different reasons. I'd love it if BD navigation (not decryption; that could be done elsewhere) were added to MPC.

Open->DVD will play the main title of a Bluray.

krbo
22nd April 2009, 07:18
In my case merging with beliyaal was too early as I reported earlier on this thread and on his also that there are problems in
reproduction while switching channels on a HDMI connected TV. (video stops or stutters like crazy after EDID)

SVN recovered from that error but beliyaal confirmed and claimed "it can't be fixed"

Now everyone with HDMI TV as only monitor (HTPC) have useless or severely crippled MPC HC.

So if I have something urgent to do during movie I must write down time mark instead of
simple pause and channel change for the rest of the family.

Mixer73
22nd April 2009, 07:21
Unfortunately I have to agree, I just installed the x64 version, changed some video settings, played one video, and it crashed.

I'll go try the 32bit version and see how it goes.

Barleyman
22nd April 2009, 07:47
Can you try to start mpc in a command line with the option "mplayerc.exe /minidump" and upload the ".dmp" file generated after crash.

Slight correction, MPC doesn't crash, it hangs. MPC window goes black and it's shown as "not responding" in task manager. Audio does keep on playing, thought. So no crash dump file, unfortunately.

Something going wrong with the VSYNC?

Jong
22nd April 2009, 09:13
Open->DVD will play the main title of a Bluray.I don't think that works with the seamless branching titles he is talking about, does it? :confused:

Jong
22nd April 2009, 09:14
I still have the problem where the internal H.264 DXVA decoder stretches any content that isn't mod8 to fill out to mod8 dimensions, resulting in distortion of the bottom mod8-n pixels. I searched the thread but couldn't find a workaround for this problem. Has anyone discovered one?

Vista SP1 x64, NV 9650M GT w/ 179.48 drivers (most recent version that won't ever BSOD), C2D @ 2.53 gHz, 4GB Ram if any of this is relevant.What renderer are you using?

Mercury_22
22nd April 2009, 09:20
In my case merging with beliyaal was too early as I reported earlier on this thread and on his also that there are problems in
reproduction while switching channels on a HDMI connected TV. (video stops or stutters like crazy after EDID)

SVN recovered from that error but beliyaal confirmed and claimed "it can't be fixed"

Now everyone with HDMI TV as only monitor (HTPC) have useless or severely crippled MPC HC.

So if I have something urgent to do during movie I must write down time mark instead of
simple pause and channel change for the rest of the family.

No problem Here so you may wanna check your system (drivers, OS, DirectX ....and reset MPC's settings by deleting the .ini or regs ! :)

krbo
22nd April 2009, 10:13
No problem Here so you may wanna check your system (drivers, OS, DirectX ....and reset MPC's settings by deleting the .ini or regs ! :)

if 1049 works OK and >= 1050 not (beliyaal's also not) then there is at least something outside my system, wouldn't you agree?


(Vista 32, HD3200, cat 9.3)

Mercury_22
22nd April 2009, 10:44
if 1049 works OK and >= 1050 not (beliyaal's also not) then there is at least something outside my system, wouldn't you agree?


(Vista 32, HD3200, cat 9.3)

Yes I agree IF is just that !

BUT how do you explain that I and others don't have such problem ? and also I don't see how the changes from SVN 1050 could have lead to this problem !
Revision 1050 - Directory Listing
Modified Sat Apr 18 22:06:04 2009 UTC (3 days, 11 hours ago) by casimir666
Changed : Libavcodec_gcc project activated for x64

Just a thought : Did you reset MPC's settings and are you using latest DirectX ?

tetsuo55
22nd April 2009, 11:57
Rewritten here:

http://forum.doom9.org/showthread.php?p=1276811#post1276811

Mercury_22
22nd April 2009, 12:50
@Everyone,

The recent changes should have fixed many bugs, and also introduced some new ones.

In order to get a clear and accurate view i invite everyone to verify that bugs can or cannot be reproduced or might even be fixed. (multiple verifications with different hardware/files are very welcome)

The full list can be found here (94 bugs)
http://sourceforge.net/tracker/?limit=100&func=&group_id=170561&atid=854651&assignee=&status=1&category=1165516&artgroup=&keyword=&submitter=&artifact_id=&limitz=100&assignee=&status=1&category=&artgroup=&submitter=&keyword=&artifact_id=&submit=Filter&limitz=100&mass_category=&mass_priority=&mass_resolution=&mass_assignee=&mass_artgroup=&mass_status=&mass_cannedresponse=

Also any new problems need to be reported there.

When reporting or confirming bugs please provide the following data:

-Windows and servicepack versions
-MPC-HC build version
-If appropriate FFdshow build number
-Videocard and driver version
-Renderer used and options selected in the output settings
-A sample so we can test it on our systems


We are unable to reproduce many of the bugs currently reported, from my limited testing i could close over half of them.

We need detailed information on how to reproduce them

I think it will be better / faster if we all take 5-10 bugs and we all test them 2-3 day then another 5-10 bugs and so on...
We can start with those you said you can close them already !

tetsuo55
22nd April 2009, 12:58
I think it will be better / faster if we all take 5-10 bugs and we all test them 2-3 day then another 5-10 bugs and so on...
We can start with those you said you can close them already !

Sounds like a good idea to me.

----------
The recent changes should have fixed many bugs, and also introduced some new ones.

In order to get a clear and accurate view i invite everyone to verify that bugs can or cannot be reproduced or might even be fixed. (multiple verifications with different hardware/files are very welcome)

Also any new problems need to be reported in the bug tracker.

When reporting or confirming bugs please provide the following data:

-Windows and servicepack versions
-MPC-HC build version
-If appropriate FFdshow build number
-Videocard and driver version
-Renderer used and options selected in the output settings
-A sample so we can test it on our systems


We are unable to reproduce many of the bugs currently reported, from my limited testing i could close over half of them.

We need detailed information on how to reproduce them

I have made batches, following Mercury's advice i would like to ask everyone to start testing batch 1:

https://sourceforge.net/tracker/?func=browse&group_id=170561&atid=854651&status=1&artgroup=898922

Mercury_22
22nd April 2009, 13:07
Sounds like a good idea to me.

----------

I have made batches, following Mercury's advice i would like to ask everyone to start testing batch 1:

https://sourceforge.net/tracker/?func=browse&group_id=170561&atid=854651&status=1&artgroup=898922

Then if no one else has something against I think you should pick some bugs and post them ! :)

Edit: looks like you beat me to it

Mixer73
22nd April 2009, 13:30
Slight correction, MPC doesn't crash, it hangs. MPC window goes black and it's shown as "not responding" in task manager. Audio does keep on playing, thought. So no crash dump file, unfortunately.

I have this symptom as well but only when I use EVR CP.

Vista x64, Q6600, 9800GT, current WHQL Nvidia drivers.

Inspector.Gadget
22nd April 2009, 14:39
What renderer are you using?

Happens with both EVR and VMR9, although latest MPC-HC build now doesn't use DXVA with VMR9 even on DXVA-compliant files and hence I've traded one problem for another.

krbo
22nd April 2009, 15:03
Yes I agree IF is just that !

BUT how do you explain that I and others don't have such problem ? and also I don't see how the changes from SVN 1050 could have lead to this problem !


Just a thought : Did you reset MPC's settings and are you using latest DirectX ?

yes, it' just and only that.


1050:


Media Player Classic HomeCinema (x86/x64), svn 1050

Supported languages: BY, CN, CZ, DE, EN, ES, FR, HU, IT, KR, PL, RU, SK, TR, UA

Changes log

Changed : merge with Beliyaal branch;
Added : support for madVR (new Video Renderer);
Resource cleanup.



Beliyaal changes are the reason, nothing else. He confirmed that in his thread.

And yes, in every time I have several builds in separate folders with their own ini files, directx si always the latest as it is required.

Why there are no more reports I don't know, maybe majority watch movies on PC monitor ?

Well, will see what will happen in future releases, I can't oppose majority.

Mercury_22
22nd April 2009, 15:17
yes, it' just and only that.


1050:



Beliyaal changes are the reason, nothing else. He confirmed that in his thread.

And yes, in every time I have several builds in separate folders with their own ini files, directx si always the latest as it is required.

Why there are no more reports I don't know, maybe majority watch movies on PC monitor ?

Well, will see what will happen in future releases, I can't oppose majority.

Where did you get the "changes" in the SVN cause your 1050 it's different from what I see in the SVN http://mpc-hc.svn.sourceforge.net/viewvc/mpc-hc/?view=log :confused:

hoborg
22nd April 2009, 15:20
Hi.
I thing there is a problem in latest MPC-HC standalone video filter - h.264 are decoded very slow in software mode (maybe caused by ffmpeg update?).
In HPC-HC player it is OK, but if you ar use standalone filter, something seems to be broken.

Using XP SP3 VMR9

To reproduce:
Register MPC-HC sandalone filter
Disable H.264 decoding in MPC-HC
Disable h.264 (DXVA) decoding in standalone decoder
Start HD or FullHD video (h.264).
Noticed since SVN 1056 standalone filters.

EDIT:
Video example: http://hobring.ic.cz/HDV_0091.zip

Beliyaal
22nd April 2009, 16:53
Beliyaal changes are the reason, nothing else. He confirmed that in his thread.

And yes, in every time I have several builds in separate folders with their own ini files, directx si always the latest as it is required.

Why there are no more reports I don't know, maybe majority watch movies on PC monitor ?

Well, will see what will happen in future releases, I can't oppose majority.

I can reproduce this problem, but have not found a way to fix it. Basically I use the driver differently than the old version, and this bug in the driver isn't exposed by the old version. I have thought of one more thing to try, so I will let you know if that pans out.

Snowknight26
22nd April 2009, 19:34
MPC-HC (64-bit at leasts) won't play E-AC3 streams. It recognizes the length but it doesn't play. When you try to seek, MPC-HC crashes.
Samples: http://stfcc.org/misc/eac3samples

Spec-Chum
22nd April 2009, 20:05
MPC-HC (64-bit at leasts) won't play E-AC3 streams. It recognizes the length but it doesn't play. When you try to seek, MPC-HC crashes.
Samples: http://stfcc.org/misc/eac3samples

Works fine here, tried with both internal and ffshow (libavcodec). Plays OK, and I can seek fine...

Tested using version 1060 compiled by me (no mods, just libavcodec compiled using -march=core2)

edit: oops, no it doesn't on x64 version, 32 bit version OK. Note to self: read previous post properly before flapping ya lips :)

clsid
22nd April 2009, 20:35
-march=core2 is not going to help much since all performance critical ffmpeg code is written in assembly.

Spec-Chum
22nd April 2009, 21:01
-march=core2 is not going to help much since all performance critical ffmpeg code is written in assembly.

Well, I know, but it makes me feel better :)

edit: lol, I just opened up one of the .asm files to have a quick peek...I had to check to make sure my font wasn't set to wingdings or something, now I know why they call it "code" :)

ADude
22nd April 2009, 22:01
yes, it' just and only that.


1050:



Beliyaal changes are the reason, nothing else. He confirmed that in his thread.

And yes, in every time I have several builds in separate folders with their own ini files, directx si always the latest as it is required.

Why there are no more reports I don't know, maybe majority watch movies on PC monitor ?

Well, will see what will happen in future releases, I can't oppose majority.

All of this sounds like horrible development methodology. :eek:

PS Note that Beliyaal has posted in this thread since three or four of us asked for a post of the instructions for the vsync changes... and still nothing...

Beliyaal
22nd April 2009, 22:24
All of this sounds like horrible development methodology. :eek:

PS Note that Beliyaal has posted in this thread since three or four of us asked for a post of the instructions for the vsync changes... and still nothing...

Well if you rather I don't post at all until I have time to sit down and write documentation I can do that.