Log in

View Full Version : ffdshow tryouts project: Discussion & Development


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

fastplayer
16th January 2008, 17:09
98 support was dropped towards the end of 2007. :search:
I don't need to search. Haruhiko brought back 98 support a few builds after he dropped it by using unicows (http://ffdshow-tryout.svn.sourceforge.net/viewvc/ffdshow-tryout?view=rev&revision=1668).

Anyway, let's see when the first Win98 users start to complain about missing 98 support :D

Edit: 3000th post in this thread :eek:

Inventive Software
16th January 2008, 17:21
Oh bugger, I shoulda known about that, cause I suggested it to him! :D How silly of me! :o

Thunderbolt8
16th January 2008, 18:19
is there any way to get to know to which changes/updates of libavcodec the number in brackets e.g. (11539) refer?

fastplayer
16th January 2008, 18:30
is there any way to get to know to which changes/updates of libavcodec the number in brackets e.g. (11539) refer?
The comments are pretty short and technical:
http://lists.mplayerhq.hu/pipermail/ffmpeg-cvslog/2008-January/thread.html

clsid
16th January 2008, 18:31
Not without comparing the actual changes. It's too much work to specify all applied ffmpeg revisions. So I stopped doing that.

Kado
16th January 2008, 20:40
Audio filter bug: winamp2 filter
"click to select winamp 2 directoty..." button is not working, it doesnt pop up a windows and let me choose directory, when configure it offline.
it does work for me when im playing video/audio, but still, it doesnt show up the plugin(Signal processing studio DSP, yes i select winamp5 as dir, not the plugin dir)

In my case if I try to configure it through mpc external filter box I cannot change it, but if I go via windows start menu I can change it.
That plugin is probably blacklisted because I have others like enhancer and dfx and they show up in the list.
Even by changing the dsp plugin name won't work.

haruhiko_yamagata
17th January 2008, 10:37
Audio filter bug: winamp2 filter
"click to select winamp 2 directoty..." button is not working, it doesnt pop up a windows and let me choose directory, when configure it offline.
it does work for me when im playing video/audio, but still, it doesnt show up the plugin(Signal processing studio DSP, yes i select winamp5 as dir, not the plugin dir)
It works for me. Which build do you use?

Yong
17th January 2008, 11:58
It works for me. Which build do you use?
clsid r1787,
sorry, the bug i encountered is same as kado one. :p

Yup, SPS DSP is not show up, but other DSPs are working.

haruhiko_yamagata
17th January 2008, 12:45
Excluded WinAmp plugins are
// DSP stacker, Adapt-X and Vst host are not compatible with ffdshow currently. Maybe ffdshow's bug, but I can't help...
if ( _strnicmp(_l("dsp_stacker.dll"),filename,16)==0
|| _strnicmp(_l("dsp_adaptx.dll"),filename,15)==0
|| _strnicmp(_l("dsp_vst.dll"),filename,12)==0
)
return;

cc979
18th January 2008, 09:37
i just tried compiling svn 1790 (because of 1788) with 'make UNICODE=yes' but it fails on makeAVIS

haruhiko_yamagata
18th January 2008, 10:18
i just tried compiling svn 1790 (because of 1788) with 'make UNICODE=yes' but it fails on makeAVISmakeAVIS is an ANSI application. Unicode build is not supported.

Inventive Software
18th January 2008, 12:31
Nice joke leeperry! :D

I think I found a bug in ffdshow's MPEG-2 decoder.

*breathe*

MPEG-2 video in Matroska with the Aspect Ratio set doesn't display correctly in Media Player Classic 6.4.9.1 or WMP11 (the only 2 media players I have). I tried both libmpeg2 and libavcodec, both wouldn't give me a correct aspect ratio (though ffdshow's Info panel gives the correct SAR and DAR), and I have no idea whether it's the splitter (Haali Media Splitter) or the decoder because using MPC's internal MPEG-2 decoder it displays correctly.

*snip*

Update: I've found the same bug with the same elementary streams. So it's not parser related.

@SMD: Regarding your workaround, that does work, but only because MPC falls back on it's internal MPEG-2 decoder (even though it's specifically disabled).

Yong
18th January 2008, 12:46
Excluded WinAmp plugins are
// DSP stacker, Adapt-X and Vst host are not compatible with ffdshow currently. Maybe ffdshow's bug, but I can't help...
if ( _strnicmp(_l("dsp_stacker.dll"),filename,16)==0
|| _strnicmp(_l("dsp_adaptx.dll"),filename,15)==0
|| _strnicmp(_l("dsp_vst.dll"),filename,12)==0
)
return;

hmm ok, i dont the SPS DSP, and the winamp2 audio filter, just reporting bug :)
should be ok if put at low priority for bug fixing.

haruhiko_yamagata
18th January 2008, 13:07
MPEG-2 video in Matroska with the Aspect Ratio set doesn't display correctly in Media Player Classic 6.4.9.1 or WMP11 (the only 2 media players I have). I tried both libmpeg2 and libavcodec, both wouldn't give me a correct aspect ratio (though ffdshow's Info panel gives the correct SAR and DAR), and I have no idea whether it's the splitter (Haali Media Splitter) or the decoder because using MPC's internal MPEG-2 decoder it displays correctly.
In 352 x 576, 4:3 movie, libmpeg2 doesn't seem to support aspect ratio handling. I have added a special workaround at rev 1791.
Please let me know if you find a similar case. 352 x 480 may be the case. If you can encode such file, could you send me a sample?

As for libavcodec mpeg2, I'm not willing to fix as usual.

Inventive Software
18th January 2008, 16:09
Thank you haruhiko. I'm just waiting for a build. I'll see about making a test 352x480 video just to see if that works or not.

SeeMoreDigital
18th January 2008, 16:12
SMD: Regarding your workaround, that does work, but only because MPC falls back on it's internal MPEG-2 decoder (even though it's specifically disabled).Hi mate,

I've just checked my filters. I can confirm that FFdshow is being used: -

http://i17.tinypic.com/8fmc1ef.jpg

EDIT: And the aspect ratio is correct also: -

http://i6.tinypic.com/8fb6715.jpg


Cheers

cc979
18th January 2008, 18:49
makeAVIS is an ANSI application. Unicode build is not supported.

maybe i misread the change logs

ImAhNoBoDy
18th January 2008, 20:25
When I encode a video using x264 or ffmpeg to 29.97fps, ffdshow shows that the fps is 25 even though I made the fps 29.97 during the encode.

I even checked it out in gspot and it does say 29.97fps. Is this a bug? I have other videos that weren't encoded by me and it shows the correct fps on them.

haruhiko_yamagata
19th January 2008, 04:28
When I encode a video using x264 or ffmpeg to 29.97fps, ffdshow shows that the fps is 25 even though I made the fps 29.97 during the encode.

I even checked it out in gspot and it does say 29.97fps. Is this a bug? I have other videos that weren't encoded by me and it shows the correct fps on them.I can't reproduce. Could you send us a sample?

Wilbert
19th January 2008, 10:42
bug report:

ffdshow spews to the debug output a LOT. It got bad enough that I hacked code into 1.7.X that allows OutputDebugString() to be intercepted and filtered by calling DLL. I submitted a bug on this and it looks like it already got assigned.
http://forums.virtualdub.org/index.php?act=ST&f=7&t=15093&

I guess this bug is also present in the tryouts?

Leak
19th January 2008, 11:43
I guess this bug is also present in the tryouts?
*cough* (http://ffdshow-tryout.svn.sourceforge.net/viewvc/ffdshow-tryout?view=rev&revision=1770) :D

np: Uusitalo - Konevitsa (Karhunainen)

SeeMoreDigital
19th January 2008, 12:28
SMD: Regarding your workaround, that does work, but only because MPC falls back on it's internal MPEG-2 decoder (even though it's specifically disabled).I've done a bit more testing using different MKV muxing applications.

It would seem .MKV muxes generated by (G)DSmux behave quite differently (when played using MediaPlayer Classic/FFdshow), compared to .MKV muxes generated by AVI-Mux GUI or VirtualDubMod.

Does anybody know why there would be such a difference? I can provide short samples.


Cheers

fastplayer
19th January 2008, 15:59
ANSI versions of ffdshow.ax and ff_wmv9.dll are included in build 1798. I guess Win98 support is back :)

Thunderbolt8
19th January 2008, 16:22
I've done a bit more testing using different MKV muxing applications.

It would seem .MKV muxes generated by (G)DSmux behave quite differently (when played using MediaPlayer Classic/FFdshow), compared to .MKV muxes generated by AVI-Mux GUI or VirtualDubMod.

Does anybody know why there would be such a difference? I can provide short samples.


Cheers
there is ineed a difference between .mkv files created with gdsmux or with mkvmerge for example. for example at least with h264 files I know its somehow not possible to seek into the file, when they are created with gdsmux. for that to work you'd have to an additional mux with mkvmerge and rewrite the timestamps.
best would be to ask madshi about that difference, he knows a bit more about it.

Jay Bee
19th January 2008, 21:28
Any news on the VC-1 in TS AV-synch bug?

haruhiko_yamagata
20th January 2008, 01:56
Any news on the VC-1 in TS AV-synch bug?It has been fixed. If you still have problems, please send me a sample.

ImAhNoBoDy
20th January 2008, 02:09
I can't reproduce. Could you send us a sample?

I use 3GPConverter just for the GUI, but I put in the ffmpeg codes myself. I am using ffmpeg v9133

Command0=""<%AppPath%>\cores\ffmpeg" -y -i "<%InputFile%>" -title "<%Title%>" -timestamp "<%TimeStamp%>" -level 30 -vcodec h264 -coder 1 -trell 2 -partitions +parti4x4+partp8x8+partb8x8 -bufsize 128 -g 250 -s 720x480 -r 30000/1001 -b 1000k -pass 1 "<%outputfile%>.mp4""
Command1=""<%AppPath%>\cores\ffmpeg" -y -i "<%InputFile%>" -title "<%Title%>" -timestamp "<%TimeStamp%>" -level 30 -vcodec h264 -coder 1 -trell 2 -partitions +parti4x4+partp8x8+partb8x8 -bufsize 128 -g 250 -s 720x480 -r 30000/1001 -b 1000k -pass 2 -acodec aac -ac 2 -ar 48000 -ab 128k "<%Outputfile%>.MP4""

I am using Gspot 2.70a to see the fps. I am using ffdshow-rev1763_20080108_xxl currently. For the current sample I had to use an avs script to encode it cause I needed to just use the first 1000 frames. Was using directshowsource(). Either way, whenever I put the file directly into 3GPConverter, ffdshow still output the fps as 25.00 than 29.97. Strange thing is if you put "Changefps(29.97)" in the avs script, ffdshow will show the correct fps. Hope that helps.

Sample: http://www.sendspace.com/file/8b1601

haruhiko_yamagata
20th January 2008, 03:07
@ImAhNoBoDy
Thanks, I reproduced using MPC's internal splitter.
If I use Haali's splitter, ffdshow shows 29.97Hz.
It's just OSD problem, low priority, but should be fixed.

ImAhNoBoDy
20th January 2008, 15:18
oh cool, thanks.

SBeaver
20th January 2008, 19:11
clsid: does your latest x64 build include the h264 multicore patch? (or does any x64 build include it?)

ShadowVlican
20th January 2008, 22:04
i've never encountered any instability or problems playing MJPEG avi made by my camera (Canon SD700IS) so why is MJPEG disabled by default in ffdshow video decoder configuration?

clsid
20th January 2008, 22:29
clsid: does your latest x64 build include the h264 multicore patch? (or does any x64 build include it?)The patch was committed to SVN some time ago. My x64 build is just a regular SVN build, so yes the patch is included.

i've never encountered any instability or problems playing MJPEG avi made by my camera (Canon SD700IS) so why is MJPEG disabled by default in ffdshow video decoder configuration?Afaik MJPEG is enabled by default in the installer.

ShadowVlican
20th January 2008, 22:45
Afaik MJPEG is enabled by default in the installer.
ok i stand corrected

apparently you need to uninstall the previous version before updating to the latest one, then you'll see the option (during setup) to enable it

if you try to install the latest version on top (without uninstalling), then it'll be off by default :)

rickardk
21st January 2008, 00:20
Done alot of remuxing (HD DVDs and Blu-rays) into mkv container. Some titles have glitches and rainbow frames when remuxed into mkv.

One example was Ratatouille Blu-ray (AVC). At 40min and 30s there are some strange frames. Looks scambled. When watching the joined m2ts there is no problem.
Same problem reported by other user who also tried to remux into mkv.

Remuxed Million Dollar Baby HD DVD (VC-1) yesterday. It did show the same problem. But when watching the evo no problems.

Is this a mux, codec or splitter problem?



Anyone who know?

Inventive Software
21st January 2008, 14:58
Combination of muxing and splitter problem. Does the problem occur with the original? :)

rickardk
21st January 2008, 16:22
Combination of muxing and splitter problem. Does the problem occur with the original? :)

Original evo and m2ts plays fine.
A more detailed description can be found here:

http://forum.doom9.org/showthread.php?t=133974

red5goahead
21st January 2008, 19:46
is it possibile to specify device monitor in use? My htpc is based in Media Portal and I use the secondary monitor but ffdshow seems from OSD to use primary monitor for resize. I set 1223x688 my custom resolution on Panasonic plasma 37pv60. Always OSD report me 1280x720 (the lcd main montor resolution). "to screen resolution size" is usefully option to use. 1223x688 seems not accepted by ffdshow, it report a red field for the horizontal value.

Jay Bee
22nd January 2008, 00:15
It has been fixed. If you still have problems, please send me a sample.

Cool. Is it supposed to be fixed in rev1798? Because if it is I'll PM you some samples that don't work.

SBeaver
22nd January 2008, 01:07
The patch was committed to SVN some time ago. My x64 build is just a regular SVN build, so yes the patch is included.

Thanks for confirming.
It is, however, terribly slow compared to 32bit still.

clsid
22nd January 2008, 13:06
Thanks for confirming.
It is, however, terribly slow compared to 32bit still.
It will continue to be slow until we are able to compile a 64-bit version of libavcodec using GCC. Currently the 64-bit mingw environment is still in early development.

fastplayer
22nd January 2008, 15:44
What does the "fast rendering" option for subtitles actually do?
All I notice is a difference in character spacing.

Zarxrax
23rd January 2008, 01:44
I was wondering if its possible for the makeavis tool to stand on its own? That is, to be able to use it on a computer that does not have ffdshow installed? I found a page saying that makeavis was previously included with Avisynth version 2.5.2, so I'm assuming that its possible to stand on its own somehow?

haruhiko_yamagata
23rd January 2008, 11:20
Cool. Is it supposed to be fixed in rev1798? Because if it is I'll PM you some samples that don't work.Thank you, I have received your sample.
It works for me. Is your problem same as this (http://ffdshow-tryout.sourceforge.net/phpBB2/viewtopic.php?t=631)?

haruhiko_yamagata
23rd January 2008, 11:33
What does the "fast rendering" option for subtitles actually do?
All I notice is a difference in character spacing.Well, fast rendering is faster.fast rendering: User: 5s, kernel: 0s, total: 6s, real: 6s, fps: 908.0, dfps: 903.5
normal rendering: User: 16s, kernel: 9s, total: 26s, real: 26s, fps: 218.8, dfps: 218.3
I filled all the screen with OSD for this test.

Fast rendering cashes characters. To render "abc cba", render "a" "b" "c" to cash and connect characters. It is fast, but spaces between characters are not precise.

Normal rendering simply render "abc cba" using GDI.

Jay Bee
23rd January 2008, 12:32
Thank you, I have received your sample.
It works for me. Is your problem same as this (http://ffdshow-tryout.sourceforge.net/phpBB2/viewtopic.php?t=631)?

Thx for testing. No, the problem you linked to seems to be a different one. Can you tell me which player and splitter you used to test the file? I've tried several combinations and AV is always out of synch until I do a seek forward into the file.

Here's the sample in case other people want to test (note: problem is only with libavcodec, not with wmv9): http://rapidshare.com/files/85756920/clip.ts.html

fastplayer
23rd January 2008, 12:39
Well, fast rendering is faster.
No kidding! :D
I filled all the screen with OSD for this test.
So this setting applies to OSD rendering as well?
Maybe you could put that info into a tooltip. Something like:
Caches characters for improved rendering performance. Applies to OSD as well.
If disabled, GDI will be used instead.
or something similar...
It is fast, but spaces between characters are not precise.
Fortunately, the spacing can be compensated with the spacing option. The difference in performance is impressive though.

Edit: Just found a bug when using italics. Take a look at the screenshot:

http://img3.freeimagehosting.net/uploads/b203500ac2.png

And here's the SRT snippet:
745
01:06:52,265 --> 01:06:54,790
a <i>soupçon</i> of intimidation, <i>et voilà!</i>
As you can see, empty space before an opening <i>-element is rendered incorrectly with fast rendering set to on.
By the way, changing the spacing has no effect on this empty space when using italics. It just affects the characters.

haruhiko_yamagata
23rd January 2008, 14:07
So this setting applies to OSD rendering as well?No, OSD has its own font configuration page.
Maybe you could put that info into a tooltip. Something like:
Caches characters for improved rendering performance. Applies to OSD as well.
If disabled, GDI will be used instead.
or something similar...My bad explanation. GDI is used either way. Fast rendering send one character per one call, while normal rendering send one sentence per one call.

It is fast, but spaces between characters are not precise.This is good.

Fortunately, the spacing can be compensated with the spacing option. The difference in performance is impressive though.

Edit: Just found a bug when using italics. Take a look at the screenshot:

http://img3.freeimagehosting.net/uploads/b203500ac2.png

And here's the SRT snippet:
745
01:06:52,265 --> 01:06:54,790
a <i>soupçon</i> of intimidation, <i>et voilà!</i>
As you can see, empty space before an opening <i>-element is rendered incorrectly with fast rendering set to on.
By the way, changing the spacing has no effect on this empty space when using italics. It just affects the characters.Thank you, I'll look into it.

haruhiko_yamagata
23rd January 2008, 14:16
Thx for testing. No, the problem you linked to seems to be a different one. Can you tell me which player and splitter you used to test the file? I've tried several combinations and AV is always out of synch until I do a seek forward into the file.

Here's the sample in case other people want to test (note: problem is only with libavcodec, not with wmv9): http://rapidshare.com/files/85756920/clip.ts.htmlI'm using MPC homecinema and latest Haali's splitter. Core2duo 6700 (2.66GHz).
Is the sync completely lost or just a few frames off?

Jay Bee
23rd January 2008, 15:04
Problem still there for me with MPC HC + all versions of Haali with AMD X2@2.8GHz.

The synch error is a constant 700ms offset (audio too late). If I set Zoom Player to playback with -700ms offset all is fine. That is until I use the seekbar because the that's when the synch error goes away. :confused:

haruhiko_yamagata
23rd January 2008, 15:27
Problem still there for me with MPC HC + all versions of Haali with AMD X2@2.8GHz.

The synch error is a constant 700ms offset (audio too late). If I set Zoom Player to playback with -700ms offset all is fine. That is until I use the seekbar because the that's when the synch error goes away. :confused:What is your audio decoder? Do you use S/PDIF?