May 2005 - ffmpeg-devel - lists.ffmpeg.org

(no subject)
by pgiacon＠ing.unife.it 02 Dec '05

02 Dec '05

hi, i am trying to use beosthread.c module, but i'm not able to find OS.h Is it a linux only file? unfortunatly because of my univer. job i must use windows...i tried to find OS.h with search but i did not find anything any suggestion? thanks a lot (as usual!) paolo -- C.N.I.T Consorzio Nazionale Interuniversitario per le Telecomunicazioni c/o Unit? di Ricerca di Ferrara Dipartimento di Ingegneria, Universita' di Ferrara Via Saragat 1, Ferrara

7 16

ratecontrol advice
by matthew＠mxtelecom.com 13 Aug '05

13 Aug '05

Hi all, I'm trying to use libavcodec to 1-pass encode an h.263 stream which has limits on both the maximum size-per-picture and total datarate. I can't work out what the right way is to respect both limitations using the ratecontrol settings, whilst maintaining the best possible QP for any given frame (so long as it doesn't end up larger than X bytes - or push the total datarate for over Y kbits/s). I assume I can do this with a custom rc_eq formula and/or adjusting qcompress - but I can't see an obvious formula for capping the picture-size. Alternatively, is there another way to force ffmpeg to adjust the qp so that pictures are never larger than a given size? Any suggestions would be very much appreciated indeed :) cheers; Matthew

6 14

MPEG-1/2 decoding benchmarks
by diego＠biurrun.de 25 Jul '05

25 Jul '05

Hi! I have repeated my MPEG-1/2 decoding benchmarks with the latest revisions of mpeg12.c. I used a current CVS snapshot of MPlayer for benchmarking with "-vc ffmpeg12 -quiet -vo null -nosound -benchmark" as common options for all runs. Benchmarks were run on the (framebuffer) console, I stopped all processes except syslogd, klogd and getty while running the benchmarks, no console output was generated. I used mplayer <options> <file> | grep BENCHMARK >> results to record the benchmarking results. I benchmarked separately with the "-lavdopts fast" option turned on. Every benchmark was run 10 times and the best result kept. Since the process is supposed to be deterministic, this should be a fair measurement. I tested revision 1.230 - 1.233 of mpeg12.c, 1.234 had no functional changes. Log messages: 1.230 avoid UPDATE_CACHE() in GET_RL_VLC() if not needed 1.231 optimizing non_intra ac coeff decode 1.232 mpeg2_fast_decode_block_intra() 1.233 avoid clearing block[] if its not used at all The movies I used can be found in http://mplayerhq.hu/~diego/benchmarks/ they were: bomb.mpeg - MPEG-1 176x144 heist60.mpeg - MPEG-1 352x240 dvd.mpeg - MPEG-2 720x576 I tested on two different systems, my desktop and notebook desktop: 500MHz AMD K6-III+ 512MB SD-RAM 133 Debian testing (up to date) gcc 2.95.4 (Debian) notebook: 1500MHz G4 with AltiVec 512MB PC2700 (333MHz) DDR SDRAM Debian testing (up to date) gcc 3.3.5 (Debian) If you have further suggestions or things I should do differently, speak up. I _could_ also test on a Pentium MMX 233, but I doubt the numbers will be very interesting, it's probably better if somebody tries a more modern processor. RESULTS: K6-III: 1.231 gave a slight speedup (2.3%) in some cases, 1.232 a slight performance degradation (1.7%) in some cases, no changes in 1.233. G4: 1.231 gave a slight speedup (1.2%) in some cases, 1.232 a slight improvement (1.2%) in one case, almost no changes in 1.233. Diego Here are the raw numbers: ############### # AMD K6-III+ # ############### 1.230: ~~~~~~ mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark bomb.mpeg BENCHMARKs: VC: 0.767s VO: 0.002s A: 0.000s Sys: 0.117s = 0.886s BENCHMARK%: VC: 86.5613% VO: 0.2794% A: 0.0000% Sys: 13.1593% = 100.0000% mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark heist60.mpeg BENCHMARKs: VC: 4.329s VO: 0.008s A: 0.000s Sys: 0.401s = 4.739s BENCHMARK%: VC: 91.3633% VO: 0.1651% A: 0.0000% Sys: 8.4716% = 100.0000% mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark dvd.mpeg BENCHMARKs: VC: 13.096s VO: 0.005s A: 0.000s Sys: 1.250s = 14.351s BENCHMARK%: VC: 91.2560% VO: 0.0348% A: 0.0000% Sys: 8.7092% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark bomb.mpeg BENCHMARKs: VC: 0.741s VO: 0.002s A: 0.000s Sys: 0.118s = 0.861s BENCHMARK%: VC: 86.0589% VO: 0.2593% A: 0.0000% Sys: 13.6818% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark heist60.mpeg BENCHMARKs: VC: 4.194s VO: 0.007s A: 0.000s Sys: 0.366s = 4.567s BENCHMARK%: VC: 91.8374% VO: 0.1504% A: 0.0000% Sys: 8.0122% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark dvd.mpeg BENCHMARKs: VC: 12.140s VO: 0.005s A: 0.000s Sys: 1.253s = 13.399s BENCHMARK%: VC: 90.6091% VO: 0.0376% A: 0.0000% Sys: 9.3533% = 100.0000% 1.231: ~~~~~~ mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark bomb.mpeg BENCHMARKs: VC: 0.757s VO: 0.002s A: 0.000s Sys: 0.120s = 0.880s BENCHMARK%: VC: 86.1142% VO: 0.2701% A: 0.0000% Sys: 13.6157% = 100.0000% mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark heist60.mpeg BENCHMARKs: VC: 4.256s VO: 0.007s A: 0.000s Sys: 0.372s = 4.636s BENCHMARK%: VC: 91.8181% VO: 0.1570% A: 0.0000% Sys: 8.0249% = 100.0000% mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark dvd.mpeg BENCHMARKs: VC: 12.772s VO: 0.005s A: 0.000s Sys: 1.255s = 14.032s BENCHMARK%: VC: 91.0222% VO: 0.0346% A: 0.0000% Sys: 8.9433% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark bomb.mpeg BENCHMARKs: VC: 0.740s VO: 0.003s A: 0.000s Sys: 0.121s = 0.864s BENCHMARK%: VC: 85.7312% VO: 0.3009% A: 0.0000% Sys: 13.9679% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark heist60.mpeg BENCHMARKs: VC: 4.173s VO: 0.007s A: 0.000s Sys: 0.384s = 4.565s BENCHMARK%: VC: 91.4272% VO: 0.1592% A: 0.0000% Sys: 8.4136% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark dvd.mpeg BENCHMARKs: VC: 12.074s VO: 0.005s A: 0.000s Sys: 1.248s = 13.327s BENCHMARK%: VC: 90.5978% VO: 0.0374% A: 0.0000% Sys: 9.3648% = 100.0000% 1.232: ~~~~~~ mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark bomb.mpeg BENCHMARKs: VC: 0.755s VO: 0.003s A: 0.000s Sys: 0.115s = 0.872s BENCHMARK%: VC: 86.5532% VO: 0.2869% A: 0.0000% Sys: 13.1600% = 100.0000% mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark heist60.mpeg BENCHMARKs: VC: 4.303s VO: 0.007s A: 0.000s Sys: 0.366s = 4.676s BENCHMARK%: VC: 92.0314% VO: 0.1393% A: 0.0000% Sys: 7.8293% = 100.0000% mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark dvd.mpeg BENCHMARKs: VC: 13.017s VO: 0.005s A: 0.000s Sys: 1.248s = 14.270s BENCHMARK%: VC: 91.2190% VO: 0.0318% A: 0.0000% Sys: 8.7492% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark bomb.mpeg BENCHMARKs: VC: 0.750s VO: 0.003s A: 0.000s Sys: 0.115s = 0.867s BENCHMARK%: VC: 86.4108% VO: 0.3270% A: 0.0000% Sys: 13.2623% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark heist60.mpeg BENCHMARKs: VC: 4.235s VO: 0.008s A: 0.000s Sys: 0.367s = 4.610s BENCHMARK%: VC: 91.8804% VO: 0.1675% A: 0.0000% Sys: 7.9521% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark dvd.mpeg BENCHMARKs: VC: 12.319s VO: 0.005s A: 0.000s Sys: 1.245s = 13.569s BENCHMARK%: VC: 90.7899% VO: 0.0354% A: 0.0000% Sys: 9.1747% = 100.0000% 1.233: ~~~~~~ mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark bomb.mpeg BENCHMARKs: VC: 0.763s VO: 0.002s A: 0.000s Sys: 0.119s = 0.884s mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark heist60.mpeg BENCHMARKs: VC: 4.290s VO: 0.007s A: 0.000s Sys: 0.373s = 4.671s BENCHMARK%: VC: 91.8507% VO: 0.1605% A: 0.0000% Sys: 7.9888% = 100.0000% mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark dvd.mpeg BENCHMARKs: VC: 12.981s VO: 0.005s A: 0.000s Sys: 1.247s = 14.233s BENCHMARK%: VC: 91.2054% VO: 0.0327% A: 0.0000% Sys: 8.7619% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark bomb.mpeg BENCHMARKs: VC: 0.751s VO: 0.003s A: 0.000s Sys: 0.119s = 0.873s BENCHMARK%: VC: 85.9708% VO: 0.3464% A: 0.0000% Sys: 13.6828% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark heist60.mpeg BENCHMARKs: VC: 4.276s VO: 0.008s A: 0.000s Sys: 0.362s = 4.646s BENCHMARK%: VC: 92.0346% VO: 0.1673% A: 0.0000% Sys: 7.7981% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark dvd.mpeg BENCHMARKs: VC: 12.291s VO: 0.005s A: 0.000s Sys: 1.246s = 13.542s BENCHMARK%: VC: 90.7621% VO: 0.0362% A: 0.0000% Sys: 9.2017% = 100.0000% ############### # Motorola G4 # ############### 1.230: ~~~~~~ mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark bomb.mpeg BENCHMARKs: VC: 0.322s VO: 0.001s A: 0.000s Sys: 0.065s = 0.388s BENCHMARK%: VC: 82.9979% VO: 0.3616% A: 0.0000% Sys: 16.6404% = 100.0000% mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark heist60.mpeg BENCHMARKs: VC: 1.755s VO: 0.003s A: 0.000s Sys: 0.158s = 1.916s BENCHMARK%: VC: 91.5799% VO: 0.1693% A: 0.0000% Sys: 8.2507% = 100.0000% mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark dvd.mpeg BENCHMARKs: VC: 5.851s VO: 0.003s A: 0.000s Sys: 0.766s = 6.620s BENCHMARK%: VC: 88.3842% VO: 0.0422% A: 0.0000% Sys: 11.5737% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark bomb.mpeg BENCHMARKs: VC: 0.306s VO: 0.001s A: 0.000s Sys: 0.065s = 0.372s BENCHMARK%: VC: 82.2443% VO: 0.4000% A: 0.0000% Sys: 17.3558% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark heist60.mpeg BENCHMARKs: VC: 1.710s VO: 0.003s A: 0.000s Sys: 0.157s = 1.870s BENCHMARK%: VC: 91.4493% VO: 0.1560% A: 0.0000% Sys: 8.3947% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark dvd.mpeg BENCHMARKs: VC: 5.555s VO: 0.003s A: 0.000s Sys: 0.765s = 6.322s BENCHMARK%: VC: 87.8600% VO: 0.0467% A: 0.0000% Sys: 12.0933% = 100.0000% 1.231: ~~~~~~ mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark bomb.mpeg BENCHMARKs: VC: 0.318s VO: 0.002s A: 0.000s Sys: 0.065s = 0.384s BENCHMARK%: VC: 82.8190% VO: 0.3904% A: 0.0000% Sys: 16.7905% = 100.0000% mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark heist60.mpeg ENCHMARKs: VC: 1.752s VO: 0.003s A: 0.000s Sys: 0.162s = 1.917s BENCHMARK%: VC: 91.3908% VO: 0.1795% A: 0.0000% Sys: 8.4297% = 100.0000% mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark dvd.mpeg BENCHMARKs: VC: 5.813s VO: 0.003s A: 0.000s Sys: 0.764s = 6.580s BENCHMARK%: VC: 88.3431% VO: 0.0473% A: 0.0000% Sys: 11.6096% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark bomb.mpeg BENCHMARKs: VC: 0.301s VO: 0.001s A: 0.000s Sys: 0.064s = 0.366s BENCHMARK%: VC: 82.0236% VO: 0.3869% A: 0.0000% Sys: 17.5894% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark heist60.mpeg BENCHMARKs: VC: 1.708s VO: 0.003s A: 0.000s Sys: 0.160s = 1.871s BENCHMARK%: VC: 91.2674% VO: 0.1611% A: 0.0000% Sys: 8.5715% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark dvd.mpeg BENCHMARKs: VC: 5.472s VO: 0.003s A: 0.000s Sys: 0.765s = 6.240s BENCHMARK%: VC: 87.6865% VO: 0.0475% A: 0.0000% Sys: 12.2660% = 100.0000% 1.232: ~~~~~~ mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark bomb.mpeg BENCHMARKs: VC: 0.318s VO: 0.001s A: 0.000s Sys: 0.064s = 0.384s BENCHMARK%: VC: 82.8557% VO: 0.3629% A: 0.0000% Sys: 16.7814% = 100.0000% mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark heist60.mpeg BENCHMARKs: VC: 1.736s VO: 0.003s A: 0.000s Sys: 0.162s = 1.901s BENCHMARK%: VC: 91.3084% VO: 0.1663% A: 0.0000% Sys: 8.5252% = 100.0000% mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark dvd.mpeg BENCHMARKs: VC: 5.794s VO: 0.003s A: 0.000s Sys: 0.765s = 6.562s BENCHMARK%: VC: 88.2966% VO: 0.0430% A: 0.0000% Sys: 11.6603% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark bomb.mpeg BENCHMARKs: VC: 0.300s VO: 0.001s A: 0.000s Sys: 0.065s = 0.366s BENCHMARK%: VC: 81.9985% VO: 0.3761% A: 0.0000% Sys: 17.6254% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark heist60.mpeg BENCHMARKs: VC: 1.689s VO: 0.003s A: 0.000s Sys: 0.157s = 1.849s BENCHMARK%: VC: 91.3421% VO: 0.1492% A: 0.0000% Sys: 8.5087% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark dvd.mpeg BENCHMARKs: VC: 5.455s VO: 0.003s A: 0.000s Sys: 0.765s = 6.223s BENCHMARK%: VC: 87.6639% VO: 0.0464% A: 0.0000% Sys: 12.2897% = 100.0000% 1.233: ~~~~~~ mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark bomb.mpeg BENCHMARKs: VC: 0.317s VO: 0.001s A: 0.000s Sys: 0.064s = 0.383s BENCHMARK%: VC: 82.7905% VO: 0.3877% A: 0.0000% Sys: 16.8218% = 100.0000% mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark heist60.mpeg BENCHMARKs: VC: 1.724s VO: 0.003s A: 0.000s Sys: 0.164s = 1.891s BENCHMARK%: VC: 91.1518% VO: 0.1577% A: 0.0000% Sys: 8.6905% = 100.0000% mplayer -vc ffmpeg12 -quiet -vo null -nosound -benchmark dvd.mpeg BENCHMARKs: VC: 5.769s VO: 0.003s A: 0.000s Sys: 0.770s = 6.542s BENCHMARK%: VC: 88.1854% VO: 0.0475% A: 0.0000% Sys: 11.7671% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark bomb.mpeg BENCHMARKs: VC: 0.299s VO: 0.001s A: 0.000s Sys: 0.064s = 0.365s BENCHMARK%: VC: 81.9638% VO: 0.3892% A: 0.0000% Sys: 17.6469% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark heist60.mpeg BENCHMARKs: VC: 1.686s VO: 0.003s A: 0.000s Sys: 0.160s = 1.849s BENCHMARK%: VC: 91.1841% VO: 0.1595% A: 0.0000% Sys: 8.6565% = 100.0000% mplayer -vc ffmpeg12 -lavdopts fast -quiet -vo null -nosound -benchmark dvd.mpeg BENCHMARKs: VC: 5.421s VO: 0.003s A: 0.000s Sys: 0.767s = 6.191s BENCHMARK%: VC: 87.5587% VO: 0.0478% A: 0.0000% Sys: 12.3935% = 100.0000%

2 3

Re: H.264 Mini Press Release
by b-o-n-d＠gmx.net 03 Jul '05

03 Jul '05

> On Mon, 23 May 2005, bond wrote: > > > i assume the main reason, why apples HD trailers dont work correctly is that > > they use multiple sliced frames, which libavcodec still doesnt seem to > > decode correctly, as can be seen too on moonlight and videosoft samples > > I tested on all of the samples in the Apple HD gallery, (which all use > multislice) and all of them worked perfectly. > That said, I have sometimes seen problems when decoding multislice from > moonlight. indeed, there are problems with the moonlight clips, but i also get exactly the same problems (artefacts looking the same way) with multislice clips from videosoft i have uploaded two such samples to incoming/avc_multislice. as you can see on the clips the errors dont occur on all frames, but only on some. cheers bond

4 6

libavformat/utils.c doxygen documentation
by danielk＠cat.nyu.edu 28 Jun '05

28 Jun '05

I just documented a bunch of functions. Hopefully it is useful. -- Daniel Kristjansson <danielk at cat.nyu.edu>

2 2

mpeg2 segdefault
by castet.matthieu＠free.fr 28 Jun '05

28 Jun '05

Hi, playing a mpeg2 stream work fine with ffplay, but ffmpeg segdefault when trying to transcoding to mpeg1 or mpeg2 (no segdefault if transcoding to mpeg4). (gdb) r -i /tmp/p.vob test.mpg Starting program: /home/mat/appli/matroska/ffmpeg/ffmpeg_g -i /tmp/p.vob test.mpg ffmpeg version 0.4.9-pre1, build 4755, Copyright (c) 2000-2004 Fabrice Bellard configuration: --prefix=/usr/local/stow/ffmpegcvs/ built on May 21 2005 11:02:50, gcc: 3.3.6 (Debian 1:3.3.6-5) Input #0, mpeg, from '/tmp/p.vob': Duration: 00:00:00.0, start: 0.287267, bitrate: 1979573 kb/s Stream #0.0: Video: mpeg2video, yuv420p, 720x576, 25.00 fps, 8000 kb/s File 'test.mpg' already exists. Overwrite ? [y/N] y Output #0, mpeg, to 'test.mpg': Stream #0.0: Video: mpeg1video, yuv420p, 720x576, 25.00 fps, q=2-31, 200 kb/s Stream mapping: Stream #0.0 -> #0.0 Press [q] to stop encoding Program received signal SIGSEGV, Segmentation fault. 0x082f7bbc in dct_unquantize_mpeg2_inter_mmx (s=0x83ec0f0, block=0x841e020, n=63, qscale=6) at i386/mpegvideo_mmx.c:411 411 else nCoeffs= s->intra_scantable.raster_end[ s->block_last_index[n] ]; (gdb) bt #0 0x082f7bbc in dct_unquantize_mpeg2_inter_mmx (s=0x83ec0f0, block=0x841e020, n=63, qscale=6) at i386/mpegvideo_mmx.c:411 #1 0x080c8b65 in encode_thread (c=0x83e3d88, arg=0x83ec0f0) at mpegvideo.c:3591 #2 0x080af143 in avcodec_default_execute (c=0x83e3d88, func=0x80c3300 <encode_thread>, arg=0x83ec1b4, ret=0x0, count=1) at utils.c:423 #3 0x080b72ac in encode_picture (s=0x83ec0f0, picture_number=0) at mpegvideo.c:5446 #4 0x080b594e in MPV_encode_picture (avctx=0x83e3d88, buf=0xb7cd5020 "", buf_size=1658880, data=0x3f) at mpegvideo.c:2339 #5 0x080af2ce in avcodec_encode_video (avctx=0x83e3d88, buf=0x3f <Address 0x3f out of bounds>, buf_size=1658880, pict=0xbfffeda0) at utils.c:584 #6 0x0805d2b0 in do_video_out (s=0x83e2e10, ost=0x83e1580, ist=0x83e14e0, in_picture=0xbffff1b0, frame_size=0xbffff0b0) at ffmpeg.c:910 #7 0x0805bec6 in output_packet (ist=0x83e14e0, ist_index=0, ost_table=0x83e1530, nb_ostreams=1, pkt=0xbffff3b0) at ffmpeg.c:1277 #8 0x08055614 in av_encode (output_files=0x837fc60, nb_output_files=1, input_files=0x837fb60, nb_input_files=1, stream_maps=0x837fcc0, nb_stream_maps=0) at ffmpeg.c:1949 #9 0x08054361 in main (argc=103984, argv=0x3f) at ffmpeg.c:4231 (gdb) disass $pc-32 $pc+32 Dump of assembler code from 0x82f7b9c to 0x82f7bdc: 0x082f7b9c <dct_unquantize_mpeg2_inter_mmx+12>: dec %ebp 0x082f7b9d <dct_unquantize_mpeg2_inter_mmx+13>: or %cl,0x8b042474(%ecx) 0x082f7ba3 <dct_unquantize_mpeg2_inter_mmx+19>: jne 0x82f7bb9 <dct_unquantize_mpeg2_inter_mmx+41> 0x082f7ba5 <dct_unquantize_mpeg2_inter_mmx+21>: mov %ebx,(%esp) 0x082f7ba8 <dct_unquantize_mpeg2_inter_mmx+24>: mov 0x1b90(%ecx),%edx 0x082f7bae <dct_unquantize_mpeg2_inter_mmx+30>: test %edx,%edx 0x082f7bb0 <dct_unquantize_mpeg2_inter_mmx+32>: jne 0x82f7bc4 <dct_unquantize_mpeg2_inter_mmx+52> 0x082f7bb2 <dct_unquantize_mpeg2_inter_mmx+34>: mov 0x10(%ebp),%ebx 0x082f7bb5 <dct_unquantize_mpeg2_inter_mmx+37>: mov 0x1514(%ecx,%ebx,4),%edx 0x082f7bbc <dct_unquantize_mpeg2_inter_mmx+44>: movzbl 0x158c(%ecx,%edx,1),%eax 0x082f7bc4 <dct_unquantize_mpeg2_inter_mmx+52>: mov 0xc(%ebp),%ebx 0x082f7bc7 <dct_unquantize_mpeg2_inter_mmx+55>: lea (%eax,%eax,1),%edx 0x082f7bca <dct_unquantize_mpeg2_inter_mmx+58>: lea 0x13c0(%edx,%ecx,1),%ecx 0x082f7bd1 <dct_unquantize_mpeg2_inter_mmx+65>: add %edx,%ebx 0x082f7bd3 <dct_unquantize_mpeg2_inter_mmx+67>: neg %edx 0x082f7bd5 <dct_unquantize_mpeg2_inter_mmx+69>: pcmpeqw %mm7,%mm7 0x082f7bd8 <dct_unquantize_mpeg2_inter_mmx+72>: psrlq $0x30,%mm7 End of assembler dump. (gdb) info all-registers eax 0x3f 63 ecx 0x83ec0f0 138330352 edx 0x8404430 138429488 ebx 0x0 0 esp 0xbfff4dd0 0xbfff4dd0 ebp 0xbfff4dd8 0xbfff4dd8 esi 0x6 6 edi 0x841e020 138534944 eip 0x82f7bbc 0x82f7bbc eflags 0x210246 2163270 cs 0x73 115 ss 0x7b 123 ds 0x7b 123 es 0x7b 123 fs 0x0 0 gs 0x33 51 st0 -nan(0x1000000010000) (raw 0xffff0001000000010000) st1 <invalid float value> (raw 0xffff0000000000000000) st2 <invalid float value> (raw 0xffff0000000000000000) st3 <invalid float value> (raw 0xffff0000000000000000) st4 <invalid float value> (raw 0xffff0000000000000000) st5 <invalid float value> (raw 0xffff0000000000000000) st6 <invalid float value> (raw 0xffff0000000000000000) ---Type <return> to continue, or q <return> to quit--- st7 <invalid float value> (raw 0xffff0000000000000000) fctrl 0x37f 895 fstat 0x120 288 ftag 0xaaaa 43690 fiseg 0x0 0 fioff 0x0 0 foseg 0x0 0 fooff 0x0 0 fop 0x0 0 xmm0 {v4_float = {0x0, 0x0, 0x0, 0x0}, v2_double = {0x0, 0x0}, v16_int8 = {0x0 <repeats 16 times>}, v8_int16 = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, v4_int32 = {0x0, 0x0, 0x0, 0x0}, v2_int64 = {0x0, 0x0}, uint128 = 0x00000000000000000000000000000000} xmm1 {v4_float = {0x0, 0x0, 0x0, 0x0}, v2_double = {0x0, 0x0}, v16_int8 = {0x0 <repeats 16 times>}, v8_int16 = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, v4_int32 = {0x0, 0x0, 0x0, 0x0}, v2_int64 = {0x0, 0x0}, uint128 = 0x00000000000000000000000000000000} xmm2 {v4_float = {0x0, 0x0, 0x0, 0x0}, v2_double = {0x0, 0x0}, v16_int8 = {0x0 <repeats 16 times>}, v8_int16 = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, v4_int32 = {0x0, 0x0, 0x0, 0x0}, v2_int64 = {0x0, 0x0}, uint128 = 0x00000000000000000000000000000000} xmm3 {v4_float = {0x0, 0x0, 0x0, 0x0}, v2_double = {0x0, 0x0}, v16_int8 = {0x0 <repeats 16 times>}, v8_int16 = {0x0, 0x0, 0x0, 0x0, 0x0, ---Type <return> to continue, or q <return> to quit--- 0x0, 0x0, 0x0}, v4_int32 = {0x0, 0x0, 0x0, 0x0}, v2_int64 = {0x0, 0x0}, uint128 = 0x00000000000000000000000000000000} xmm4 {v4_float = {0x0, 0x0, 0x0, 0x0}, v2_double = {0x0, 0x0}, v16_int8 = {0x0 <repeats 16 times>}, v8_int16 = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, v4_int32 = {0x0, 0x0, 0x0, 0x0}, v2_int64 = {0x0, 0x0}, uint128 = 0x00000000000000000000000000000000} xmm5 {v4_float = {0x0, 0x0, 0x0, 0x0}, v2_double = {0x0, 0x0}, v16_int8 = {0x0 <repeats 16 times>}, v8_int16 = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, v4_int32 = {0x0, 0x0, 0x0, 0x0}, v2_int64 = {0x0, 0x0}, uint128 = 0x00000000000000000000000000000000} xmm6 {v4_float = {0x0, 0x0, 0x0, 0x0}, v2_double = {0x0, 0x0}, v16_int8 = {0x0 <repeats 16 times>}, v8_int16 = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, v4_int32 = {0x0, 0x0, 0x0, 0x0}, v2_int64 = {0x0, 0x0}, uint128 = 0x00000000000000000000000000000000} xmm7 {v4_float = {0x0, 0x0, 0x0, 0x0}, v2_double = {0x0, 0x0}, v16_int8 = {0x0 <repeats 16 times>}, v8_int16 = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, v4_int32 = {0x0, 0x0, 0x0, 0x0}, v2_int64 = {0x0, 0x0}, uint128 = 0x00000000000000000000000000000000} mxcsr 0x1f80 8064 mm0 {uint64 = 0x1000000010000, v2_int32 = {0x10000, 0x10000}, v4_int16 = {0x0, 0x1, 0x0, 0x1}, v8_int8 = {0x0, 0x0, 0x1, 0x0, 0x0, 0x0, 0x1, 0x0}} mm1 {uint64 = 0x0, v2_int32 = {0x0, 0x0}, v4_int16 = {0x0, 0x0, ---Type <return> to continue, or q <return> to quit--- 0x0, 0x0}, v8_int8 = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}} mm2 {uint64 = 0x0, v2_int32 = {0x0, 0x0}, v4_int16 = {0x0, 0x0, 0x0, 0x0}, v8_int8 = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}} mm3 {uint64 = 0x0, v2_int32 = {0x0, 0x0}, v4_int16 = {0x0, 0x0, 0x0, 0x0}, v8_int8 = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}} mm4 {uint64 = 0x0, v2_int32 = {0x0, 0x0}, v4_int16 = {0x0, 0x0, 0x0, 0x0}, v8_int8 = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}} mm5 {uint64 = 0x0, v2_int32 = {0x0, 0x0}, v4_int16 = {0x0, 0x0, 0x0, 0x0}, v8_int8 = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}} mm6 {uint64 = 0x0, v2_int32 = {0x0, 0x0}, v4_int16 = {0x0, 0x0, 0x0, 0x0}, v8_int8 = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}} mm7 {uint64 = 0x0, v2_int32 = {0x0, 0x0}, v4_int16 = {0x0, 0x0, 0x0, 0x0}, v8_int8 = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}} the stream may be still at ftp://ftp.mplayerhq.hu/MPlayer/incoming/mpeg2_demux_bug.{vob,txt} if it weren't removed. Matthieu

2 1

[patch] *BSD bktr grab support
by jakemsr＠jakemsr.com 25 Jun '05

25 Jun '05

here is the cleaned up patch for gab support for the bktr driver found on *BSD systems. the license is now LGPL, I changed the tabs to 4 spaces, added a couple comments, and took out an #if 0 block and a CVS Id that were accidently left in the last version. -- <jakemsr at jakemsr.com>

4 17

[Ffmpeg-devel-old] LinuxTAG 2005: Linux Video Booth
by alex＠fsn.hu 17 Jun '05

17 Jun '05

Hi, We at the MPlayer team are organizing this time the booth aswell. Linuxtag will be held between June 22-25. Right now we have a booth, some fancy posters and we plan to tell people about our "products", new technologies (such as nut, snow, etc) and maybe give them CD-s with cool stuff on it. Also a "hacking contest" is planned, that means everyone is applying patches or writing code in his project while being at Linuxtag. We except quiet much work, as we did last year. Right now I invite all of you to participate, bring along your poster, beamer, notebook, whatever and spread the word. If you know someone deeply interested in this, but didnt get this mail, please make him/her aware. Thanks and awaiting all of you! -- Alex Beregszaszi e-mail: alex at fsn.hu Free Software Network cell: +36 70 3144424 ------------------------------------------------------- This SF.Net email is sponsored by Yahoo. Introducing Yahoo! Search Developer Network - Create apps using Yahoo! Search APIs Find out how you can build Yahoo! directly into your own Applications - visit http://developer.yahoo.net/?fr=offad-ysdn-ostg-q22005 _______________________________________________ Ffmpeg-devel mailing list Ffmpeg-devel at lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ffmpeg-devel

2 1

mpeg transport streams
by eric＠hernes.ws 13 Jun '05

13 Jun '05

Hi all, I've been working on encoding transport streams. I am playing the files through a digital cable distribution, hardware multiplexer, QAM mod, etc. Attached are three sets of I've been using to generate transport streams that play on the set top. The changes to mpegtsenc.c are * Send full frames as PES packets rather than breaking them into transport sized blocks. * Properly stuff the adaptation field if a packet is short. * Calculate a PCR from the PTS - framesize * Stuff the stream with null packets up to satisfy mux_rate The changes to mpeg12.c / mpegvideo.c are: * Uncomment and finish the option for mpeg2 streams to use the mpeg2 VLC table The changes to ffmpeg.c are: * Specify mux_rate in Kbps (like other rates) * Allow b_frames to be 0. I have a specific decoder target that requires there not be any B-Frames in the source (I realize it is not a fully compliant decoder), and there was nothing in the rest of the code that broke if b_frames was zero. (At least in the MPEG2 codec). -Eric <cid:part1.04050901.05060506 at hernes.ws>

12 51

libavformat MPEG 1/2 "buffer underflow" message
by amayer＠cosy.sbg.ac.at 06 Jun '05

06 Jun '05

Hello all, I'm using FFMPEG to read MPEG-1/2 files, extract the frames (but don't decode them), do some processing, and write them out again using av_write_frame (I don't fully decode the frames). However, when writing out the video frames, FFMPEG produces the message "buffer underflow" several times. This happens with the CVS version from 2005-04-09 (Debian package libavcodeccvs-dev), but not with the released version 0.4.9-pre1 ; the output also looks correct for the latter version (using a self-made MPEG dumper tool); the CVS version puts the audio frames into an earlier position. I've tried av_interleaved_write_frame, but it doesn't help. Actually, how is this function different? I don't really understand the doxygen description. I have looked for the responsible piece of code; it's in libavformat/mpeg.c, line 1019 (e.g. http://www.mplayerhq.hu/~michael/ffmpeg-doxy/mpeg_8c-source.html ): 01019 static int remove_decoded_packets(AVFormatContext <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/structAVFormatContext.html> *ctx, int64_t scr){ 01020 // MpegMuxContext *s = ctx->priv_data; 01021 int i; 01022 01023 for(i=0; i<ctx->nb_streams <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/structAVFormatContext.html#o5>; i++){ 01024 AVStream <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/structAVStream.html> *st = ctx->streams <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/structAVFormatContext.html#o6>[i]; 01025 StreamInfo <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/structStreamInfo.html> *stream = st->priv_data <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/structAVStream.html#o4>; 01026 PacketDesc <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/structPacketDesc.html> *pkt_desc= stream->predecode_packet <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/structStreamInfo.html#o4>; 01027 01028 while(pkt_desc && scr > pkt_desc->dts <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/structPacketDesc.html#o1>){ //FIXME > vs >= 01029 if(stream->buffer_index <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/structStreamInfo.html#o3> < pkt_desc->size <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/structPacketDesc.html#o2> || 01030 stream->predecode_packet <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/structStreamInfo.html#o4> == stream->premux_packet <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/structStreamInfo.html#o5>){ 01031 av_log <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/libavcodec_2utils_8c.html#a58>(ctx, AV_LOG_ERROR, "buffer underflow\n"); 01032 break; 01033 } 01034 stream->buffer_index <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/structStreamInfo.html#o3> -= pkt_desc->size <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/structPacketDesc.html#o2>; 01035 01036 stream->predecode_packet <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/structStreamInfo.html#o4>= pkt_desc->next <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/structPacketDesc.html#o5>; 01037 av_freep <http://www.mplayerhq.hu/%7Emichael/ffmpeg-doxy/libavcodec_2utils_8c.html#a21>(&pkt_desc); 01038 } 01039 } 01040 01041 return 0; 01042 } Is it intended that the while loop can only execute once? av_freep sets the pkt_desc to NULL, so the loop quits after that. I suspect that a "pkt_desc = stream->predecode_packet;" is missing, but then I don't really know what this function actually should do. (By the way, does av_write_frame correctly split the video frames into multiple MPEG PS packets, or does it put the complete frame into one packet? Also, av_read_frame splits up the audio packets into several frames; does av_write_frame correctly put these together again?) thanks & greetings August Mayer

2 3