[Ffmpeg-devel-irc] ffmpeg-devel.log.20120514

[burek]

[01:00] <lemonzest> how do i get ffmpeg to put seek points into an ogg with theora/vorbis in, it lags sometimes when seeking in the media player, yet is fine if i just let it play
[01:43] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * r555e2921dd 10ffmpeg/libavcodec/flac_parser.c: 
[01:43] <CIA-63> ffmpeg: flac_parser: speedup fifo realloc
[01:43] <CIA-63> ffmpeg: Fixes Ticket1121
[01:43] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[03:05] <CIA-63> ffmpeg: 03Diego Elio Pettenò 07master * r7b089b79e6 10ffmpeg/doc/fate.texi: 
[03:05] <CIA-63> ffmpeg: doc: fix link to fate.libav.org
[03:05] <CIA-63> ffmpeg: Signed-off-by: Diego Elio Pettenò <flameeyes at flameeyes.eu>
[03:05] <CIA-63> ffmpeg: Signed-off-by: Luca Barbato <lu_zero at gentoo.org>
[03:05] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * r0db047cd0f 10ffmpeg/: 
[03:05] <CIA-63> ffmpeg: Merge remote-tracking branch 'qatar/master'
[03:05] <CIA-63> ffmpeg: * qatar/master:
[03:05] <CIA-63> ffmpeg:  doc: fix link to fate.libav.org
[03:05] <CIA-63> ffmpeg: Conflicts:
[03:05] <CIA-63> ffmpeg:  doc/fate.texi
[03:05] <CIA-63> ffmpeg: Merged-by: Michael Niedermayer <michaelni at gmx.at>
[03:05] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * r04064e1c20 10ffmpeg/libavformat/mpegts.c: 
[03:05] <CIA-63> ffmpeg: mpegts: fix handling of files smaller than 8kb
[03:05] <CIA-63> ffmpeg: Fixes Ticket1094
[03:05] <CIA-63> ffmpeg: Analyzed-by: mjs973
[03:05] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[06:47] <burek> does ffmpeg support PJPG pixel format input from webcam?
[06:48] <burek> http://pastebin.com/theLdVk6
[06:57] <floater_f432c> hi
[06:57] <floater_f432c> http://pastebin.com/theLdVk6
[06:58] <floater_f432c> i see burek pasted it, anybody can give a hint?
[07:01] <floater_f432c> was thinking about helping in the selection of formats, but i'm confused ffmpeg -f v4l2 -input_format mpjpeg -pix_fmt rgb24 -s 640x480 -r 30 -i /dev/video1 -vcodec libx264 test.mkv
[12:28] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * r4991cbafaf 10ffmpeg/libavcodec/mpegvideo_enc.c: 
[12:28] <CIA-63> ffmpeg: mpegvideo_enc: Fix chroma edge size
[12:28] <CIA-63> ffmpeg: Fixes ticket1303
[12:28] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[13:01] <CIA-63> ffmpeg: 03Stefano Sabatini 07master * ra0fd322195 10ffmpeg/ffprobe.c: 
[13:01] <CIA-63> ffmpeg: ffprobe: remove unused print_format_entry() function
[13:01] <CIA-63> ffmpeg: The function was introduced in 653d117c, but is not required.
[13:01] <CIA-63> ffmpeg: 03Stefano Sabatini 07master * r6cd06bd22c 10ffmpeg/ffprobe.c: 
[13:01] <CIA-63> ffmpeg: ffprobe: simplify/fix logic for the -show_format_entry option
[13:01] <CIA-63> ffmpeg: Do not skip chapter header/footer printing, which breaks the parsing for
[13:01] <CIA-63> ffmpeg: most writers.
[13:01] <CIA-63> ffmpeg: 03Stefano Sabatini 07master * r3946187d60 10ffmpeg/ (doc/ffprobe.texi ffprobe.c): 
[13:01] <CIA-63> ffmpeg: ffprobe: add "nokey" option to default writer
[13:01] <CIA-63> ffmpeg: Help simplifying parsing in certain cases.
[13:01] <CIA-63> ffmpeg: 03Stefano Sabatini 07master * rf48f03a400 10ffmpeg/ (doc/ffprobe.texi ffprobe.c tests/fate-run.sh): 
[13:01] <CIA-63> ffmpeg: ffprobe: add "noprint_wrappers" option to default writer
[13:01] <CIA-63> ffmpeg: The option is useful for simplifying parsing.
[13:01] <CIA-63> ffmpeg: Also use the new option in fate, in order to fix the regression
[14:16] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * r5931c754b8 10ffmpeg/libavformat/oggdec.c: 
[14:16] <CIA-63> ffmpeg: ogg: in ogg_get_length() start from data_offset not 0.
[14:16] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[14:16] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * r8443082db6 10ffmpeg/libavformat/oggdec.c: 
[14:16] <CIA-63> ffmpeg: ogg: rewrite first timestamp reading code.
[14:16] <CIA-63> ffmpeg: this is simpler and closer to how timestamps are found for seeking.
[14:16] <CIA-63> ffmpeg: Fixes Ticket1186
[14:16] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[14:36] <michaelni> ubitux, on the sharedlib fate box libavresample is missing in Wl,-rpath= i think, or alternatively you can disable avresample
[14:38] <ubitux> ah, right
[14:38] <ubitux> forgot to change that
[14:39] <ubitux> should be fixed in the next run
[14:39] <ubitux> thanks :)
[17:44] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * rc69b6649cf 10ffmpeg/libavformat/utils.c: 
[17:44] <CIA-63> ffmpeg: update_initial_durations: add some checks to ensure things match up.
[17:44] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[17:44] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * r6d49e819be 10ffmpeg/ffmpeg.c: 
[17:44] <CIA-63> ffmpeg: ffmpeg: revert 492cc1
[17:44] <CIA-63> ffmpeg: This restores correct block_align values in case of stream copy.
[17:44] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[17:48] <ubitux> michaelni: could you use av_ts2str() for the pts/dts?
[17:48] <ubitux> (i'm thinking of c69b6649cf)
[17:50] <michaelni> sure, ill chnage it
[17:52] <ubitux> btw, sharedlib fate fixed
[18:01] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * r381cc4b1e6 10ffmpeg/libavformat/utils.c: 
[18:01] <CIA-63> ffmpeg: update_initial_durations: use av_ts2str()
[18:01] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[18:05] <ubitux> oh i didn't know about the -fflags +genpts
[18:05] <ubitux> thanks michaelni, i'll test that soon :)
[18:05] <michaelni> please do, iam sure there are cases left that fail ...
[18:06] <ubitux> btw, ruggles had a look to the block_align stuff
[18:06] <michaelni> also about block_align, IMHO reverting the commit was correct, we shouldnt feed a muxer with nonsense
[18:06] <ubitux> and iirc the outcome was that not copying the block_align at all solved the issue
[18:06] <ubitux> but broke for wma for instance
[18:07] <ubitux> (so basically the "fix" was to drop codec->block_align        = icodec->block_align;)
[18:07] <ubitux> i don't understand well the issue so it's hard to tell
[18:09] <michaelni> we will see what works best
[18:09] <michaelni> whichever it is ill make sure ffmpeg.c will use it :)
[18:11] <ubitux> it doesn't look like they care much about the regression anyway so i'm fine with the revert anyway
[18:57] <CIA-63> ffmpeg: 03Nicolas George 07master * r7a44223319 10ffmpeg/ (doc/filter_design.txt libavfilter/avfilter.h): 
[18:57] <CIA-63> ffmpeg: lavfi: document filter design subtleties.
[18:57] <CIA-63> ffmpeg: The details on reference ownership and permissions are missing.
[19:20] <CIA-63> ffmpeg: 03Clément BSsch 07master * r9548deeea9 10ffmpeg/libavutil/ (avutil.h bprint.c bprint.h): lavu/bprint: add av_bprint_reset().
[19:20] <CIA-63> ffmpeg: 03Clément BSsch 07master * rb545b947dd 10ffmpeg/ffprobe.c: 
[19:20] <CIA-63> ffmpeg: ffprobe: replace fast_asprintf() with bprint utils.
[19:20] <CIA-63> ffmpeg: Also remove the unused print_fmt_opt() in the process.
[19:51] <CIA-63> ffmpeg: 03Nicolas George 07master * r83f954e859 10ffmpeg/configure: configure: check for pthread_cancel.
[19:51] <CIA-63> ffmpeg: 03Nicolas George 07master * r281bde2789 10ffmpeg/libavformat/udp.c: 
[19:51] <CIA-63> ffmpeg: udp: check for HAVE_PTHREAD_CANCEL instead of HAVE_PTHREADS.
[19:51] <CIA-63> ffmpeg: Some environments, for example Android, pretend to have pthreads
[19:51] <CIA-63> ffmpeg: but actually have only a partial implementation.
[21:12] <CIA-63> ffmpeg: 03Mans Rullgard 07master * r1c0e8b94ad 10ffmpeg/tests/ (fate/demux.mak ref/fate/mtv): 
[21:12] <CIA-63> ffmpeg: fate: mtv: disable video decoding
[21:12] <CIA-63> ffmpeg: This test contains raw rgb565le video. Converting to rgb24 serves
[21:12] <CIA-63> ffmpeg: no useful purpose here.
[21:12] <CIA-63> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[21:12] <CIA-63> ffmpeg: 03Mans Rullgard 07master * r0c26380f4f 10ffmpeg/tests/ (fate/pcm.mak ref/fate/ea-mad-pcm-planar ref/fate/pcm-planar): 
[21:12] <CIA-63> ffmpeg: fate: pcm-planar: disable video
[21:12] <CIA-63> ffmpeg: This file has eamad video which is tested elsewhere.
[21:12] <CIA-63> ffmpeg: Also rename the test to reflect this change.
[21:12] <CIA-63> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[21:12] <CIA-63> ffmpeg: 03Mans Rullgard 07master * r63322d8c89 10ffmpeg/tests/ (fate/qtrle.mak ref/fate/qtrle-1bit ref/fate/qtrle-2bit): 
[21:12] <CIA-63> ffmpeg: fate: qtrle: disable audio in all tests
[21:12] <CIA-63> ffmpeg: These files contain mace6 audio which is tested elsewhere.
[21:12] <CIA-63> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[21:12] <CIA-63> ffmpeg: 03Mans Rullgard 07master * rf4d2b93be4 10ffmpeg/tests/ (3 files in 2 dirs): 
[21:12] <CIA-63> ffmpeg: fate: truemotion1: disable audio
[21:12] <CIA-63> ffmpeg: These tests include adpcm-ima-dk3 audio which is tested elsewhere.
[21:12] <CIA-63> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[21:12] <CIA-63> ffmpeg: 03Mans Rullgard 07master * rbd7ac952b1 10ffmpeg/tests/ (4 files in 2 dirs): 
[21:12] <CIA-63> ffmpeg: fate: separate delphine-cin audio and video tests
[21:12] <CIA-63> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[21:12] <CIA-63> ffmpeg: 03Mans Rullgard 07master * r95939bf565 10ffmpeg/tests/ (4 files in 2 dirs): 
[21:12] <CIA-63> ffmpeg: fate: split off adpcm-ima-ws test from vqa-cc
[21:12] <CIA-63> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[21:12] <CIA-63> ffmpeg: 03Mans Rullgard 07master * r838521e110 10ffmpeg/tests/ (fate/adpcm.mak fate/demux.mak ref/fate/funcom-iss): 
[21:12] <CIA-63> ffmpeg: fate: rename funcom-iss to adpcm-ima-iss
[21:12] <CIA-63> ffmpeg: This matches the name of the codec.
[21:26] <ubitux> michaelni: do you mind having a look to the checkheader patchset ("[PATCHSET] fix make checkheaders + misc header fixes") if that looks better to you?
[21:26] <ubitux> (i know it's a bit early for a ping but well... :))
[21:46] <CIA-63> ffmpeg: 03Nicolas George 07master * r440af105f2 10ffmpeg/libavfilter/ (avfiltergraph.c avfiltergraph.h version.h): lavfi: add avfilter_graph_set_auto_convert().
[21:46] <CIA-63> ffmpeg: 03Nicolas George 07master * r9cf5bcded0 10ffmpeg/ (Changelog doc/ffmpeg.texi ffmpeg.c): ffmpeg: accept + prefix to -pix_fmt option to disable automatic conversions.
[23:37] <ubitux> beastd: i think we must not change the av_fifo_realloc2() because it will affect av_fifo_size()
[23:37] <ubitux> (i guess)
[23:37] <ubitux> the grow is in order to limit realloc call by automatically growing "faster"
[23:37] <ubitux> afaict
[23:37] <beastd> hmm. seems so
[23:38] <beastd> i meant are the old vs req size checks needed in av_fifo_grow too?
[23:39] <beastd> I might well be missing one or more points. But I decided to ask as I probably would have not found the time to reply soon.
[23:44] <CIA-63> ffmpeg: 03Matthieu Bouron 07master * r4c73b3b1e0 10ffmpeg/ffmpeg.c: 
[23:44] <CIA-63> ffmpeg: ffmpeg: affect computed filtered frame pts after filling frame from buffer
[23:44] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:44] <CIA-63> ffmpeg: 03Marton Balint 07master * r94a9ac1277 10ffmpeg/ (4 files in 2 dirs): (log message trimmed)
[23:44] <CIA-63> ffmpeg: lavf: add av_guess_frame_sample_aspect_ratio function
[23:44] <CIA-63> ffmpeg: Guesses the sample aspect ratio of a frame, based on both the stream and the
[23:44] <CIA-63> ffmpeg: frame aspect ratio.
[23:44] <CIA-63> ffmpeg: Since the frame aspect ratio is set by the codec but the stream aspect ratio
[23:44] <CIA-63> ffmpeg: is set by the demuxer, these two may not be equal. This function tries to
[23:44] <CIA-63> ffmpeg: return the value that you should use if you would like to display the frame.
[23:51] <ubitux> hey saste :)
[23:51] <saste> hey clement
[23:51] <saste> internet is back since this morning
[23:52] <ubitux> yes but internet is slow :(
[23:52] <ubitux> !@#$ isp.
[23:52] <saste> i have a patch for removing packing negotiation
[23:53] <saste> i don't know what's the status about >=8 channels and if the audio filters code should be adjusted
[23:53] <saste> (uhm condition is strictly >)
[23:58] <ubitux> michaelni: i don't know if that's only because of the block_align revert, but it seems to solve all my a/v desync issues with -c copy
[23:58] <ubitux> so, thanks :)
[00:00] --- Tue May 15 2012