[Ffmpeg-devel-irc] ffmpeg-devel.log.20120520

[burek]

[12:36] <CIA-105> ffmpeg: 03Anton Khirnov 07master * r39885a4bd7 10ffmpeg/avconv.c: avconv: reindent
[12:36] <CIA-105> ffmpeg: 03Mans Rullgard 07master * r63dcd16d56 10ffmpeg/tests/ (38 files in 2 dirs): 
[12:36] <CIA-105> ffmpeg: fate: fix dependencies on SAMPLES being set
[12:36] <CIA-105> ffmpeg: This allows fate to run without errors with or without SAMPLES
[12:36] <CIA-105> ffmpeg: being set.
[12:36] <CIA-105> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[12:36] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r31f3c94760 10ffmpeg/: (log message trimmed)
[12:36] <CIA-105> ffmpeg: Merge remote-tracking branch 'qatar/master'
[12:36] <CIA-105> ffmpeg: * qatar/master:
[12:36] <CIA-105> ffmpeg:  fate: fix dependencies on SAMPLES being set
[12:36] <CIA-105> ffmpeg:  fate: remove unnecessary libavfilter dependency
[12:36] <CIA-105> ffmpeg: Conflicts:
[12:36] <CIA-105> ffmpeg:  ffmpeg.c
[12:36] <CIA-105> ffmpeg: 03Anton Khirnov 07master * r74b961db77 10ffmpeg/ (48 files in 2 dirs): 
[12:36] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r3cf5591344 10ffmpeg/ffmpeg.c: 
[12:36] <CIA-105> ffmpeg: ffmpeg: reintroducing check_recording_time()
[12:36] <CIA-105> ffmpeg: This reduces the difference to qatar by 27 lines
[12:36] <CIA-105> ffmpeg: No functional change, the code should be unused
[12:36] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[12:36] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * ref7ac8e8ec 10ffmpeg/ffmpeg.c: 
[12:36] <CIA-105> ffmpeg: ffmpeg: set the video filters timebase to the demuxers.
[12:36] <CIA-105> ffmpeg: Inspired-by a change from Anton Khirnov
[12:36] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[12:36] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r9ce3d03fc2 10ffmpeg/ffmpeg.c: 
[12:36] <CIA-105> ffmpeg: ffmpeg: cosmetics to reduce difference to qatar by 101 lines
[12:36] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[12:36] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r13c254a216 10ffmpeg/ffmpeg.c: 
[12:36] <CIA-105> ffmpeg: ffmpeg: random cosmectics to reduce diff to qatar
[12:36] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[12:36] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r3ba90d9e97 10ffmpeg/ffmpeg.c: 
[12:37] <CIA-105> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[12:37] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r953df5d962 10ffmpeg/ffmpeg.c: 
[12:37] <CIA-105> ffmpeg: ffmpeg: drop pkt_pts
[12:37] <CIA-105> ffmpeg: Reduces diff to qatar by 9 lines
[12:37] <CIA-105> ffmpeg: Should not change anything
[12:37] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[12:37] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * rc53326afad 10ffmpeg/ffmpeg.c: 
[12:37] <CIA-105> ffmpeg: ffmpeg: move ost->frame_rate setting code before the filtergraph stuff
[12:37] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[14:34] <ubitux> voice synthesis *_*
[15:54] <CIA-105> ffmpeg: 03Nicolas George 07master * r6adefeef45 10ffmpeg/libavfilter/buffersrc.c: buffersrc: cosmetic: reindent.
[15:55] <CIA-105> ffmpeg: 03Nicolas George 07master * rea07063fd8 10ffmpeg/ (cmdutils.c doc/avtools-common-opts.texi): 
[15:55] <CIA-105> ffmpeg: cmdutils: allow to specify a stream by its ID.
[15:55] <CIA-105> ffmpeg: The stream can be specified as "#129" or "#0x81".
[15:55] <CIA-105> ffmpeg: It is especially useful for VOBs dumped from a DVD,
[15:55] <CIA-105> ffmpeg: where the language-id mapping is available externally
[15:55] <CIA-105> ffmpeg: and the probing can find the streams in a random order.
[15:55] <ubitux> saste: i really have a/v sync troubles with the showwaves filter
[15:55] <ubitux> at least with the example you provide
[15:55] <ubitux> (using ffplay)
[15:56] <saste> what is the example?
[15:56] <saste> yes i noticed it may have problems when the size is small
[15:56] <saste> usually increasing the output size it does better
[15:56] <ubitux> ./ffplay -f lavfi "amovie=a.mp3,asplit[out0],showwaves[out1]"
[15:57] <saste> yes, try to increase the size from the default 320x240 -> 800x200 e.g.
[15:57] <ubitux> i wonder if we won't need to resample what we display
[15:57] <saste> why?
[15:58] <ubitux> isn't this caused by the fact we don't have enough samples (or too much) to display?
[15:59] <saste> yes
[15:59] <saste> you could define the accepted sample rate depending on the size
[16:00] <saste> some kind of optional heuristic
[16:01] <ubitux> or just do some resampling for what's being displayed?
[16:02] <saste> that's also possible but looks more complicated (and add a dependency on libswr in case you don't want to do it by hand)
[16:04] <ubitux> (that's indeed better with a larger window)
[16:06] <saste> if the windows is small you force a very high frame rate, and SDL can't cope well with that
[16:07] <saste> resampling to a proper value should definitively help with that
[16:11] <saste> well i suppose we have to add a proper INI ffprobe writer
[16:13] <ubitux> yep
[16:13] <saste> ubitux: do you have any comment on syntax.texi?
[16:13] <ubitux> i was wondering about that
[16:14] <saste> apart that it misses examples (which I plan to add soon or later)
[16:16] <ubitux> i'm going to have a look
[16:17] <saste> regarding aflite I dislike the fact that it synthetizes audio during the init stage, which is not very nice in case the file to synth is big (e.g. a book text) and you want to play/stream in RT
[16:18] <saste> but maybe i just need to check better the flite API, hopefully there is already some sort of "streaming" API
[16:19] <ubitux> i think i would mention the exception of the scale filter in the video size syntax
[16:21] <ubitux> otherwise it LGTM
[16:21] <saste> dunno, i don't like to document special case in generic documentation, this way you create untraceable dependencies
[16:21] <ubitux> good point, ok :)
[16:21] <saste> better to document the special case in the scale docs
[16:21] <ubitux> libaflite isn't packaged in my distro; where does it come from?
[16:21] <ubitux> asterisk?
[16:22] <saste> uhmm no debian/ubuntu?
[16:22] <saste> here it is flite1-dev
[16:22] <saste> stands for festival-lite
[16:23] <ubitux> ah festival right
[16:23] <ubitux> got it now
[16:23] <ubitux> extra/festival 2.1-3 A general multi-lingual speech synthesis system
[16:23] <ubitux> (archlinux)
[16:28] <CIA-105> ffmpeg: 03Stefano Sabatini 07master * rcd945d2112 10ffmpeg/libavutil/tree.h: lavu/tree: perform minor fixes to av_tree_insert() documentation
[16:29] <ubitux> about the INI writer
[16:29] <ubitux> can't it be achieved with some options in the default writer?
[16:29] <ubitux> like "ini_compliant", "libav_compliant" or sth?
[16:30] <ubitux> (i wonder if the way the subsections are done in the "valid" libav ini format really is standard)
[16:34] <ubitux> saste: btw, Luca is pretty well aware of the current state of ffprobe since long, and i tried again to remind him that he could have just pick the stuff in ffmpeg, but i get ignored again
[16:35] <ubitux> feel free to try by yourself
[16:35] <ubitux> i think i'll stop trying some coop for now.
[16:35] <saste> maybe he don't want to pick dozens patches for implementing the only feature that he needs
[16:36] <ubitux> well, it looks like he's rewriting the writers system
[16:36] <ubitux> not just a simple json hack
[16:36] <ubitux> he could have reconstruct a history without much pain
[16:37] <saste> NIH syndrome?
[16:37] <ubitux> yeah exactly...
[16:37] <ubitux> also it's too bad he is breaking the default output
[16:37] <ubitux> where we tried to keep it to not break user scripts..
[16:40] <CIA-105> ffmpeg: 03Piotr Bandurski 07master * rcb7b0f3573 10ffmpeg/libavcodec/flicvideo.c: 
[16:40] <CIA-105> ffmpeg: flicvideo: support 256-bytes extradata
[16:40] <CIA-105> ffmpeg: sample:
[16:40] <CIA-105> ffmpeg: http://www.datafilehost.com/download-ab1156a9.html
[16:40] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:40] <CIA-105> ffmpeg: 03Piotr Bandurski 07master * r718b90d4b9 10ffmpeg/libavcodec/aasc.c: 
[16:40] <CIA-105> ffmpeg: aasc: support 16bpp
[16:40] <CIA-105> ffmpeg: sample:
[16:40] <CIA-105> ffmpeg: http://www.datafilehost.com/download-b881f3f6.html
[16:40] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:40] <CIA-105> ffmpeg: 03Piotr Bandurski 07master * r4d18f4c15b 10ffmpeg/libavformat/id3v2.c: 
[16:40] <CIA-105> ffmpeg: id3v2: reject APICs with size of 0
[16:40] <CIA-105> ffmpeg: fixes ticket #1324
[16:40] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:40] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * rcf9812d22b 10ffmpeg/ffmpeg.c: 
[16:40] <CIA-105> ffmpeg: Fixes Ticket122
[16:40] <CIA-105> ffmpeg: Reviewed-by: Marton Balint <cus at passwd.hu>
[16:40] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:40] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r9225513242 10ffmpeg/libavfilter/vf_scale.c: 
[16:40] <CIA-105> ffmpeg: vf_scale: support dynamically changing input parameters.
[16:40] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:40] <CIA-105> ffmpeg: 03Jean First 07master * rf7985f348f 10ffmpeg/libavcodec/exr.c: (log message trimmed)
[16:40] <CIA-105> ffmpeg: exr: fix header parsing
[16:40] <CIA-105> ffmpeg: the header in the sample provided for ticket #1306 is not parsed correctly and thus
[16:40] <CIA-105> ffmpeg: ffmpeg tries to decode the sample instead of abording the decoding.
[16:40] <CIA-105> ffmpeg: I tested it with two other exr samples I have - one float, one half float - and
[16:41] <CIA-105> ffmpeg: they still decode correctly.
[16:41] <CIA-105> ffmpeg: Signed-off-by: Jean First <jeanfirst at gmail.com>
[17:12] <ubitux> saste: would it make sense to make a single filter for showwaves and showspectrum?
[17:12] <ubitux> or at least share the same file?
[17:13] <ubitux> maybe not a good idea since they can get bigger...
[17:13] <ubitux> well, we'll see.
[17:15] <saste> they may use a different display logic, so the options may be different
[17:16] <saste> but sure i suppose some parts of the code may be shared, so having them in the same file may help
[18:45] <CIA-105> ffmpeg: 03Nicolas George 07master * r2ec3e5767e 10ffmpeg/libavutil/ (opt.c opt.h): opt: add AV_OPT_TYPE_IMAGE_SIZE.
[18:45] <CIA-105> ffmpeg: 03Nicolas George 07master * r16a44b41ab 10ffmpeg/libavfilter/vsrc_testsrc.c: vsrc_testsrc: use AV_OPT_TYPE_IMAGE_SIZE.
[18:45] <CIA-105> ffmpeg: 03Nicolas George 07master * r0ef4924db4 10ffmpeg/libavdevice/sdl.c: sdl: use AV_OPT_TYPE_IMAGE_SIZE.
[18:45] <CIA-105> ffmpeg: 03Nicolas George 07master * r3e63071115 10ffmpeg/libavdevice/v4l2.c: v4l2: use AV_OPT_TYPE_IMAGE_SIZE.
[18:45] <CIA-105> ffmpeg: 03Nicolas George 07master * r5732f40e3f 10ffmpeg/libavdevice/x11grab.c: x11grab: use AV_OPT_TYPE_IMAGE_SIZE.
[18:45] <CIA-105> ffmpeg: 03Nicolas George 07master * rdc6a02de01 10ffmpeg/libavfilter/vsrc_life.c: vsrc_life: use AV_OPT_TYPE_IMAGE_SIZE.
[18:45] <CIA-105> ffmpeg: 03Nicolas George 07master * r4a5d18b6e2 10ffmpeg/libavfilter/vsrc_mandelbrot.c: vsrc_mandelbrot: use AV_OPT_TYPE_IMAGE_SIZE.
[18:45] <CIA-105> ffmpeg: 03Nicolas George 07master * r9bcd70aa23 10ffmpeg/libavdevice/bktr.c: bktr: use AV_OPT_TYPE_IMAGE_SIZE.
[18:45] <CIA-105> ffmpeg: 03Nicolas George 07master * rca554fbc2f 10ffmpeg/libavfilter/vsrc_cellauto.c: vsrc_cellauto: use AV_OPT_TYPE_IMAGE_SIZE.
[18:45] <CIA-105> ffmpeg: 03Nicolas George 07master * r2987efe9ba 10ffmpeg/libavfilter/vf_ass.c: vf_ass: use AV_OPT_TYPE_IMAGE_SIZE.
[19:37] <CIA-105> ffmpeg: 03Marton Balint 07master * rb36a0a80fc 10ffmpeg/libavformat/ (utils.c version.h): 
[19:37] <CIA-105> ffmpeg: lavf: also use codec aspect ratio in av_guess_sample_aspect_ratio
[19:37] <CIA-105> ffmpeg: Use codec aspect ratio for frame aspect ratio if AVFrame is NULL.
[19:37] <CIA-105> ffmpeg: Signed-off-by: Marton Balint <cus at passwd.hu>
[19:37] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:37] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r318e395764 10ffmpeg/ffmpeg.c: 
[19:37] <CIA-105> ffmpeg: ffmpeg: drop get_sync_ipts() and drop related special case for video
[19:37] <CIA-105> ffmpeg: Idea found in a commit from Anton Khirnov
[19:37] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:37] <CIA-105> ffmpeg: 03Matthieu Bouron 07master * r432fe9a38a 10ffmpeg/ffmpeg.c: 
[19:37] <CIA-105> ffmpeg: ffmpeg: fix a crash with complex filters when pix_fmt is not specified
[19:37] <CIA-105> ffmpeg: ffmpeg -i in.mxf -filter_complex "[0:0]fieldorder=tff" out.wav will
[19:37] <CIA-105> ffmpeg: fail with an error message instead of crashing.
[19:37] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:00] <CIA-105> ffmpeg: 03Mans Rullgard 07master * r81ad97eeda 10ffmpeg/libavformat/utils.c: 
[23:00] <CIA-105> ffmpeg: lavf: make output format matching case insensitive
[23:00] <CIA-105> ffmpeg: This is consistent with how input formats are matched.
[23:00] <CIA-105> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[23:00] <CIA-105> ffmpeg: 03Mans Rullgard 07master * r68aef0b481 10ffmpeg/ (6 files in 2 dirs): 
[23:00] <CIA-105> ffmpeg: lavf: change some (de)muxer names to lowercase
[23:00] <CIA-105> ffmpeg: This is consistent with other format names.
[23:00] <CIA-105> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[23:00] <CIA-105> ffmpeg: 03Kostya Shishkov 07master * rb37d945dd4 10ffmpeg/libavcodec/mpegaudiodec.c: 
[23:00] <CIA-105> ffmpeg: mp3: fix start band index for block type 2 in 8kHz audio
[23:00] <CIA-105> ffmpeg: In hybrid frames long window part ends at 36 samples for most of the cases
[23:00] <CIA-105> ffmpeg: but at 72 for 8kHz case. For some reason decoder assumed it's 48 or even 36
[23:00] <CIA-105> ffmpeg: samples, which caused wrong bitstream decoding for such blocks.
[23:00] <CIA-105> ffmpeg: l3_25207.mpg from conformance suite demonstrates it the best.
[23:00] <CIA-105> ffmpeg: 03Kostya Shishkov 07master * raa372cf470 10ffmpeg/libavcodec/ (indeo4.c indeo5.c ivi_common.c ivi_common.h): factor out common decoding code for Indeo 4 and Indeo 5
[23:00] <CIA-105> ffmpeg: 03Kostya Shishkov 07master * ra6e4ac40a6 10ffmpeg/libavcodec/ (ivi_common.c ivi_common.h): indeo: track tile macroblock size
[23:00] <CIA-105> ffmpeg: 03Kostya Shishkov 07master * rfe7a37c36f 10ffmpeg/libavcodec/ivi_common.c: indeo: check custom Huffman tables for errors
[23:00] <CIA-105> ffmpeg: 03Kostya Shishkov 07master * r9603738221 10ffmpeg/libavcodec/ivi_common.c: indeo: check that band output buffer exists
[23:00] <CIA-105> ffmpeg: 03Kostya Shishkov 07master * rcf61aaaca1 10ffmpeg/libavcodec/ (ivi_common.c ivi_common.h): indeo: check for invalid motion vectors
[23:01] <CIA-105> ffmpeg: 03Mans Rullgard 07master * r4852cb37ce 10ffmpeg/tests/Makefile: 
[23:01] <CIA-105> ffmpeg: fate: fix dependencies for non-SAMPLES avconv tests
[23:01] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r349c62410b 10ffmpeg/: (log message trimmed)
[23:01] <CIA-105> ffmpeg: Merge remote-tracking branch 'qatar/master'
[23:01] <CIA-105> ffmpeg: * qatar/master:
[23:01] <CIA-105> ffmpeg:  indeo: Make ivi_calc_band_checksum() static, it is only used in one file.
[23:01] <CIA-105> ffmpeg:  indeo: Drop unused debug function ivi_check_band().
[23:01] <CIA-105> ffmpeg:  avcodec/utils: cast a function argument to shut up a compiler warning
[23:01] <CIA-105> ffmpeg:  truemotion1: remove disabled code
[23:01] <CIA-105> ffmpeg: 03Kostya Shishkov 07master * r23ba1503f2 10ffmpeg/libavcodec/ivi_common.c: indeo: clear allocated band buffers
[00:00] --- Mon May 21 2012