[Ffmpeg-devel-irc] ffmpeg-devel.log.20120624

[burek]

[02:09] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * rcaf7381d50 10ffmpeg/ffmpeg.c: 
[02:09] <CIA-119> ffmpeg: ffmpeg: fix mixup of old and new sinks
[02:09] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[02:09] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * rda728d5d2e 10ffmpeg/ffmpeg.c: 
[02:09] <CIA-119> ffmpeg: ffmpeg: fix FF_API_OLD_VSINK_API=0 case
[02:09] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[02:20] <CIA-119> ffmpeg: 03Anton Khirnov 07master * rfa06623930 10ffmpeg/libavfilter/ (af_aformat.c af_anull.c audio.c audio.h): (log message trimmed)
[02:20] <CIA-119> ffmpeg: lavfi/audio: eliminate ff_default_filter_samples().
[02:20] <CIA-119> ffmpeg: It currently does the following:
[02:20] <CIA-119> ffmpeg: 1) get a zeroed audio buffer
[02:20] <CIA-119> ffmpeg: 2) copy some properties (but not the data) of the input buffer to it
[02:20] <CIA-119> ffmpeg: 3) pass this buffer to the output filter
[02:20] <CIA-119> ffmpeg: This looks useless and is indeed not used by any filters, therefore
[02:20] <CIA-119> ffmpeg: 03Anton Khirnov 07master * r58b049f2fa 10ffmpeg/ (6 files in 2 dirs): 
[02:20] <CIA-119> ffmpeg: lavfi: support automatically inserting the fifo filter when needed.
[02:20] <CIA-119> ffmpeg: This breaks libavfilter ABI.
[02:20] <CIA-119> ffmpeg: 03Anton Khirnov 07master * rf75be9856a 10ffmpeg/libavfilter/ (avfilter.h fifo.c): 
[02:20] <CIA-119> ffmpeg: lavfi: allow audio filters to request a given number of samples.
[02:20] <CIA-119> ffmpeg: This makes synchronization simpler for filters with multiple inputs.
[02:21] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r1c60088885 10ffmpeg/: (log message trimmed)
[02:21] <CIA-119> ffmpeg: Merge remote-tracking branch 'qatar/master'
[02:21] <CIA-119> ffmpeg: * qatar/master:
[02:21] <CIA-119> ffmpeg:  x86: Only use optimizations with cmov if the CPU supports the instruction
[02:21] <CIA-119> ffmpeg:  x86: Add CPU flag for the i686 cmov instruction
[02:21] <CIA-119> ffmpeg:  x86: remove unused inline asm macros from dsputil_mmx.h
[02:21] <CIA-119> ffmpeg:  x86: move some inline asm macros to the only places they are used
[02:21] <CIA-119> ffmpeg: 03Alex Converse 07master * r41e637e449 10ffmpeg/ (5 files in 3 dirs): 
[02:21] <CIA-119> ffmpeg: lavfi: Add the af_channelmap audio channel mapping filter.
[02:21] <CIA-119> ffmpeg: Inspired by MPlayer's af_channels filter and SoX's remix effect.
[02:23] <Compn> too many filters
[02:23] <Compn> is af_channels like af pan ? :P
[02:23] Action: Compn cant recall
[02:24] <Compn> mplayer has bad naming of filters, that is
[02:24] <Compn> 'extrastereo' 'channels' 'pan' etc
[02:32] <durandal_1707> michaelni: In file included from libavcodec/h263dec.c:30: ./libavutil/cpu.h:45:5: warning: 'LIBAVUTIL_VERSION_MAJOR' is not defined, evaluates to 0 [-Wundef]
[02:39] <durandal_1707> oh bunch of warnings pops up
[02:43] <michaelni> moment, ill fix it
[03:21] <michaelni> durandal_1707, patchset to fix the warning is on the ML
[03:25] <durandal_1707> well if it compiles without warnings it is ok
[03:26] <durandal_1707> perhaps lavu needs minor bump because new header is installed?
[03:27] <durandal_1707> and mentined in doc/APIchanges?
[03:44] <CIA-119> ffmpeg: 03Hendrik Leppkes 07master * rb5e6aa6c33 10ffmpeg/libavcodec/dxva2_h264.c: 
[03:44] <CIA-119> ffmpeg: dxva2_h264: fix signaling of mbaff frames
[03:44] <CIA-119> ffmpeg: The MBAFF flag may only be signaled if we're actually dealing with
[03:44] <CIA-119> ffmpeg: a full frame, and not singular fields, as it can happen in mixed content.
[03:44] <CIA-119> ffmpeg: Reviewed-by: Derek Buitenhuis <derek.buitenhuis at gmail.com>
[03:44] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[11:17] <ubitux> anyone to have a look to patch 5-11 in the subtitles patchset?
[11:19] <ubitux> i mean, those not previously reviewed already
[11:20] <ubitux> basically 6, 10 and 11
[11:49] <icenix> hey guys, i have a question that is prolly pretty simple about using ffserver to stream, would anyone be able to help me for a second?
[12:08] <burek> icenix, try #ffmpeg channel, this one is for ffmpeg development, not ffmpeg usage :)
[12:37] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * rc41899a377 10ffmpeg/ (4 files in 2 dirs): 
[12:37] <CIA-119> ffmpeg: lavu: split version info into version.h
[12:37] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[12:37] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * reeddd3962e 10ffmpeg/libavutil/cpu.h: 
[12:37] <CIA-119> ffmpeg: cpu.h: include version.h for LIBAVUTIL_VERSION_MAJOR
[12:37] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[14:48] <CIA-119> ffmpeg: 03Mashiat Sarker Shakkhar 07master * r98b3f17bd7 10ffmpeg/libavformat/img2dec.c: 
[14:48] <CIA-119> ffmpeg: image2: Add "start_number" private option to the demuxer
[14:48] <CIA-119> ffmpeg: Currently if a pattern is given we search for up to the fifth file name in
[14:48] <CIA-119> ffmpeg: that sequence. This option sets that limit to an arbitrary number.
[14:48] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[14:48] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r11f2cad127 10ffmpeg/doc/examples/filtering_audio.c: 
[14:48] <CIA-119> ffmpeg: examples/filtering_audio: update for avfilter API "cleanup"
[14:48] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[14:48] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * rfbf03ee391 10ffmpeg/ffmpeg.c: 
[14:48] <CIA-119> ffmpeg: ffmpeg: fix indent of 1 line
[14:48] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[14:48] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r3013722c13 10ffmpeg/libavfilter/sink_buffer.c: 
[14:48] <CIA-119> ffmpeg: sink_buffer: assert that functions and sinks arent mixed
[14:48] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[15:19] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r62975afa68 10ffmpeg/libavfilter/af_aresample.c: 
[15:19] <CIA-119> ffmpeg: af_aresample: fix sample_rate type.
[15:19] <CIA-119> ffmpeg: should fix fate failures on 32bit x86 linux
[15:19] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[15:31] <LunaVorax> Hello everyone!
[15:31] <LunaVorax> I'm trying to cross-compile ffmpeg for Linux on MIPS32
[15:32] <LunaVorax> I set the flags --enable-cross-compile --target-os=linux and --cpu=mips (among others) and it tells me back "Must specify target arch and OS when cross-compiling"
[15:33] <LunaVorax> I'm a bit confused, I've checked how configure was made and I don't see what I'm doing wrong, I've tried other name of the cpu witout success.
[15:33] <LunaVorax> By the way, I'm using the lastest version from Git
[15:58] <michaelni> LunaVorax, see "http://fate.ffmpeg.org/report.cgi?time=20120624061051&slot=mips-ubuntu-qemu-gcc-4.4" for working mips configure parameters
[16:00] <LunaVorax> Thank you michaelni, I'm going to read this
[16:02] <LunaVorax> Ok replacing --cpu by --arch fixed it (apparently). Let's see if it compiles all the way down :)
[17:11] <jesk> all movies i muxed from mkv to m4v/mp4 aren't seekable in iTunes or Airplay
[17:11] <jesk> movie starts to play just fine
[17:12] <jesk> but as soon as I skip a bit forward audio keeps starting again correctly but video never comes again
[17:12] <jesk> i also tried muxing via mp4box, but same problem
[17:13] <jesk> in VLC seeking just works blameless
[17:14] <jesk> what could i try to get that working in future?
[18:37] <Nickname123> Is it possible to access frame count in AVOutputFormat.write_packet?
[18:38] <Nickname123> I am trying to make the segmenter format observe a valid frame whitelist so that segments start with specific frames
[18:39] <Nickname123> But I am having trouble getting the frame numbers in the packet
[18:40] <Nickname123> I thought I was doing it correctly with AVStream->nb_frames but I am getting really large frame counts from it on a short 30 second test clip
[18:57] <michaelni> Nickname123, does the nb_frames you see mismatch the number of packets written ?
[19:03] <Nickname123> Maybe I am doing something wrong, but I was keeping a sum of st->nb_frames thinking it was the # of frames in the packet.  This sum is 1398795 at the end of a 30 second video with a frame rate of 30
[19:04] <Nickname123> st->nb_frames is between 800 and 1400 on every call to write_packet
[19:04] <Nickname123> I think the video has 901 or so frames according to the last ffmpeg cli output
[19:05] <Nickname123> for the record... this is the first time I am working on ffmpeg internals so I am sure I am just doing something wrong
[19:17] <Nickname123> How would I get the frame number of the first frame in the packet?
[19:31] <michaelni> each write_packet() call will "contain" 1 frame normally
[19:32] <Nickname123> Oh I thought I read a comment that it could contain more than one
[19:35] <Nickname123> Thanks.  That gives me the same frame count that the ffmpeg cli output gives
[19:41] <jesk> is my problem more likely to be a problem of the h264 data or a mp4 container problem?
[19:43] <Compn> jesk : did you try running qt-faststart on them files ?
[20:05] <uau> michaelni: compiling against ffmpeg libavutil is broken; your change creating separate version.h broke libavutil.pc generation as it tries to read version from old location
[20:06] <jesk> Compn: yes, both ways (so either with mdata at the beginning or at the end) results in same problem
[20:12] <michaelni> uau, thanks, ill fix it in a moment
[20:13] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * ree51565e8e 10ffmpeg/configure: 
[20:13] <CIA-119> ffmpeg: configure: fix libavutil.pc generation
[20:13] <CIA-119> ffmpeg: Found-by: uau
[20:13] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:31] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * rf0a3259f9a 10ffmpeg/libavcodec/rangecoder.h: 
[20:31] <CIA-119> ffmpeg: rangecoder.h: use av_assert
[20:31] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:31] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r5c44c2de80 10ffmpeg/libavcodec/mlp_parser.c: 
[20:31] <CIA-119> ffmpeg: mlp_parser: use av_assert
[20:31] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[21:20] <jesk> is it possible to write mdata at the end and atoms at the beginning with those -moov_* flags?
[21:20] <jesk> so without using qt-faststart
[21:20] <jesk> understood it that way that it reserve space for atoms when starting to write
[21:21] <jesk> maybe i'am wrong
[21:22] <jesk> when just using -moov_size it writes no mdat at all :O
[21:26] <Compn> jesk : you said you tried mp4box ? did you try libquicktime or whatever the name of the other mp4 project was
[21:27] <Compn> gpac
[21:27] <Compn> what was your source h264 again ? are you sure its not strange h264 without sps or some nonsense flag that allows seeking
[21:27] <Compn> e.g. do you need to specify an h264 bsf ?
[21:29] <jesk> isn't mp4box gpac?
[21:30] <Compn> is it ?
[21:30] <jesk> seems that the problem only arise when extracting h264 with mkvextract and then muxing into mp4 with ffmpeg/mp4box
[21:31] <jesk> Compn, ok i confused those two projects
[21:31] <jesk> but they are from the same guy it seems
[21:37] <jesk> GPAC is Osmo+MP4Box
[21:41] <CIA-119> ffmpeg: 03Mans Rullgard 07master * r0b6f973635 10ffmpeg/libavcodec/ (h264_cabac.c x86/h264_i386.h): 
[21:41] <CIA-119> ffmpeg: h264: use asm cabac reader under a generic condition
[21:41] <CIA-119> ffmpeg: This removes a dependency on implementation details from generic
[21:41] <CIA-119> ffmpeg: code and allows easy addition of the equivalent optimisation for
[21:41] <CIA-119> ffmpeg: other architectures than x86.
[21:41] <CIA-119> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[21:41] <CIA-119> ffmpeg: 03Anton Khirnov 07master * r7a8059eb9c 10ffmpeg/doc/filters.texi: doc/filters: fix typo.
[21:41] <CIA-119> ffmpeg: 03Anton Khirnov 07master * r6fc7d9a078 10ffmpeg/avconv.c: 
[21:41] <CIA-119> ffmpeg: avconv: remove redundant handling of async.
[21:41] <CIA-119> ffmpeg: Because of a mistake during merging the code for simple and complex
[21:41] <CIA-119> ffmpeg: filtergraphs, -async inserts an asyncts filter both on input and output.
[21:41] <CIA-119> ffmpeg: Remove the output hunk.
[21:41] <CIA-119> ffmpeg: 03Anton Khirnov 07master * rdf98bf22cb 10ffmpeg/avconv.c: 
[21:41] <CIA-119> ffmpeg: avconv: use more descriptive names for hardcoded filters.
[21:41] <CIA-119> ffmpeg: Also ensure that all such filters get a non-NULL name. Should fix FATE
[21:41] <CIA-119> ffmpeg: failures on some architectures after
[21:41] <CIA-119> ffmpeg: 58b049f2fa4f192b00baadb5f1f32ca366f936ea.
[23:59] <CIA-119> ffmpeg: 03Paul B Mahol 07master * r7ebe28c23c 10ffmpeg/libavcodec/xwdenc.c: 
[23:59] <CIA-119> ffmpeg: xwdenc: support 8bpp grayscale
[23:59] <CIA-119> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[23:59] <CIA-119> ffmpeg: 03Piotr Bandurski 07master * rb9c94e826e 10ffmpeg/libavcodec/xwddec.c: xwddec: support 8bpp grayscale
[00:00] --- Mon Jun 25 2012