[Ffmpeg-devel-irc] ffmpeg-devel.log.20120116

[burek]

[00:24] <ubitux> saste: you have a small typo in the ifzero patch: "togheter"
[00:25] <saste> ubitux: that's one of my favourite typos
[00:26] <ubitux> if you love them, then i guess you have to keep it
[00:26] <ubitux> but just add a comment so it isn't fixed later
[00:26] <saste> i should change the dictionaries instead
[00:27] <ubitux> i'm sure the higher institutions will agree with this
[00:27] <ubitux> ffdictionary
[00:28] <ubitux> let's also rewrite dictionnaries in a better way
[00:28] <ubitux> heh, we already have AVDictionary.
[00:32] Action: iive handles the price "Best Joke Of The Year" to ubitux 
[00:33] <ubitux> sorry :(
[00:33] <ubitux> i won't do it again.
[00:34] <iive> why, it was really nice joke.
[00:35] <ubitux> i'm not the original author of the AVDictionary joke
[00:35] <ubitux> :-°
[00:40] <ubitux> btw, am i the only one to be ignored by most of the libav people?
[00:42] <michaelni> ubitux, no
[00:42] <ubitux> ok
[00:42] <ubitux> i guess it's because i'm definitely part of the dark side
[00:53] <beastd> It is not dark. It is just FFmpeg, fast and forward you know :D
[00:54] <ubitux> i like when it's dark
[00:54] <ubitux> it feels safer.
[00:55] <ubitux> saste: i like michael' suggestion about the if() instead of ifzero
[00:58] <saste> ubitux: +1
[01:06] <Compn> ubitux : i am also ignored by most of libav 
[01:07] <ubitux> it feels like they want us to despise them so they could have a reasons to be
[01:08] <ubitux> it sucks :(
[01:08] <Compn> i would recommend dropping those patches
[01:08] <ubitux> yes, i'll ping them another time and then drop them
[01:44] <CIA-31> ffmpeg: 03Laurentiu Ion 07master * r4fccc7c46c 10ffmpeg/libavcodec/vb.c: 
[01:44] <CIA-31> ffmpeg: vb: Use bytestream2 functions
[01:44] <CIA-31> ffmpeg: Signed-off-by: Justin Ruggles <justin.ruggles at gmail.com>
[01:44] <CIA-31> ffmpeg: 03Kostya Shishkov 07master * re6e7bfc11e 10ffmpeg/libavcodec/wavpack.c: 
[01:44] <CIA-31> ffmpeg: wavpack: fix clipping for 32-bit lossy mode
[01:44] <CIA-31> ffmpeg: Reference decoder clips data before shifting it to final range and also
[01:44] <CIA-31> ffmpeg: forces 32-bit lossy mode to be actually 24-bit lossy mode in order to be
[01:44] <CIA-31> ffmpeg: able to perform proper clipping.
[01:44] <CIA-31> ffmpeg: 03Daniel Huang 07master * rbcd5d979aa 10ffmpeg/libavformat/ (mov.c mxfdec.c): 
[01:44] <CIA-31> ffmpeg: mov, mxfdec: Employ more meaningful return values.
[01:44] <CIA-31> ffmpeg: Signed-off-by: Diego Biurrun <diego at biurrun.de>
[01:44] <CIA-31> ffmpeg: 03Diego Biurrun 07master * r2df5f59ad0 10ffmpeg/tests/ (Makefile fate/adpcm.mak fate/dpcm.mak): fate: split ADPCM and DPCM test references into separate files.
[01:44] <CIA-31> ffmpeg: 03Michael Niedermayer 07master * r4a4c4278b7 10ffmpeg/: (log message trimmed)
[01:44] <CIA-31> ffmpeg: Merge remote-tracking branch 'qatar/master'
[01:44] <CIA-31> ffmpeg: * qatar/master:
[01:44] <CIA-31> ffmpeg:  fate: split ADPCM and DPCM test references into separate files.
[01:44] <CIA-31> ffmpeg:  mov, mxfdec: Employ more meaningful return values.
[01:44] <CIA-31> ffmpeg:  lavc: Relax API strictness in avcodec_decode_audio3 with a custom get_buffer()
[01:44] <CIA-31> ffmpeg:  wavpack: fix clipping for 32-bit lossy mode
[01:44] <CIA-31> ffmpeg: 03Reinhard Tartler 07master * re2ff436ef6 10ffmpeg/libavcodec/ (avcodec.h utils.c): (log message trimmed)
[01:44] <CIA-31> ffmpeg: lavc: Relax API strictness in avcodec_decode_audio3 with a custom get_buffer()
[01:44] <CIA-31> ffmpeg: Do not fail audio decoding with avcodec_decode_audio3 if user has set a
[01:44] <CIA-31> ffmpeg: custom get_buffer. Strictly speaking, this was never allowed by the API,
[01:44] <CIA-31> ffmpeg: but it seems that some software packages did so anyways. In order to
[01:44] <CIA-31> ffmpeg: unbreak applications (cf. http://bugs.debian.org/655890), this change
[01:44] <CIA-31> ffmpeg: clarifies the API and overrides the custom get_buffer() with the defaults.
[03:40] <CIA-31> ffmpeg: 03Jean First 07master * r0c10885c39 10ffmpeg/libavformat/movenc.c: 
[03:40] <CIA-31> ffmpeg: movenc: indent cosemtics
[03:40] <CIA-31> ffmpeg: Signed-off-by: Jean First <jeanfirst at gmail.com>
[03:40] <CIA-31> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[03:40] <CIA-31> ffmpeg: 03Jean First 07master * r60168f9319 10ffmpeg/libavformat/movenc.c: 
[03:40] <CIA-31> ffmpeg: movenc: add FIXME to add alis and rsrc atoms
[03:40] <CIA-31> ffmpeg: Signed-off-by: Jean First <jeanfirst at gmail.com>
[03:40] <CIA-31> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[03:40] <CIA-31> ffmpeg: 03Jean First 07master * rfa169d5650 10ffmpeg/libavformat/mov.c: 
[03:40] <CIA-31> ffmpeg: mov: cosmetics - move a line to a better position and add a comment
[03:40] Last message repeated 1 time(s).
[03:40] <CIA-31> ffmpeg: Oana Andreea Stratulat submitted a similar patch to trac, but forgot
[03:40] <CIA-31> ffmpeg: to notify the ML about it.
[03:40] <CIA-31> ffmpeg: Signed-off-by: Jean First <jeanfirst at gmail.com>
[03:41] <CIA-31> ffmpeg: This is required by the spec and fixes video-1frag.ogg.48.ogg. (FPE)
[03:41] <CIA-31> ffmpeg: Based on the debuging work of Oana Stratulat and ubitux.
[03:41] <CIA-31> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[03:53] <CIA-31> ffmpeg: 03Panagiotis H.M. Issaris 07master * r811de351ea 10ffmpeg/libavformat/applehttp.c: 
[03:53] <CIA-31> ffmpeg: Enhance HLS seeking.
[03:53] <CIA-31> ffmpeg: Enhances seeking by demuxing until the requested timestamp is reached within
[03:53] <CIA-31> ffmpeg: the segment selected by the seek code using the playlist info.
[03:53] <CIA-31> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[04:13] <troy_s> Quick question... according to the transfer enum in avcodec.h, which would be the proper bt601.7 implementation?
[04:16] <troy_s> Am I to assume it is 2.2 bt470m?
[10:33] <ubitux> cbsrobot: i will send the first draft of the new public timecode api today, or tomorrow
[10:34] <ubitux> so you can move on with your mov work
[10:34] <ubitux> feel free to comment it on the mailing list (or here) when i post it btw
[10:59] <CIA-31> ffmpeg: 03Carl Eugen Hoyos 07master * r203ea375ca 10ffmpeg/libavcodec/dpxenc.c: Support rgb64 dpx encoding.
[11:00] <CIA-31> ffmpeg: 03Carl Eugen Hoyos 07master * ra8b117f800 10ffmpeg/tests/ (lavf-regression.sh ref/lavf/png): Add gray16 and rgb48 png encoding regression tests.
[11:10] <cbsrobot> ubitux: great
[11:10] <cbsrobot> btw here's the WIP http://pastebin.com/Pm6JaG1Q
[11:11] <cbsrobot> it's not worth sending to the ML because it's rather basic
[11:11] <ubitux> the new API should be way more consistent and flexible
[11:11] <cbsrobot> ok
[11:11] <ubitux> and i'm introducing the flags indeed
[11:12] <cbsrobot> i'm not sure if we should add metadata to ff_timecode ....
[11:14] <ubitux> good point
[12:01] <kriegerod> hi, devs - in libav-user at ffmpeg.org there's interesting question about usage of context pointer argument for logging callback function. Title is "[Libav-user] Questions about using libav in a multi-threaded application". Could anybody shed some light?
[12:02] <kriegerod> I mean answering in that thread
[12:57] <CIA-31> ffmpeg: 03Carl Eugen Hoyos 07master * rd93b6340d7 10ffmpeg/tests/ (lavf-regression.sh ref/lavf/dpx): Add dpx encoding regression test.
[16:38] <durandal_1707> how to handle formats that put data in single one big chunk after headers?
[16:38] <av500> durandal_1707: run from them as fast as possible
[16:39] <durandal_1707> our api cant handle them?
[16:39] <Tjoppen> what format?
[16:39] <Tjoppen> mov would fit that vague description
[16:39] <durandal_1707> brstm
[16:40] <durandal_1707> http://wiibrew.org/wiki/BRSTM_file
[16:41] <av500> ah, so it's like WAV
[16:41] <Tjoppen> write a demuxer for it? should be less than 200 lines
[16:41] <Tjoppen> yeah, could probably share code with wav.c
[16:42] <av500> refactor into header_and_large_chunk.c
[16:42] <Tjoppen> wav.c ought to be split into wavdec.c and wavenc.c btw
[16:42] <merbanan> durandal_1707: just return a suitable amount of data
[16:44] <merbanan> durandal_1707: do you have complete specs for the adpcm format ?
[16:45] <durandal_1707> you have implementation in vgmstream
[16:45] <merbanan> ok
[16:50] <merbanan> durandal_1707: yupp, it's all there, write demuxer and pass the adpcm coeffs as extradata to the decoder
[17:48] <CIA-31> ffmpeg: 03Clément BSsch 07master * r5fc51c3155 10ffmpeg/libavutil/audioconvert.c: lavu: add 2.1 channel layout.
[19:35] <CIA-31> ffmpeg: 03Reimar Döffinger 07master * r6003ec880b 10ffmpeg/libavformat/asfdec.c: 
[19:35] <CIA-31> ffmpeg: asfdec: Move variable declarations into the blocks they are used in.
[19:35] <CIA-31> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[19:35] <CIA-31> ffmpeg: 03Reimar Döffinger 07master * r203f94fd7c 10ffmpeg/libavformat/mxfdec.c: 
[19:35] <CIA-31> ffmpeg: mxfdec: Add missing free.
[19:35] <CIA-31> ffmpeg: Should fix valgrind FATE.
[19:35] <CIA-31> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[19:35] <CIA-31> ffmpeg: 03Reimar Döffinger 07master * r7594791541 10ffmpeg/libavcodec/vp8.c: (log message trimmed)
[19:35] <CIA-31> ffmpeg: Rework VP8 decoder frame skipping code.
[19:35] <CIA-31> ffmpeg: This reworks the frame skipping code such that the reference
[19:35] <CIA-31> ffmpeg: buffers are still updated according to the header.
[19:35] <CIA-31> ffmpeg: However it also ensures that the current frame will not end
[19:35] <CIA-31> ffmpeg: up in any reference buffer.
[19:35] <CIA-31> ffmpeg: Also fixes a hang with frame-multithreading, probably because
[19:45] <CIA-31> ffmpeg: 03Reimar Döffinger 07master * rc203044f37 10ffmpeg/tests/ref/lavf/dpx: 
[19:45] <CIA-31> ffmpeg: Remove extraneous newline from dpx reference which breaks FATE.
[19:45] <CIA-31> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[20:43] <CIA-31> ffmpeg: 03Maksym Veremeyenko 07master * r94bf9ac473 10ffmpeg/libavcodec/parser.c: 
[20:43] <CIA-31> ffmpeg: fix av_dlog call with non-AVClass struct
[20:43] <CIA-31> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[00:00] --- Tue Jan 17 2012