[Ffmpeg-devel-irc] ffmpeg-devel.log.20111210

[burek]

[00:03] <Tjoppen> I'll take a look in the morning
[01:08] <Tjoppen> took a quick peek. patches look decent I guess. haven't look at the offendine sample though
[01:09] <Tjoppen> *looked *offending
[01:22] <Compn> anyone remember the email when ruby said he would work on j2k even tho he missed GSoC ?
[01:22] <Compn> digging in the archives i cant find it
[01:22] <Compn> maybe he said in on irc
[01:32] <pasteeater> Compn: when was this?
[01:57] <Compn> pasteeater : 2011 soc , so march/april/may
[01:58] <Compn> i see a bunch of posts where he asks about gsoc, and his mails have that in the subject
[01:58] <Compn> then ... just him working on soc
[01:58] <Compn> brb, must restart compootor
[02:08] <CIA-41> ffmpeg: 03Luca Barbato 07master * rb73a01eaf1 10ffmpeg/Doxyfile: doxy: change hue value to match our green
[02:08] <CIA-41> ffmpeg: 03Luca Barbato 07master * r1ebbdda1d2 10ffmpeg/doc/ (Makefile fate.texi fate.txt): 
[02:08] <CIA-41> ffmpeg: doc: document fate in a texinfo
[02:08] <CIA-41> ffmpeg: Summarize the information provided in the wiki and the one provided
[02:08] <CIA-41> ffmpeg: in fate.txt
[02:08] <CIA-41> ffmpeg: 03Luca Barbato 07master * rc1f5da698b 10ffmpeg/doc/t2h.init: 
[02:08] <CIA-41> ffmpeg: doc: provide fallback css and customize @float
[02:08] <CIA-41> ffmpeg: Add "note" or "important" class to the generated div.
[02:08] <CIA-41> ffmpeg: 03Luca Barbato 07master * r2f017d9791 10ffmpeg/doc/ (Makefile git-howto.texi): doc: port the git-howto to texinfo
[02:08] <CIA-41> ffmpeg: 03Luca Barbato 07master * r1de6e14efe 10ffmpeg/doc/ (Makefile general.texi platform.texi): 
[02:08] <CIA-41> ffmpeg: doc: split platform specific information
[02:08] <CIA-41> ffmpeg: Move the platform-specific information in its own file.
[02:08] <CIA-41> ffmpeg: 03Luca Barbato 07master * r9bdfe9ac1c 10ffmpeg/doc/general.texi: doc: extend external library coverage
[02:08] <CIA-41> ffmpeg: 03Luca Barbato 07master * rca410b4eb0 10ffmpeg/doc/demuxers.texi: 
[02:08] <CIA-41> ffmpeg: doc: update demuxers section
[02:09] <CIA-41> ffmpeg: Cleanup tags usage and update commandline examples.
[02:09] <CIA-41> ffmpeg: 03Luca Barbato 07master * rd5837d7fe9 10ffmpeg/doc/ (6 files): doc: update documentation to use avconv
[02:09] <CIA-41> ffmpeg: 03Justin Ruggles 07master * r8e8c51318c 10ffmpeg/ (6 files in 3 dirs): 
[02:09] <CIA-41> ffmpeg: movenc: simplify handling of pcm vs. adpcm vs. other compressed codecs
[02:09] <CIA-41> ffmpeg: Use Sound Sample Description Version 2 for all MOV files.
[02:09] <CIA-41> ffmpeg: Updated FATE references accordingly.
[02:09] <CIA-41> ffmpeg: Note that ADPCM is treated as compressed audio in version 2.
[02:09] <CIA-41> ffmpeg: 03Justin Ruggles 07master * rb2890f5ed6 10ffmpeg/libavutil/ (audioconvert.h avutil.h): audioconvert: add some additional channel and channel layout macros
[02:09] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rc59b80c8d3 10ffmpeg/: (log message trimmed)
[02:09] <CIA-41> ffmpeg: Merge remote-tracking branch 'qatar/master'
[02:09] <CIA-41> ffmpeg: * qatar/master:
[02:09] <CIA-41> ffmpeg:  isom: sort and pretty-print codec_movaudio_tags[]
[02:09] <CIA-41> ffmpeg:  isom: remove pointless comments in codec_movaudio_tags[]
[02:09] <CIA-41> ffmpeg:  isom: remove commented-out tag for vorbis
[02:09] <CIA-41> ffmpeg:  movenc: write 'chan' tag for AC-3 in MOV
[02:09] <CIA-41> ffmpeg: 03Justin Ruggles 07master * r60cd802daf 10ffmpeg/libavformat/movenc.c: movenc: write 'chan' tag for AC-3 in MOV
[02:09] <CIA-41> ffmpeg: 03Justin Ruggles 07master * r21ebf2d493 10ffmpeg/libavformat/isom.c: isom: remove commented-out tag for vorbis
[02:25] <Compn> pasteeater : i looked all over ffmpeg-devel , libav-devel and such to no luck
[02:25] <Compn> must have been on irc or some soc list i cant find
[02:30] <michaelni> btw, can someone add links to the splited out htmls like http://ffmpeg.org/platform.html ?
[02:31] <michaelni> http://ffmpeg.org/fate.html
[02:31] <michaelni> http://ffmpeg.org/git-howto.html
[02:31] <michaelni> possibly more are missing links
[02:32] <pasteeater> Compn: i found several email threads [JPEG2000 decoder, GSoC project (JPEG 2000), JPEG 2000, Regarding GSoC], but i didn't see a specific mention of missing gsoc in a quick search
[03:05] <Compn> michaelni : ping me later about it
[03:05] <Compn> also we need to link to the wiki, remember that thread died ?
[03:08] Action: michaelni will forget
[03:59] <ubitux> what is the unit name for the step between 2 close component?
[03:59] <ubitux> for instance, the red component goes from 0xAA to 0xAB, it is increased by one what?
[04:01] <michaelni> redness unit ;)
[04:01] <ubitux> okay =)
[04:07] <durandal_1707> does ljpeg encoder actually works?
[04:17] <michaelni> durandal_1707, tests/data/vsynth1/ljpeg.avi looks ok so it works for that case at least
[04:17] <michaelni> what problem do you see with it ?
[04:17] <durandal_1707> well i just use avconv to create avi with ljpeg stream
[04:18] <michaelni> well ...
[04:18] <durandal_1707> that is probably dumbest idea ever
[04:18] <michaelni> you should be using ffmpeg
[04:23] <durandal_1707> amv encoder produce strange stuff on top pixels
[04:25] <michaelni> i dont see anything like this in tests/data/vsynth1/amv.avi (which is generated by ffmpeg)
[04:25] <michaelni> do you have a reproduceable testcase ?
[04:27] <durandal_1707> yes
[04:30] <michaelni> can you show us that testcase ?
[04:37] <durandal_1707> hmm, it seems to happen with h264 (High) as input
[04:41] <michaelni> durandal_1707, are you using latest ffmpeg git master ?
[04:42] <michaelni> if yes, i probably need a short piece of the source file with which it can be reproduced
[04:42] <durandal_1707> it does not happens with mpeg4
[04:43] <durandal_1707> i'm not sure if this is heightxweight issue actually
[04:44] <durandal_1707> 720x404
[04:45] <michaelni> we need a testcase not random values
[04:45] <durandal_1707> you want me to upload input video?
[04:46] <michaelni> yes, preferable a small part of it (the smaller the better)
[04:46] <michaelni> small in terms of length
[04:46] <michaelni> that is if 1second is enough then just upload that
[04:50] <ubitux> Compn: wasn't ruby a girl?
[04:50] Action: Compn eats a pomelo
[04:50] <Compn> possibly
[04:50] <Compn> but i'd rather call a girl as a guy than call a guy as a girl ;)
[04:51] <ubitux> that's because you're not a girl
[04:51] <Compn> its because i'd rather insult a girl than to insult a guy who may use fists in a fit of rage :P
[04:51] <ubitux> :D
[04:52] <ubitux> you should defend yourself by saying calling him a man is actually a compliment
[04:52] <ubitux> calling her*
[04:52] <ubitux> :p
[04:53] <Compn> did someone reply to my mail ?
[04:53] <ubitux> not yet afaik
[04:53] <ubitux> anyway it's late, 'night ppl :)
[04:54] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r4ac862f43c 10ffmpeg/Makefile: 
[04:54] <CIA-41> ffmpeg: Makefile: disable compiling avconv by default.
[04:54] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[04:54] <CIA-41> ffmpeg: 03Hendrik Leppkes 07master * rb7bf93b923 10ffmpeg/libavcodec/mlp_parser.c: 
[04:54] <CIA-41> ffmpeg: mlp: Use correct speaker locations in TrueHD channel layout.
[04:54] <CIA-41> ffmpeg: (cherry picked from commit d7787835ad737d4f4d8b6f17e04d981632eeaa91)
[04:54] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[05:01] <burek> is there a way to tell ffmpeg to ignore a duration, that is found in WAV header?
[05:01] <burek> because some ip cameras seem to send stream of audio
[05:02] <burek> in wav format, with duration set to 32 seconds
[05:03] <CIA-41> ffmpeg: 03Paul B Mahol 07master * r85cd1eb12f 10ffmpeg/libavcodec/ (dcaenc.c mjpegenc.c): 
[05:03] <CIA-41> ffmpeg: add missing long_name for amv and dca encoder
[05:03] <CIA-41> ffmpeg: Reviewed-by: Carl Eugen Hoyos
[05:03] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[05:03] <Compn> you mean the audio stream stops but video continues ?
[05:03] <Compn> burek : ffmpeg -an -i ... ?
[05:04] <burek> why -an
[05:04] <burek> no no
[05:04] <burek> only audio
[05:04] <burek> http://pastie.org/2994411
[05:04] <burek> you see "Duration: 00:00:32.76, bitrate: N/A"
[05:05] <burek> but it's not 32 seconds, it's an endless stream
[05:05] <Compn> does it stop after 32 seconds ?
[05:05] <burek> yes
[05:05] <Compn> just making sure :D
[05:05] <burek> :))
[05:05] <michaelni> burek, this is not latest ffmpeg
[05:05] <Compn> do you want the audio ?
[05:06] <burek> michaelni, how can we tell the latest ffmpeg to ignore the duration?
[05:06] <burek> Compn, yes
[05:06] <michaelni> well, if its reproduceable with latest ffmpeg submit a bugreport and ill look into it
[05:07] <durandal_1707> i uploaded amv_in.mp4 to incoming on upload.ffmpeg.org
[05:07] <burek> ok
[05:07] <michaelni> burek, or just give me the file and skip the report
[05:07] <burek> ok just a second
[05:07] <burek> hackeron, if your webcam is accessible via internet
[05:07] <burek> could you leave your url here
[05:08] <hackeron> burek: IP camera, it isn't, but give me a minute :)
[05:08] <burek> ok
[05:10] <hackeron> help yourself: curl http://admin:@82.13.197.168:82/audio.cgi
[05:11] <burek> michaelni, will the url be enough?
[05:11] <hackeron> working command: curl http://admin:@82.13.197.168:82/audio.cgi | sox --ignore-length - -o test.wav   -- I'm looking for something similar in ffmpeg to be able to ignore the header length/duration
[05:12] <burek> ok, hackeron, please wait a little bit now :)
[05:12] <hackeron> :)
[05:16] <durandal_1707> michaelni: can you reproduce amv encoder bug?
[05:17] <michaelni> durandal_1707, yes
[05:21] <burek> michaelni, I've tried the ip cam url above, and loop_input doesn't help.. should it be reported as a bug or as a feature or how?
[05:24] <burek> also, when I use "-loop 1" it says "Option loop not found." :) and when I use "-loop_input" it says "-loop_input is deprecated, use -loop 1" :) so I'll report this too :)
[05:40] <hackeron> burek: any luck?
[05:44] <michaelni> hackeron, ive added a ignore-length option, ill push it in a few moments
[05:46] <hackeron> michaelni: fantastic, thank you :)
[05:46] <hackeron> that was quick, heh
[05:48] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r02acfb1c01 10ffmpeg/libavformat/wav.c: 
[05:48] <CIA-41> ffmpeg: wavdec: add ignore length option to ignore incorrect lengthes
[05:48] <CIA-41> ffmpeg: like SOX
[05:48] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[05:48] <hackeron> thank you :)
[05:50] <hackeron> one more thing since I'm here - in gstreamer, I can do src do-timestamp=true ! decodebin ! encoder - this is very useful with IP cameras that have an mjpeg stream (no timestampts) - I couldn't find anything equivalent in ffmpeg - am I missing something?
[05:54] <michaelni> hackeron, i dont really know gstreamer, what does that do ?
[05:56] <hackeron> michaelni: the problem is it's a live source - http multipart. When I open it with ffmpeg, it sets an fps and generates timestamps based on the fps - but the source is variable framerate with no timestamps (mjpeg) - anyway to make ffmpeg treat it like a live source and set timestamps of "now" for every incoming frame? - I tried -timestamp now but it had no effect
[05:58] <hackeron> do-timestamp=true description on the gstreamer website is: Timestamps are set on the buffers as they arrive from the camera. These are used by the mime/multipart demultiplexer to emit timestamps on the JPEG-encoded video frame buffers.
[05:58] <hackeron> This allows the multiplexer to timestamp the frames in the resulting file.
[05:59] <michaelni> not sure if we support  that but it should be quite easy to add
[06:00] <michaelni> probably best if you add feature request to our bug tracker otherwise i might forget
[06:00] <hackeron> would be very useful, pretty much every mjpeg IP camera I tried has this problem with ffmpeg - I will file a feature request, thank you :)
[06:01] <burek> is it possible to get a reencoded file after -vcodec copy -acodec copy ?
[06:02] <hackeron> ffmpeg bugs/requests go here, right? < http://ffmpeg.org/trac/ffmpeg/report
[06:02] <burek> yes
[06:03] <burek> anyway, a guy used -acodec copy -vcodec copy and got a video output that is not the same as the original (he says it's fuzzy): http://ffmpeg.org/trac/ffmpeg/ticket/728
[06:03] <burek> so I was wondering if ffmpeg did reencode the video or what could have happened
[06:05] <michaelni> burek, SVN-r0.5.1 is too old
[06:05] <michaelni> thats years old
[06:05] <burek> oh sorry wrong link
[06:05] <burek> http://pastebin.com/F08P2ugT
[06:06] <burek> I made the guy compile the latest git ^^
[06:08] <michaelni> better update the ticket before someone closes it in rage, screaming that its too old ;)
[06:08] <durandal_1707> well streams can be same but muxing does not need to be
[06:09] <burek> well I understand that part, but can the muxing cause the video to become "fuzzy" ? :)
[06:10] <durandal_1707> what is "fuzzy"?
[06:11] <durandal_1707> well with copy you can trivialy make file which will not play with ffplay (because you can copy unsupported codecs)
[06:11] <hackeron> hope this is descriptive enough: https://ffmpeg.org/trac/ffmpeg/ticket/730
[06:12] <burek> I asked him to provide a sample of a "fuzzy" file, so I will know what to answer :)
[06:14] <durandal_1707> here is stupid example: ffmpeg -i escape130.rpl -vcodec copy out.avi
[06:15] <durandal_1707> in theory this could be supported because escape130 codec have header which define width and height and it could be played from avi ;)
[06:51] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r02ce07f8c4 10ffmpeg/libavcodec/mpegvideo_enc.c: 
[06:51] <CIA-41> ffmpeg: amvenc: support videos with heights that are not a multiple of 16.
[06:51] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[07:13] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rcd519b7e83 10ffmpeg/libavformat/riff.c: 
[07:13] <CIA-41> ffmpeg: avi: add escape130 support
[07:13] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[07:13] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rf5ecc4c616 10ffmpeg/libavcodec/avcodec.h: 
[07:13] <CIA-41> ffmpeg: lavc: 1000l, unbreak ABI
[07:13] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[10:03] <jeremyS> What's the deal with libavfilter planar audio?
[10:03] <jeremyS> How do the channels map to the data[i] buffers?
[11:07] <Tjoppen> so, do any of the mxf patches on the list fix #721?
[11:07] <Tjoppen> looking quickly it seems it uses the same essence_container_ul for both streams?
[13:03] <mateo`> hello, i am currently writing a filter to merge two audio tracks. I could not find a way to detect inputs sample rate or force a given sample on inputs. Any ideas ?
[14:23] <CIA-41> ffmpeg: 03Reimar Döffinger 07master * r2483f53693 10ffmpeg/libavformat/mxfdec.c: 
[14:23] <CIA-41> ffmpeg: mxfdec: do not try to use an obviously broken index.
[14:23] <CIA-41> ffmpeg: Also fixes a memleak.
[14:23] <CIA-41> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[14:43] <michaelni> mateo`, AVFilterLink contains a sample_rate field, which should tell you the input sample rate
[14:45] <GordonFreeman> hi guys
[14:46] <GordonFreeman> what is your opinion about Fractal Image Compression ?
[14:55] <mateo`> michaelni: AVFilterLink sample_rate is not set at init time
[14:56] <mateo`> michaelni: in init function i tried to use ctx->inputs[0]->sample_rate but it set to 0 :(
[14:57] <mateo`> i didn't find a way to autoinsert a aconvert filter by requesting a specific sample rate in query_formats
[14:59] <michaelni> mateo`, why do you need the sample rate at init time ?
[15:00] <michaelni> GordonFreeman, my opinion is that you should implement it and send a patch
[15:00] <GordonFreeman> :Ö
[15:00] <GordonFreeman> :)
[15:00] <michaelni> mateo`, sample rate can change anyway later :)
[15:01] <GordonFreeman> well i have tested a fractal image compression program 10 years ago
[15:01] <GordonFreeman> the compression is slow, but image quality and compression ratio is superior
[15:01] <GordonFreeman> http://en.wikipedia.org/wiki/Fractal_compression
[15:01] <mateo`> michaelni: my goal was to avoid doing the resampling job myself
[15:02] <mateo`> and let an autoinserted aconvert doing it for me
[15:05] <michaelni> i still dont understand why you need any resample at all, why does the sample rate matter to your filter ?
[15:06] <mateo`> to merge two audio tracks at two different sample rates.
[15:06] <CIA-41> ffmpeg: 03Carl Eugen Hoyos 07master * r5207f9597e 10ffmpeg/libavformat/mxfdec.c: 
[15:06] <CIA-41> ffmpeg: Separate mxf_essence_container_uls for audio and video.
[15:06] <CIA-41> ffmpeg: It is a really bad idea to assign a video codec id
[15:06] <CIA-41> ffmpeg: when we have set codec_type to audio and vice versa.
[15:06] <CIA-41> ffmpeg: Fixes ticket #721.
[15:06] <GordonFreeman> if i'd implement a fractal image compression algorithm then i'd say it's free until you use it in a software you sell, then you have to give me share, is it correct?
[15:23] <GordonFreeman> i'm doing bilinear interpolation for fun with matrices now :)
[15:48] <michaelni> GordonFreeman, if you implement something, then you can do so under any license you choose
[15:49] <GordonFreeman> and how to enforce my rights for some code?
[15:49] <michaelni> lawyer ;)
[15:50] <GordonFreeman> hehe, but it has to be registered somewhere to be correct i think
[15:50] <GordonFreeman> like with patents
[15:50] <GordonFreeman> what if the judge does not believe our database is real?
[15:50] <GordonFreeman> ;/
[15:51] <michaelni> you of course must pay the judge more than your opponent pays him ;)
[15:51] <GordonFreeman> and like the source of my code has date defined in my filesystem
[15:51] <GordonFreeman> ahaha
[15:52] <GordonFreeman> hmm though i could probably prove my rights if i show my whole development tree to the judge no?
[15:53] <GordonFreeman> like 10000 versions backed up
[15:55] <GordonFreeman> it'd be hard to believe someone created the exact same code without development
[15:56] <michaelni> just put the code on a public github repo then you have many witnesses that can say it was there and was put there by you
[15:58] <michaelni> mateo`, such filter probably is easiest to implement with some extensions to the framework
[15:58] <michaelni> so that its possible for a filter to request equal sample rates on input
[15:59] <michaelni> i mean such merge filter could probably even handle changing sample rates just not mismatching between inputs
[16:07] <GordonFreeman> michaelni<< well the witness thing is ok, though then anybody can copy all sources that i was wiling to show the judge to be my development history :)
[16:08] <michaelni> GordonFreeman, you should not write code but become lawyer
[16:09] <GordonFreeman> hehe :)
[16:09] <GordonFreeman> you'r right, its lawyer's duty...
[16:10] <GordonFreeman> i'm just bothered by ms like companies stealing free code and making money out of it
[16:11] <GordonFreeman> they'd just copy paste my work and don't give a fuck referring to my
[17:13] <CIA-41> ffmpeg: 03Nicolas George 07master * r337ce558b6 10ffmpeg/ (5 files in 2 dirs): fftools: add -report option.
[18:49] <ubitux> mmpf we really need a ffserver maintainer
[18:50] <ubitux> libav ppl are just removing it step by step instead of fixing the regressions :(
[18:54] <michaelni> ubitux, libav plans to drop ffserver ?
[18:54] <ubitux> well, from what i see they are trying to drop some broken features
[18:55] <michaelni> :)
[18:55] <ubitux> http://lists.libav.org/pipermail/libav-devel/2011-December/016292.html
[18:57] <michaelni> and yes we need a ffserver maintainer
[18:58] <michaelni> also the code elenril drops has been partially fixed by me already :)
[18:59] Action: michaelni kicks CIA-41 
[18:59] <CIA-41> ow
[19:00] <michaelni> CIA-41, where are the commits i just pushed ?
[19:50] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r257c85cddd 10ffmpeg/libavcodec/adpcm.c: 
[19:50] <CIA-41> ffmpeg: adpcm: fix division by zero in fate/creative/intro-partial.wav with -s 2 and -r 0.001:1
[19:50] <CIA-41> ffmpeg: Bug-Found-by: Shitiz Garg
[19:50] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:50] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r5b2873c1d5 10ffmpeg/libavcodec/ (h264.c h264_refs.c): 
[19:50] <CIA-41> ffmpeg: h264: switch from qatars has_b_frames calculation back to mine.
[19:50] <CIA-41> ffmpeg: Their 2nd try does like the first not work at all.
[19:50] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:50] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rcc61643123 10ffmpeg/libavcodec/h264.c: 
[19:50] <CIA-41> ffmpeg: h264: improve "no picture" debug output.
[19:50] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:51] <CIA-41> ffmpeg: 03Reimar Döffinger 07master * r0849a0ebb2 10ffmpeg/libavformat/rtmpproto.c: 
[19:51] <CIA-41> ffmpeg: rtmpproto: make condition work in overflow case.
[19:51] <CIA-41> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[19:51] <CIA-41> ffmpeg: 03Reimar Döffinger 07master * r8b8a47f6fb 10ffmpeg/libavformat/flvdec.c: 
[19:51] <CIA-41> ffmpeg: Make detection of single-stream flv work again.
[19:51] <CIA-41> ffmpeg: It had become dead code when code was added to avoid
[19:51] <CIA-41> ffmpeg: exporting audio and video codec id as metadata.
[19:51] <CIA-41> ffmpeg: Untested due to lack of sample.
[19:51] <CIA-41> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[19:51] <CIA-41> ffmpeg: 03Reimar Döffinger 07master * r78fde9323d 10ffmpeg/tests/ (4 files in 3 dirs): 
[19:51] <CIA-41> ffmpeg: Run zmbv regression without md5sum.
[19:51] <CIA-41> ffmpeg: Judging by the differences this should make all FATE
[19:51] <CIA-41> ffmpeg: machines pass the test.
[19:51] <CIA-41> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[21:04] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r934fe00680 10ffmpeg/ffmpeg.c: 
[21:04] <CIA-41> ffmpeg: ffmpeg: better warning about loop_input
[21:04] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[21:05] <michaelni> burek, see above and if you have a better idea for a warning, please suggest it, iam happy to change it
[21:52] <J_Darnley> Is there a check in configure for SSE4.1 which sets a define like HAVE_SSSE3?
[21:52] <J_Darnley> I guess there isn't since there ist a HAVE_SSE4 define
[21:52] <J_Darnley> Can I use HAVE_AVX instead?
[21:58] <burek> michaelni, it's better now :) also, maybe the documentation should be updated too?
[21:59] <michaelni> J_Darnley, yes, use something close, or add a configure check but these HAVE_X86_CPU_FEATURES are not very usefull
[22:00] <michaelni> burek, sure, do you have a suggestion or ... better ... a patch :)
[22:03] <burek> :)
[22:03] <burek> Well, just add the same line, like you did in the source code
[22:03] <burek> just mention that the option is meant for -f image2, that'll do
[23:39] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r8743400961 10ffmpeg/libavcodec/h264.c: 
[23:39] <CIA-41> ffmpeg: h264: reset prev_poc_msb to the same value everywhere.
[23:39] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:39] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r3dea28cc2e 10ffmpeg/libavcodec/h264.c: 
[23:39] <CIA-41> ffmpeg: h264: Fix concealment regression introduced by babf4fe01a808327d53977ba319c113a930180b1
[23:39] <CIA-41> ffmpeg: Limit the new case to when the decoder is flushed instead of at each idr frame
[23:39] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:39] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rcdba9b617f 10ffmpeg/doc/eval.texi: 
[23:39] <CIA-41> ffmpeg: eval.texi: document that variables are not shared between expressions.
[23:39] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[00:00] --- Sun Dec 11 2011