[Ffmpeg-devel-irc] ffmpeg-devel.log.20130113

[burek]

[06:38] <cone-658> ffmpeg.git 03Michael Niedermayer 07master:a6e4796fbf0a: pthread: Do not use a half updated context as master for deallocation.
[12:25] <durandal_1707> it is really disapointer that nobody answers in ml
[12:28] <saste> durandal_1707, which patch?
[12:29] <durandal_1707> no patch, but user questions on user* ml
[12:49] <durandal_1707> ATTENTION!!!
[12:49] <durandal_1707> http://forum.doom9.org/showthread.php?p=1610364&highlight=ffmpeg#post1610364
[12:56] <wm4> ffmpeg helping applications having to add horrible workarounds
[12:56] <wm4> because people will use it anyway, and such mismuxed files might make it into the wild
[12:58] <durandal_1707> michaelni: the commit that allows muxing opus before matroska even defined how it shoul be done shoule be reverted asap
[13:24] <funman> durandal_1707: how about adding EXPERIMENTAL tag like mkvtoolnix does?
[13:40] <nevcairiel> but its not compatible with mkvtoolnix experimental version =p
[13:42] <durandal_1707> so remove it, feel free to properly implement experimental version
[13:47] <cone-153> ffmpeg.git 03Diego Biurrun 07master:4d3b144c5ea8: fate: cosmetics: Order some test entries
[13:47] <cone-153> ffmpeg.git 03Michael Niedermayer 07master:841a288cb327: Merge commit '4d3b144c5ea824193019019d33740a1ae9e0bb69'
[13:49] <michaelni> durandal_1707, ive mailed nicolas about the opus issue, i suspect he will look into it soon
[13:50] <michaelni> but dont hesitate to disable it before nicolas has time
[13:59] <cone-153> ffmpeg.git 03Diego Biurrun 07master:0f94c0b1b68a: fate: vp3: Fix fate-vp3-coeff-level64 test dependencies
[13:59] <cone-153> ffmpeg.git 03Martin Storsjö 07master:abae27ed3acd: rtpdec: Fix the calculation of expected number of packets
[13:59] <cone-153> ffmpeg.git 03Michael Niedermayer 07master:6d6eb7c12c83: Merge commit 'abae27ed3acd0a7c54f11760c5be2d2653c4edf8'
[14:01] <durandal_1707> so dolby-e is not supported?
[14:02] <kierank> no
[14:04] <durandal_1707> but multimedia.cx claims it is
[14:05] <cone-153> ffmpeg.git 03Martin Storsjö 07master:e568db40258d: rtpdec: Calculate and report packet reception jitter
[14:05] <cone-153> ffmpeg.git 03Michael Niedermayer 07master:6cd1dbe6dffe: Merge commit 'e568db40258d549777ac1c16971678e18a18f5f5'
[14:06] <kierank> it is writing about my unreleased decoder
[14:08] <durandal_1707> the sample in mxf is reported as pcm
[14:08] <kierank> that's the point
[14:09] <kierank> the point of dolby e is to masquarade as pcm to pass through legacy equipment
[14:10] <nevcairiel> does it have the same silly rules as DTS to be like, only 14-bit or so, so its not too loud when played accidentally like this? .d
[14:11] <kierank> they seem to run some sort of XOR on some variants of the bitstream which might be for that reason
[14:12] <cone-153> ffmpeg.git 03Martin Storsjö 07master:22c436c85e0d: rtpdec: Send a valid "delay since SR" value in the RTCP RR packets
[14:12] <cone-153> ffmpeg.git 03Michael Niedermayer 07master:3bcf443f91d8: Merge remote-tracking branch 'qatar/master'
[14:13] <durandal_1707> kierank: so 3 years passed and nobody interested in donating some money?
[14:13] <kierank> not about money
[14:13] <durandal_1707> but original issue on roundup mentioned it
[14:13] <kierank> things change
[14:14] <kierank> the issue is dolby will sue anyone who uses it professionally (including myself)
[14:14] <JEEB> ya
[14:14] <JEEB> dolby and dts -- the folk who'll never license anything (F)OSS
[14:15] <durandal_1707> but we have dts encoder
[14:15] <kierank> durandal_1707: doesn't mean you can use it professionally
[14:15] <kierank> same with ac3
[14:15] <durandal_1707> so even decoder is no-go?
[14:16] <kierank> i will release it when "the process" (which i can't elaborate on) that will allow me and others to use it professionally ends
[14:17] <kierank> also fwiw there are problems in lavf/lavc with regards to demux that make dolby e integration hard without hacks
[14:17] <kierank> i guess less of a problem these days because you can just put one track as a substream in an avframe
[14:18] <kierank> not 100% correct mind you
[14:19] <kierank> getting it to demux in mxf will be a total pita
[14:25] <cone-153> ffmpeg.git 03Paul B Mahol 07master:5b9126ba845d: lavfi/gradfun: support YUV440P
[14:30] <pross-au> kierank: just put your encoder in international waters. license prblm solved.
[15:14] <cone-153> ffmpeg.git 03Stefano Sabatini 07master:c499d45c6bea: doc/texipod: add rule to correctly interpret @ref{ANCHOR,XREF,SECTION_NAME,...}
[15:14] <cone-153> ffmpeg.git 03Stefano Sabatini 07master:89ef5868679f: doc/ffmpeg: extend documentation for the -filter options and -af/vf aliases
[15:14] <cone-153> ffmpeg.git 03Stefano Sabatini 07master:78e27c44c3c2: lavf/utils: clarify/extend messages in avformat_find_stream_info()
[15:14] <cone-153> ffmpeg.git 03Stefano Sabatini 07master:98247e3368fb: lavf: clarify/extend documentation for AVFormatContext::analyzeduration
[15:32] <cone-153> ffmpeg.git 03Michael Niedermayer 07master:dc6588421e09: swr: move buffer used to discard sample to context
[15:32] <cone-153> ffmpeg.git 03Michael Niedermayer 07master:db4e0ecaf0e9: swr: move silence buffer to context to avoid per use malloc/free
[16:04] <cone-153> ffmpeg.git 03Stefano Sabatini 07master:e8bab4c6fca0: ffplay: extend documentation for the -vf option
[16:04] <cone-153> ffmpeg.git 03Michael Niedermayer 07master:b481d09bd9c7: swr: limit buffer size for discarding.
[16:04] <cone-153> ffmpeg.git 03Michael Niedermayer 07master:a34aee464628: swr: limit buffer size for silence injection
[17:57] <rik__> I have a question
[17:58] <jmbclegend> it was a dream
[17:58] <rik__> in ffio_limit I see a reference to the maxsize field of an AVIOContext. In some of the demuxers, I see that value set when the header is read and sometimes it's not mentioned anywhere in the demuxer at all.
[17:59] <rik__> So I'm assuming that there's some default value that it's set to. Where does that happen if it does?
[18:05] <kierank> cbsrobot: ping
[18:11] <michaelni> rik__, default is 0 set during allocation
[18:11] <rik__> michaelni, thanks a lot
[19:09] <cone-484> ffmpeg.git 03Michael Niedermayer 07master:d676598f879b: swr: fix handling of timestamps that cause multiple drops or silence injections
[20:04] <cbsrobot> kierank: pong
[20:05] <kierank> cbsrobot: what can decode that sample of yours
[20:05] <kierank> with cid 1244
[20:05] <kierank> dnhxd-tr
[20:05] <cbsrobot> quicktime can
[20:06] <cbsrobot> with dnxhd codec of course
[20:07] <cbsrobot> I guess it's AvidAVdnCodec.component version 2.1.0
[20:09] <kierank> i reckon it's just dnhxd with a  certain bitrate/resolution
[00:00] --- Mon Jan 14 2013