[Ffmpeg-devel-irc] ffmpeg-devel.log.20160614

[burek]

[02:22:46 CEST] <cone-356> ffmpeg 03Michael Niedermayer 07master:00c4861f13d9: avformat/mpegts: adjust probe score for low check_count
[04:30:14 CEST] <cone-356> ffmpeg 03Michael Niedermayer 07master:228568fc8ac4: fate: fate-filter-acrossfade needs samples
[05:00:09 CEST] <kylophone> If anyone else has an opinion or wants to review to ebur128 patches on the ML, please do so.
[05:00:20 CEST] <kylophone> I'm planning on sending an updated patchset later tomorrow and want to make sure I address concerns if there are any.
[08:05:51 CEST] <cone-956> ffmpeg 03Martin Storsjö 07master:03aa4cd22b4b: rtpdec_vp9: Make sure to free the temp buffer on close
[08:05:51 CEST] <cone-956> ffmpeg 03Martin Storsjö 07master:43880f971662: rtpdec_vp9: Update header parsing to spec draft 02
[08:05:51 CEST] <cone-956> ffmpeg 03Martin Storsjö 07master:94dee9067acd: rtpdec: Use AVERROR_PATCHWELCOME instead of AVERROR(ENOSYS) for unimplemented features
[08:05:51 CEST] <cone-956> ffmpeg 03Martin Storsjö 07master:d50b3fe46313: rtpdec_h264: Use avpriv_report_missing_feature instead of a manual av_log
[08:05:51 CEST] <cone-956> ffmpeg 03Martin Storsjö 07master:cab237953ea6: rtpdec_vp9: Support parsing the scalability structure
[08:24:35 CEST] <cone-956> ffmpeg 03Clément BSsch 07master:fef89aae98a0: lavc/h264: move green meta logging out of the sei decoding
[08:24:35 CEST] <cone-956> ffmpeg 03Clément BSsch 07master:b0ec36c9b168: lavc/h264_sei: reduce scope of parameters for green meta decode
[08:24:36 CEST] <cone-956> ffmpeg 03Clément BSsch 07master:5342bb624dd9: lavc/h264: rename GreenMetaData to H264SEIGreenMetaData
[08:24:37 CEST] <cone-956> ffmpeg 03Clément BSsch 07master:7dd8cd193329: lavc/h264_sei: fix broken style around green metadata code
[12:43:02 CEST] <cone-956> ffmpeg 03Stefano Sabatini 07master:4b5620446ada: lavf/utils: fix timestamp trace log in estimate_timings() function
[12:43:03 CEST] <cone-956> ffmpeg 03Stefano Sabatini 07master:6bcc6fd55d28: doc/muxers/mpegts: apply misc consistency/formatting fixes
[17:03:46 CEST] <cone-956> ffmpeg 03Carl Eugen Hoyos 07master:d5ded429f25c: lavc/dpx: Support decoding 10 bit colourspace with transparency information.
[17:10:34 CEST] <pfelt> greatest of mornings all!
[17:43:41 CEST] <cone-956> ffmpeg 03James Almer 07master:ede4ec1f8f5f: x86/aacpsdsp: optimize add_squares loop
[18:16:02 CEST] <cone-956> ffmpeg 03Michael Niedermayer 07master:fa0097e008a1: avfilter/vf_telecine: Make frame writable before writing into it
[18:18:39 CEST] <ubitux> btw, next merge in progress @ https://github.com/ubitux/FFmpeg/commits/merge-libav
[18:18:43 CEST] <ubitux> it compiles but doesn't work yet
[18:19:05 CEST] <ubitux> (see notes/TODOs in the WIP message)
[18:19:23 CEST] <ubitux> ofc, help welcome, but i haven't really debugged anything yet
[18:24:52 CEST] <durandal_1707> ubitux: what's it about?
[18:26:06 CEST] <ubitux> splitting h264context out of the sei decoding
[18:26:23 CEST] <ubitux> (so the sei decoding in the parser doesn't depend on the h264context)
[18:26:34 CEST] <ubitux> first one was the pps/sps for the same reasons
[18:26:45 CEST] <ubitux> and next one will be the extradata, for the same reasons as well
[18:46:21 CEST] <rcombs> ubitux: oh nice
[18:48:24 CEST] <JoshX> ubitux: you might be able to get me in the right direction talking about SEI's :-)
[18:49:16 CEST] <JoshX> ubitux: only i want to encode userdata in the SEI headers, so i'm trying to figure out where i need to hook this into the ffmpeg sources
[19:32:56 CEST] <ubitux> JoshX: encoder ’ libx264?
[19:43:30 CEST] <JoshX> ubitux: yes that much is clear :-)
[19:44:00 CEST] <JoshX> what i want is the system timestamp as epoch in a sei userdata field every second
[19:44:33 CEST] <ubitux> when i said libx264, i meant the libx264 sources, so not in ffmpeg
[19:44:34 CEST] <JoshX> what i managed is to find the position where to do it (I guess) and print the epoch to the screen in the routine that encodes the SEI headers..
[19:44:42 CEST] <JoshX> ah!
[19:48:32 CEST] <JoshX> ubitux: and then probably in encoder.c?
[19:48:57 CEST] <ubitux> x264 is not part of the ffmpeg project, i can't help you
[19:49:03 CEST] <ubitux> "git grep -i sei" is your friend
[19:49:07 CEST] <ubitux> can't say more
[19:49:19 CEST] <JoshX> thanks anyway for pointing in this direction :)
[21:17:27 CEST] <chovy> does anyone know the answer to this? https://www.mail-archive.com/ffmpeg-user%40ffmpeg.org/msg09688.html
[21:25:03 CEST] <Compn> chovy : sorry, i do not. hmm there should be someway to choose the playlist...
[21:25:19 CEST] <Compn> probably just defaults to the first stream :\
[21:25:32 CEST] <Compn> wonder if it uses -map ?
[21:32:07 CEST] <chovy> Compn: i can't find anything regarding it
[21:32:29 CEST] <chovy> i thought maybe something like -index 'index_2400_p.m3u3' or whatever it would be.
[21:32:40 CEST] <chovy> i'm not familiar with map or how it would help
[21:34:29 CEST] <chovy> but i think you're right. in my case it choses the first stream probably which is the worst one.
[21:37:19 CEST] <cone-956> ffmpeg 03Aman Gupta 07master:e9391ab121f9: avformat/mpegts: enhance logging in trace mode
[21:45:35 CEST] <chovy> Compn: i tried -map 0:11 to get the 12th stream in the manifest but it seems to be disregarded. 
[21:49:25 CEST] <cone-956> ffmpeg 03Aman Gupta 07master:ddffafc0b049: avformat/utils: in debug mode, print number of streams found before avformat_find_stream_info()
[23:50:33 CEST] <cone-956> ffmpeg 03Hendrik Leppkes 07master:cb46b78b2ec0: os_support: use the appropriate stat functions matching the stat type
[00:00:00 CEST] --- Wed Jun 15 2016