[Ffmpeg-devel-irc] ffmpeg-devel.log.20140201

[burek]

[02:30] <llogan> kierank: can i leave your name as backup mentor for H.264 MVC? also, do you have any additional project ideas?
[07:06] <cone-546> ffmpeg.git 03Michael Niedermayer 07master:85d51d8e327c: avcodec/vc1dec: field pictures with direct mode MBs, followed by frame pictures are not supported
[07:06] <cone-546> ffmpeg.git 03Michael Niedermayer 07master:f4b288a639bb: avcodec/vc1dec: Print warning for frame pictures with direct mode MBs, followed by field pictures
[17:17] <cone-434> ffmpeg.git 03Michael Niedermayer 07master:388b4cf86ed5: avcodec/vc1: factor read_bfraction() out
[17:17] <cone-434> ffmpeg.git 03Michael Niedermayer 07master:dcf5bfbdb613: avcodec/vc1: Check bfraction_lut_index
[19:22] <cone-434> ffmpeg.git 03Michael Niedermayer 07master:b51e9354772d: avcodec/vc1: reset fcm/field_mode in non advanced header parsing
[20:44] <JEEB> btw, what's this start parameter that lavf/ffmpeg is outputting Duration: 00:01:36.10, start: 7079.000000, bitrate: 33811 kb/s ?
[20:52] <nevcairiel> JEEB: start_time in AVFormatContext i assume, ie the offset of the first timestamp
[20:52] <JEEB> ah
[20:53] <JEEB> and just got to avformatcontext->start_time
[20:54] <JEEB> I was just looking at it because someone was complaining that their ~01:30 mpeg-ts played for longer than that
[20:54] <JEEB> and that start variable looked suspiciously close to that
[21:05] <nevcairiel> mpegts very rarely starts at exactly 0
[21:05] <nevcairiel> at least if its from broadcast
[22:39] <lukaszmluki> Hi, does anyone know how to present AVRational in  AVOptionRange struct?
[22:49] <JEEB> any simple way of listing all tracks from an MPEG-TS file?
[22:49] <JEEB> uhh, not TS
[22:49] <JEEB> PS
[22:49] <JEEB> even ones with unknown identifiers
[22:56] <nevcairiel> i thought ffmpeg would show all, even if unknown tracks with none codec
[23:01] <JEEB> either this file is muxed in a weird way or that doesn't seem to be so :s
[23:06] <wm4> lukaszmluki: a rational range?
[23:08] <cone-542> ffmpeg.git 03addr-see-the-website at aetey.se 07master:80a79f264136: RoQ decoder: Remove misleading diagnostics.
[23:08] <cone-542> ffmpeg.git 03addr-see-the-website at aetey.se 07master:392b0345d6d2: RoQ-encoder: introducing Quake 3 compatibility option
[23:41] <michaelni> lukaszmluki, value_min/max are limits on the value of the rational, component_min/max would limit the 2 integers
[23:43] <ubitux> michaelni: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=718805  IIRC you kept those assert; it seems they're triggerable somehow with v4l2
[23:43] <ubitux> given how much changes we have in v4l2, that might not be true though
[23:49] <michaelni> ubitux, that assert fails when the timebase is negative or something like that, removing the assert is not likely going to make that work
[23:50] <ubitux> yes i know, i'm not asking to remove those assert
[23:50] <ubitux> but it might reveal a real bug in v4l2
[23:51] <michaelni> s/is negativ/is negative or 0/
[23:53] <wm4> why is it an assert instead of returning an error?
[23:53] <ubitux> because you can't return an error
[23:53] <michaelni> and because nothing checks the return code
[23:59] <JEEB> hmm, "Found stream id 0xBD  = Private Stream 1" , where do I add this in libavformat/mpeg.c to make it try and select this stream?
[00:00] --- Sun Feb  2 2014