[Ffmpeg-devel-irc] ffmpeg-devel.log.20141105

[burek]

[00:44] <pross> wm: for BGGR8, odd lines have 'BGBG' components, even lines have 'GRGR' components. There are no blue on the even lines.
[00:44] <cehoyos> pross: Note that scaling fails badly: https://trac.ffmpeg.org/ticket/4053
[00:45] <pross> noted. why on earth would you want to scale bayer!
[00:47] <pross> but i agree it should not crash.
[00:49] <cehoyos> If it refuses to scale, FFmpeg will happily choose yuv420p, if it crashes, this is difficult...
[00:49] <cehoyos> But as said, I still don't understand how to get useful output at all;-(
[00:49] <pross> Do you need samples?
[00:52] <cehoyos> I have two samples with two output tiff files: How can I produce similar looking output?
[00:58] <pross> cehoyos: thats the white balance issue right?
[00:59] <pross> bayer data is (usually) unprocessed: it is the raw data returned by the image sensor.
[01:03] <cehoyos> I guess it is the white balance issue: I think we should document how to get useful output from Phantom Cine input (if this cannot be done automatically).
[01:38] <cone-311> ffmpeg.git 03Carl Eugen Hoyos 07master:f2ffaae9ac93: Use -fno-optimize-sibling-calls on parisc also for gcc 4.9.
[02:09] <cone-311> ffmpeg.git 03Michael Niedermayer 07master:77f1199e8fd9: avcodec/mpeg12dec: do not trust AVCodecContext input dimensions
[02:17] <cone-311> ffmpeg.git 03Rémi Denis-Courmont 07master:4ff670d99beb: hwaccel: Deinitialize hardware acceleration early enough
[02:17] <cone-311> ffmpeg.git 03Michael Niedermayer 07master:43fee7ad920d: Merge commit '4ff670d99bebd97429322719089363d83143477d'
[02:26] <cone-311> ffmpeg.git 03Lou Logan 07master:37425fcb0474: doc: clarify -frames options behavior
[09:58] <cone-95> ffmpeg.git 03Michael Niedermayer 07release/2.4:cd57d608a412: avcodec/mpeg12dec: do not trust AVCodecContext input dimensions
[09:58] <cone-95> ffmpeg.git 03Michael Niedermayer 07release/2.4:f9ca1fecb0fd: ffmpeg_opt: store canvas size in decoder context
[10:32] <cone-95> ffmpeg.git 03Michael Niedermayer 07release/2.4:a5cc8775cf1d: avcodec/h264_sei: ff_h264_decode_sei: dont try to parse trailing zeroes
[10:32] <cone-95> ffmpeg.git 03Carl Eugen Hoyos 07release/2.4:9798dc8061c9: Use -fno-optimize-sibling-calls on parisc also for gcc 4.9.
[10:34] <cone-95> ffmpeg.git 03Martin Storsjö 07master:a490391157dc: rtmpproto: Ignore errors from the getStreamLength method
[10:34] <cone-95> ffmpeg.git 03Michael Niedermayer 07master:513d57cc4d33: Merge commit 'a490391157dcf4dc6b65352ec3eea2781dd0a404'
[11:25] <cone-95> ffmpeg.git 03Thomas Volkert 07master:07c3a4f69336: avformat/udp: UDP-Lite (RFC 3828) support added
[18:11] <cone-95> ffmpeg.git 03Kevin Mitchell 07master:fdf22f973d41: avfilter/vf_idet: add a repeated field detection
[19:00] <cone-95> ffmpeg.git 03Michael Niedermayer 07master:a604de4fd89e: avutil/time_internal: do not attempt to override *time_r() macros
[19:43] <cone-95> ffmpeg.git 03Andrey Utkin 07master:4fd70679877b: v4l2: support MPEG4 compressed streams
[20:08] <cone-95> ffmpeg.git 03Reimar Döffinger 07master:a26d0ffa18db: vdpau_mpeg4: Do not fail on unknown profile.
[21:05] <ubitux> i think i will still keep a N dialogues per subtitles
[21:05] <ubitux> it can be useful in some situations
[21:06] <ubitux> typically if you have a markup that can't be expressed in a single dialog
[21:06] <ubitux> any objection?
[21:11] <cone-95> ffmpeg.git 03Nicolas George 07master:90cdec5e2698: ffmpeg: init sub2video.last_pts.
[21:27] <ubitux> "Consider using a tool like VirtualDub or avidemux to fix it."
[21:27] <ubitux> ffmpeg can't do it?
[21:29] <Daemon404> at least the ydidnt say nandub
[21:35] <ubitux> maybe we should mention mencoder
[21:40] <Compn> fix it with dd and a hexeditor
[23:58] <J_Darnley> What's the usual command to run to create configure script for an autotools project?  ./autogen.sh?
[23:59] <kepstin-laptop> if there is an autogen.sh then yeah
[23:59] <kepstin-laptop> in most projects you can use "autoreconf"
[00:00] --- Thu Nov  6 2014