[Ffmpeg-devel-irc] ffmpeg-devel.log.20120102

burek burek021 at gmail.com
Tue Jan 3 02:05:03 CET 2012


[00:19] <durandal_1707> done
[03:25] <CIA-101> ffmpeg: 03Janne Grunau 07master * r60d626650a 10ffmpeg/tests/ (12 files in 2 dirs): 
[03:25] <CIA-101> ffmpeg: fate: use rgb24 as output format for dfa tests
[03:25] <CIA-101> ffmpeg: Palette is as supposed in native endianness. Converting the pal8 output
[03:25] <CIA-101> ffmpeg: to rgb24 is thus necessary for identical CRCs on big and little endian
[03:25] <CIA-101> ffmpeg: systems.
[03:25] <CIA-101> ffmpeg: 03Aneesh Dogra 07master * rda7c65f0ce 10ffmpeg/avplay.c: 
[03:25] <CIA-101> ffmpeg: avplay: K&R cleanup
[03:25] <CIA-101> ffmpeg: Signed-off-by: Anton Khirnov <anton at khirnov.net>
[03:25] <CIA-101> ffmpeg: 03Janne Grunau 07master * rb6064d9a59 10ffmpeg/libavcodec/utils.c: threads: set thread_count to 1 when thread support is disabled
[03:25] <CIA-101> ffmpeg: 03Janne Grunau 07master * rb12d217339 10ffmpeg/libavcodec/pthread.c: 
[03:25] <CIA-101> ffmpeg: threads: limit the number of automatic threads to MAX_AUTO_THREADS
[03:25] <CIA-101> ffmpeg: The extra thread added in {frame_}*thread_init was not taken into
[03:25] <CIA-101> ffmpeg: account. Explicitly sets thread_count to 1 if only one CPU core was
[03:25] <CIA-101> ffmpeg: detected. Also fixes two typos in comments.
[03:25] <CIA-101> ffmpeg: 03Janne Grunau 07master * rb4d44a45f9 10ffmpeg/libavcodec/ (avcodec.h libx264.c pthread.c version.h): 
[03:25] <CIA-101> ffmpeg: threads: introduce CODEC_CAP_AUTO_THREADS and add it to libx264
[03:25] <CIA-101> ffmpeg: Some external codecs have their own code to determine the best number
[03:26] <CIA-101> ffmpeg: bfi: K&R cosmetics
[03:26] <CIA-101> ffmpeg: Signed-off-by: Anton Khirnov <anton at khirnov.net>
[03:26] <CIA-101> ffmpeg: 03Anton Khirnov 07master * rf15f02c204 10ffmpeg/avconv.c: avconv: make -frames work for all types of streams, not just video.
[03:26] <CIA-101> ffmpeg: 03Michael Niedermayer 07master * r4121148388 10ffmpeg/: (log message trimmed)
[03:26] <CIA-101> ffmpeg: Merge remote-tracking branch 'qatar/master'
[03:26] <CIA-101> ffmpeg: * qatar/master:
[03:26] <CIA-101> ffmpeg:  avconv: make -frames work for all types of streams, not just video.
[03:26] <CIA-101> ffmpeg:  bfi: K&R cosmetics
[03:26] <CIA-101> ffmpeg:  bgmc: K&R cleanup
[03:26] <CIA-101> (18 lines omitted)
[04:03] <CIA-101> ffmpeg: 03Paul B Mahol 07master * rdfa77dead2 10ffmpeg/ (10 files in 4 dirs): 
[04:03] <CIA-101> ffmpeg: y41p encoder and decoder
[04:03] <CIA-101> ffmpeg: y41p is a packed 12-bit 4:1:1 YUV format used by Brooktree.
[04:03] <CIA-101> ffmpeg: Fixes issue 1123 / ticket #102.
[04:03] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[04:03] <CIA-101> ffmpeg: 03Jean First 07master * rb103156235 10ffmpeg/libavcodec/j2kdec.c: 
[04:03] <CIA-101> ffmpeg: j2kdec: av_log formatting use %tx instead of %x
[04:03] <CIA-101> ffmpeg: Signed-off-by: Jean First <jeanfirst at gmail.com>
[04:03] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[04:03] <CIA-101> ffmpeg: 03Jean First 07master * re2c1d3ec41 10ffmpeg/libavformat/flvdec.c: 
[04:04] <CIA-101> ffmpeg: flvdec: use av_uninit to silence compiler warnings
[04:04] <CIA-101> ffmpeg: Signed-off-by: Jean First <jeanfirst at gmail.com>
[04:04] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[07:22] <ubitux> michaelni: i need to get the drawtext lavc dep patch done; i'll do it today, and will certainly commit it in the following days; please wait until then for the 0.9.1
[08:45] <ubitux> mmh, strange thing, i can't play anymore mp3 encoded with ffmpeg on my mp3 player
[08:46] <ubitux> i wonder if it's an issue in lame
[08:55] <ubitux> ok it seems to come from ffmpeg
[09:02] <ubitux> it works with latest ffmpeg 0.8
[09:03] <ubitux> and outputs totally differs mmmh.
[09:12] <ubitux> ok it seems 1fea432b7a6957b800006fd7c4bd9b4076b0cc06 is at fault
[09:17] <ubitux> looks sane to me though...
[09:24] <ubitux> it's indeed this commit; output differs after while they shouldn't
[09:24] <kcm1700> hmm...
[09:27] <ubitux> ah i know
[09:27] <ubitux> header is modified in the loop
[09:28] <ubitux> and reused after
[09:28] <ubitux> thus, the mask need to be kept
[09:28] <kcm1700> can 12th~15th bits in header already be set?
[09:28] <ubitux> mmh or maybe not.
[09:30] <ubitux> ok the "if (needed <= c.frame_size)" doesn't trigger the mask disabling
[09:30] <ubitux> so the header shouldn't have its mask disabled
[09:30] <kcm1700> aha.
[09:35] <ubitux> didn't test it properly, but this should do the trick: http://pastie.org/3111074
[09:36] <ubitux> but now i wonder if 40176fc3149bc72c1309f93989a57aec2322e626 is still necessary
[09:38] <ubitux> ok i confirm it fixes the playback issue
[09:46] <ubitux> patch sent.
[10:00] <kcm1700> how about or-ing after the loop?
[10:00] <kcm1700> I like the simplified version.
[10:01] <ubitux> i prefer the working one ;)
[10:02] <ubitux> kcm1700: i thought about that, but the or-ing seems to depend on if it trigger the case or not 
[10:02] <ubitux> (the if needed...)
[10:03] <ubitux> i let the ppl knowing what does this code mean decide :p
[10:03] <ubitux> i don't want to assume wrong things
[10:04] <kcm1700> indeed.
[14:23] <ubitux> does anyone have mov samples with timecode tracks in "frame count" and in "mov" formats?
[14:24] <ubitux> i added the timecode extraction from mov, but i have samples where the header indicates the timecode is in mov format while it is actually in "frame count" format
[14:25] <ubitux> i have a hacky heuristic to fix the behaviour but it sucks a bit&
[14:25] <ubitux> if i could get some more samples, that would be nice ;p
[15:16] <ubitux> btw, any way to do some emphasis in texi?
[15:17] <ubitux> @emph{}, ok.
[15:35] <CIA-101> ffmpeg: 03Clément BSsch 07master * r940d3cc0b0 10ffmpeg/ (doc/filters.texi libavfilter/vf_drawtext.c): drawtext: fix hard dependency to lavc (timecode).
[15:35] <CIA-101> ffmpeg: 03Clément BSsch 07master * re1a8164d62 10ffmpeg/libavformat/mp3enc.c: 
[15:35] <CIA-101> ffmpeg: mp3enc: fix playback with mp3 encoded files on some devices.
[15:35] <CIA-101> ffmpeg: This partially reverts 1fea432b; header variable needs to keep its value
[15:35] <CIA-101> ffmpeg: in case it triggers "if (needed <= c.frame_size)".
[15:39] <CIA-101> ffmpeg: 03Michael Niedermayer 07master * ra2a12e3358 10ffmpeg/libavcodec/vp3.c: 
[15:39] <CIA-101> ffmpeg: vp3: fix regression with mplayer-crash.ogv
[15:39] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[17:19] <CIA-101> ffmpeg: 03Michael Niedermayer 07master * r0af34c4d2c 10ffmpeg/MAINTAINERS: 
[17:19] <CIA-101> ffmpeg: MAINTAINERS: update release maintainers
[17:19] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[17:36] <CIA-101> ffmpeg: 03Michael Niedermayer 07release/0.7 * r1218f8ed49 10ffmpeg/libavcodec/vp3.c: 
[17:36] <CIA-101> ffmpeg: vp3: fix regression with mplayer-crash.ogv
[17:36] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[17:36] <CIA-101> ffmpeg: (cherry picked from commit a2a12e3358c3bbdc0246ffc94973e58eba50ee30)
[17:36] <CIA-101> ffmpeg: 03Michael Niedermayer 07release/0.8 * rc409ac5adc 10ffmpeg/libavcodec/vp3.c: 
[17:36] <CIA-101> ffmpeg: vp3: fix regression with mplayer-crash.ogv
[17:36] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[17:36] <CIA-101> ffmpeg: (cherry picked from commit a2a12e3358c3bbdc0246ffc94973e58eba50ee30)
[20:22] <CIA-101> ffmpeg: 03Michael Niedermayer 07release/0.7 * r870e74dc43 10ffmpeg/ (Doxyfile RELEASE VERSION): 
[20:22] <CIA-101> ffmpeg: Update for 0.7.10
[20:22] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:22] <CIA-101> ffmpeg: 03Michael Niedermayer 07release/0.8 * rcee1568ae1 10ffmpeg/ (Doxyfile RELEASE VERSION): 
[20:22] <CIA-101> ffmpeg: Update for 0.8.9
[20:22] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:47] <ubitux> bcoudurier: hi; what do you mean by "yet"? no app generates such media atm?
[22:50] <bcoudurier> doesn't matter for now
[22:50] <bcoudurier> once a sample is found, it will be added
[22:50] <ubitux> ok& so basically i have to assume frame num format all the time?
[22:51] <ubitux> sounds wrong to me but well, should do the trick for all the samples i have
[22:52] <bcoudurier> as long as it doesn't fail it's fine it would just show a wrong timecode
[22:53] <bcoudurier> same for the check framerate, if it's wrong it shouldn't fail demuxing, it should be fine though the ret value is not considered in your patch
[22:54] <ubitux> yes return values in read_timecode_track are ignored
[22:54] <ubitux> it just displays err/warn
[23:10] <durandal_1707> why our libpostproc version is different from libav? latest mplayer2 configure fails because of this
[00:00] --- Tue Jan  3 2012


More information about the Ffmpeg-devel-irc mailing list