[Ffmpeg-devel-irc] ffmpeg-devel.log.20120118

burek burek021 at gmail.com
Thu Jan 19 02:05:03 CET 2012

[00:00] <durandal_1707> there is single corrupted frame near at end with second sample(did not tried first one)
[00:01] <durandal_1707> cbsrobot: so make sure to test both samples
[00:05] <cbsrobot> durandal_1707: first sample has artifacts at the bottom in vlc, but not in ffplay
[00:05] <cbsrobot> second sample is as you say - but no artifacts
[00:08] <durandal_1707> second sample in mplayer have strange last line, so maybe that is regression...
[00:09] <durandal_1707> it appears to have no color at all
[00:10] <CIA-31> ffmpeg: 03Stefano Sabatini 07master * r54661219c1 10ffmpeg/ffprobe.c: 
[00:10] <CIA-31> ffmpeg: ffprobe: fix printing of unit values which cannot be contained in an int
[00:10] <CIA-31> ffmpeg: Use long long int to contain such values instead of an int, which is
[00:10] <CIA-31> ffmpeg: required to contain at least 64 bits, so it is guaranteed to contain also
[00:10] <CIA-31> ffmpeg: int64_t values, which are used by some fields.
[00:10] <CIA-31> ffmpeg: In particular, should fix trac ticket #921.
[00:17] <cbsrobot> it looks like there are some color artifacts at the top row, left side 
[00:17] <cbsrobot> michaelni: do you see that too ?
[00:25] <michaelni> cbsrobot, which video with which player and where on the timeline ?
[00:25] <cbsrobot> both from https://trac.videolan.org/vlc/ticket/5872
[00:26] <cbsrobot> a bit everywhere - just the top row left side
[00:26] <cbsrobot> ffplay
[00:26] <michaelni> there are 2 videos in that ticket
[00:26] <michaelni> cmi & outlaw
[00:26] <cbsrobot> both
[00:27] <michaelni> is this also vissible if you convert to jpg/png with ffmpeg ? 
[00:27] <michaelni> because if yes, please upload a picture of these artifacts
[00:29] <cbsrobot> no I cant see them
[00:35] <michaelni> i cant reproduce this :(
[00:45] <CIA-31> ffmpeg: 03Stefano Sabatini 07master * r140a9afcf2 10ffmpeg/ffprobe.c: 
[00:45] <CIA-31> ffmpeg: ffprobe: add more safe casts in value_string()
[00:45] <CIA-31> ffmpeg: Second attempt at fixing ticket #921.
[00:45] <CIA-31> ffmpeg: 03Stefano Sabatini 07master * r2c8d37c2fb 10ffmpeg/ffprobe.c: 
[00:45] <CIA-31> ffmpeg: ffprobe: use safer type for index variable in value_string()
[00:45] <CIA-31> ffmpeg: There is no guarantee that the casted double which is assigned to the
[00:45] <CIA-31> ffmpeg: variable will be contained in an int (also if it is almost sure for most
[00:45] <CIA-31> ffmpeg: non-alien architectures).
[02:37] <CIA-31> ffmpeg: 03Anton Khirnov 07master * r9993284793 10ffmpeg/avconv.c: 
[02:37] <CIA-31> ffmpeg: avconv: fix -frames for video encoders with delay.
[02:37] <CIA-31> ffmpeg: Frames must be counted when they are passed to the encoder, not when
[02:37] <CIA-31> ffmpeg: they come out.
[02:37] <CIA-31> ffmpeg: Fixes Bug 202.
[02:37] <CIA-31> ffmpeg: 03Justin Ruggles 07master * r01ed1c390d 10ffmpeg/libavcodec/tta.c: 
[02:37] <CIA-31> ffmpeg: tta: cast output data pointer to the correct type
[02:37] <CIA-31> ffmpeg: fixes "warning: assignment from incompatible pointer type"
[02:37] <CIA-31> ffmpeg: 03Michael Niedermayer 07master * re191f1f414 10ffmpeg/: 
[02:37] <CIA-31> ffmpeg: Merge remote-tracking branch 'qatar/master'
[02:37] <CIA-31> ffmpeg: * qatar/master:
[02:37] <CIA-31> ffmpeg:  tta: cast output data pointer to the correct type
[02:37] <CIA-31> ffmpeg:  avconv: fix -frames for video encoders with delay.
[02:37] <CIA-31> ffmpeg: Merged-by: Michael Niedermayer <michaelni at gmx.at>
[04:46] <piece3> Hi all, does ffmpeg currently offer muxing more than 2 video streams simultaneously? or is this feature not yet implemented?
[04:49] <piece3> Excuse my last question, wrong channel, plus I have just confirmed it does :) have a good day
[06:32] <Compn> wikipedia is still up
[06:32] <Compn> what a liar jimbo is
[06:32] <Compn> unless i'm stuck in some cache 
[06:36] <t1g3r> blockd from here
[07:07] <michaelni> Compn, the wiki blackout only works with javascript enabled AFAIK
[07:08] <michaelni> also its just the english wikipedia
[08:28] <av500> ffmpeg.org is also blacked out?
[08:50] <ubitux> making wikipedia available when you have no js is a real joke&
[08:51] <av500> ubitux: its not about making it unavailable
[08:51] <av500> they even state in the help section how to disable the blackout
[08:51] <av500> like using m.wikipedia
[08:51] <ubitux> yes i read it
[08:52] <av500> and most non-geek users have js enabled and these are the ones they try to reach
[08:52] <ubitux> if it doesn't annoy people i guess the effect will be way less effective, but just my PoV
[08:53] <av500> it will annoy enough people
[08:53] <ubitux> it would have been nice to blacklist USA instead
[08:53] <ubitux> to force them to use another country as proxy
[08:53] <ubitux> could have been fun :)
[08:56] <av500> ubitux: https://twitter.com/#!/herpderpedia
[08:57] <ubitux> haha
[09:00] <ubitux> "I think Wikipedia planned this shit."
[09:00] <ubitux> there are good ones :')
[09:00] <av500> yep
[13:42] <Compn> michaelni : ah, i usually browse without javascript
[13:42] <Compn> since wikipedia pages have become so bloated
[14:33] Action: Compn writes another run on sentence
[14:33] <Compn> grammar getting worse :)
[16:33] <durandal_1707> hmm, ffplay gives bmp little darker and less sharp than feh/ImageMagick
[16:34] <durandal_1707> happens with any image format, so sws is really busted
[16:39] <CIA-31> ffmpeg: 03Dmitry Volyntsev 07master * rb29ec002bb 10ffmpeg/libavformat/ (rtsp.c rtsp.h): 
[16:39] <CIA-31> ffmpeg: fix rtp 916 issue
[16:39] <CIA-31> ffmpeg: Reviewed-by: Luca Abeni <lucabe72 at email.it>
[16:39] <CIA-31> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:39] <CIA-31> ffmpeg: 03Dmitry Volyntsev 07master * r4259d1eb99 10ffmpeg/libavformat/rtsp.c: 
[16:39] <CIA-31> ffmpeg: fix rtp 916 issue, patch 2
[16:39] <CIA-31> ffmpeg: Reviewed-by: Luca Abeni <lucabe72 at email.it>
[16:39] <CIA-31> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:39] <CIA-31> ffmpeg: 03Michael Niedermayer 07master * re49780f749 10ffmpeg/libavcodec/utils.c: (log message trimmed)
[16:39] <CIA-31> ffmpeg: lavc: call ff_init_buffer_info() from default reget buffer.
[16:39] <CIA-31> ffmpeg: This is needed in case the get_buffer() callback doesnt set
[16:39] <CIA-31> ffmpeg: width/height.
[16:39] <CIA-31> ffmpeg: Ideally all decoders would make calls through some wraper
[16:39] <CIA-31> ffmpeg: to the callbacks and that wraper would call ff_init_buffer_info()
[16:39] <CIA-31> ffmpeg: But until thats done, the default reget buffer must call this
[16:39] <CIA-31> ffmpeg: 03Michael Niedermayer 07master * r2fb65d3e55 10ffmpeg/libavcodec/utils.c: 
[16:40] <CIA-31> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:42] <michaelni> durandal_1707, ffplay uses SDL fixed in YUV420 mode so RGB gets its chroma downconverted which surely doesnt help sharpness
[16:43] <michaelni> the darker could have various reasons from bugs in sws over bugs in SDL to gamma correction differences
[16:43] <ubitux> (hey fate is almost all green!)
[16:44] <ubitux> (time to add a valgrind tracking undef read)
[16:45] Action: michaelni is working on adding asan
[16:45] <ubitux> oh? great :)
[16:45] <ubitux> are you expecting it to raise things valgrind doesn't handle?
[16:46] <michaelni> a quick preliminary tests had something in eac3 failing IIRC
[16:46] <ubitux> oh? ok
[16:46] <michaelni> asan is able to detect stack issues better than valgrind and valgrind is better at leaks and undef reads AFAIK
[16:47] <michaelni> i also tried safecode yesterday
[16:47] <ubitux> this make me thing we should have one build with -fstack-protector-all or sth like that
[16:47] <ubitux> but i think it's the default on openbsd
[16:47] <michaelni> but it deadlocked somehow in mpeg2 -therads > 1
[17:24] <durandal_1707> anybody have working Xorg 1 or 4 depth display please upload xwd -root dump
[17:24] <durandal_1707> or find xwd with gray colorspace
[18:35] <durandal_1707> xy image format is ugly to support in ffmpeg
[21:22] <CIA-31> ffmpeg: 03Reimar Döffinger 07master * rf9ced97543 10ffmpeg/libavcodec/ (fraps.c mpc.h mpc7.c): 
[21:22] <CIA-31> ffmpeg: Use av_fast_padded_malloc in fraps and mpc decoders.
[21:22] <CIA-31> ffmpeg: Fixes FATE failures due to uninitialized reads under valgrind
[21:22] <CIA-31> ffmpeg: for these two codecs.
[21:22] <CIA-31> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[21:23] <j-b> m
[21:23] <j-b> oops
[21:41] <gnafu> Moops?
[21:42] <funman> moo process show : a CowOS command
[21:44] Action: gnafu prefers BullOS.
[21:46] <CIA-31> ffmpeg: 03Michael Niedermayer 07master * r126b04196e 10ffmpeg/libpostproc/postprocess.h: 
[21:46] <CIA-31> ffmpeg: pp: bump major in git master to match what debian contains.
[21:46] <CIA-31> ffmpeg: The ABI differences are negligible, but its easier for all if
[21:46] <CIA-31> ffmpeg: all distros have libpostproc HEAD under the same soname and
[21:46] <CIA-31> ffmpeg: debian bumped soname without consulting upstream, so as silly as
[21:46] <CIA-31> ffmpeg: it is following this is probably the least pain for all.
[21:46] <CIA-31> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:04] <CIA-31> ffmpeg: 03Reimar Döffinger 07master * re8b060f7e7 10ffmpeg/libavcodec/ (mimic.c truemotion2.c): 
[22:04] <CIA-31> ffmpeg: Use av_fast_padded_malloc in mimic and truemotion2 decoders.
[22:04] <CIA-31> ffmpeg: Fixes use of uninitialized data errors under valgrind.
[22:04] <CIA-31> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[22:04] <CIA-31> ffmpeg: 03Reimar Döffinger 07master * r6fd4b8a6ef 10ffmpeg/libavcodec/fraps.c: 
[22:04] <CIA-31> ffmpeg: Replace a av_fast_malloc with av_fast_padded_malloc.
[22:04] <CIA-31> ffmpeg: This one was missed in the previous fraps fix, the
[22:04] <CIA-31> ffmpeg: allocation is exactly the same in both cases.
[22:04] <CIA-31> ffmpeg: Fixes fraps-v5 under valgrind.
[22:04] <CIA-31> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[00:00] --- Thu Jan 19 2012

More information about the Ffmpeg-devel-irc mailing list