[Ffmpeg-devel-irc] ffmpeg-devel.log.20120312

[burek]

[03:45] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r9fcc62ed9b 10ffmpeg/ffserver.c: 
[03:45] <CIA-17> ffmpeg: Revert "ffserver: remove unused variable."
[03:45] <CIA-17> ffmpeg: This reverts commit f02f745e4aa070e5fabf1156b75c340171304933.
[03:45] <CIA-17> ffmpeg: 03Miroslav SlugeH 07master * r00969376ce 10ffmpeg/ffserver.c: 
[03:45] <CIA-17> ffmpeg: ffserver: add missing buffer set after rewrite av_open_input_file to avformat_open_input.
[03:45] <CIA-17> ffmpeg: Fixes Ticket346
[03:45] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[04:49] <CIA-17> ffmpeg: 03Ronald S. Bultje 07master * rc95fefa042 10ffmpeg/libavcodec/dsicinav.c: 
[04:49] <CIA-17> ffmpeg: dsicinvideo: validate buffer offset before copying pixels.
[04:49] <CIA-17> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[04:49] <CIA-17> ffmpeg: CC: libav-stable at libav.org
[04:50] <CIA-17> ffmpeg: 03Ronald S. Bultje 07master * r97e48b2f54 10ffmpeg/libavcodec/cook.c: 
[04:50] <CIA-17> ffmpeg: cook: error out on quant_index values outside [-63, 63] range.
[04:50] <CIA-17> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[04:50] <CIA-17> ffmpeg: CC: libav-stable at libav.org
[04:50] <CIA-17> ffmpeg: 03Ronald S. Bultje 07master * rd7eabd5042 10ffmpeg/libavcodec/ (mpc.c mpc7.c mpcdata.h): 
[04:50] <CIA-17> ffmpeg: mpc: pad mpc_CC/SCF[] tables to allow for negative indices.
[04:50] <CIA-17> ffmpeg: MPC8 allows indices of mpc_CC up to -1, and mpc_SCF up to -6, thus pad
[04:50] <CIA-17> ffmpeg: the tables by that much on the left end.
[04:50] <CIA-17> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[04:50] <CIA-17> ffmpeg: CC: libav-stable at libav.org
[04:50] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r1c27359867 10ffmpeg/: (log message trimmed)
[04:50] <CIA-17> ffmpeg: Merge remote-tracking branch 'qatar/master'
[04:50] <CIA-17> ffmpeg: * qatar/master:
[04:50] <CIA-17> ffmpeg:  dsicinvideo: validate buffer offset before copying pixels.
[04:50] <CIA-17> ffmpeg:  cook: error out on quant_index values outside [-63, 63] range.
[04:50] <CIA-17> ffmpeg:  mpc: pad mpc_CC/SCF[] tables to allow for negative indices.
[04:50] <CIA-17> ffmpeg: Conflicts:
[06:05] <CIA-17> ffmpeg: 03C Chatterjee 07master * r3c9a86df0e 10ffmpeg/libavformat/flvenc.c: 
[06:05] <CIA-17> ffmpeg: flvenc: Remove apparently unneeded AAC check.
[06:05] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[06:05] <CIA-17> ffmpeg: 03ami_stuff 07master * r86b6e49d92 10ffmpeg/libavcodec/iff.c: 
[06:05] <CIA-17> ffmpeg: iff: check for pbm tag
[06:05] <CIA-17> ffmpeg: more robust
[06:05] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[06:05] <CIA-17> ffmpeg: 03Tudor.Suciu 07master * r9abd47fe90 10ffmpeg/libavformat/udp.c: 
[06:05] <CIA-17> ffmpeg: udp: add multicast input interface selection
[06:05] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[07:02] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r2367b2a502 10ffmpeg/libavformat/libavformat.v: 
[07:02] <CIA-17> ffmpeg: libavformat.v: add ffio_set_buf_size() for ffserver.
[07:02] <CIA-17> ffmpeg: its inevitable with the current API
[07:02] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[07:02] <CIA-17> ffmpeg: 03Hendrik Leppkes 07master * ra3c5aefff2 10ffmpeg/libavcodec/dxva2_vc1.c: 
[07:02] <CIA-17> ffmpeg: dxva2_vc1: fix decoding of BI frames
[07:02] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[07:02] <CIA-17> ffmpeg: 03Hendrik Leppkes 07master * r0aedd8c0bd 10ffmpeg/libavcodec/dxva2_vc1.c: 
[07:02] <CIA-17> ffmpeg: dxva2_vc1: pass the overlap flag to the decoder
[07:02] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[07:02] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r88c76c7d22 10ffmpeg/libavformat/udp.c: 
[07:02] <CIA-17> ffmpeg: Revert "udp: add multicast input interface selection"
[07:02] <CIA-17> ffmpeg: This reverts commit 9abd47fe90a592822b326345726626a018641d04.
[07:02] <CIA-17> ffmpeg: Broke compilation on openbsd and others
[14:24] <kriegerod> Hi. Please look at Pavel Sokolov's mail http://ffmpeg.org/pipermail/ffmpeg-devel/2012-March/121930.html . He pays for solution.
[14:32] <cbsrobot>  bcoudurier: looks like encoding to dnxhd using -mdb 2 is broken on ffmpeg and ffmbc
[14:33] <cbsrobot> *-mbd 2
[14:34] <durandal_1707> does ffmpeg supports converting betwen unplanar and planar sample fmts?
[14:36] <cbsrobot> unplanar = packed ?
[14:36] <durandal_1707> packed /unplanar channels are interleaved
[14:37] <durandal_1707> planar channel is on own plane
[14:59] <michaelni> durandal_1707, swresample supports it, ffmpeg should too
[15:00] <durandal_1707> michaelni: than i can remove stupid hacks in pcm decoder?
[15:02] <michaelni> not sure which hacks you mean but if it works without the hacks sure you can remove them
[15:11] <iive> durandal_1707: the question is, would it break some programs that use it?
[15:11] <durandal_1707> come one grow up
[15:39] <kriegerod> what is the reason that UDP protocol reader buffers datagrams instead of bytestream?
[15:42] <michaelni> i think there was some protocol which needed the packets
[15:43] <kriegerod> michaelni, thanls. Another question: what do you think about patch adding option that turns udp "OVERRUN" fatal error to warning?
[15:47] <michaelni> could be useful ...
[16:44] <funman> http://mailman.videolan.org/pipermail/vlc-devel/2012-March/087489.html
[16:52] <ubitux> funman: what do you meant by "support is easier to get from Libav devs"? :p
[16:52] <ubitux> is there something we can improve here?
[16:52] <funman> i think mostly faster IRC answer around the days when i commited this
[16:53] <JEEBsv> I think it depends heavily on what exactly you are in need of
[16:53] <bcoudurier> I don't think so
[16:54] <JEEBsv> ...and on one's luck
[16:54] <JEEBsv> (whether or not people are AFK)
[16:54] <funman> yeah, mostly i wanted VLC devs to react but they are as annoyed as me by the fork so it was just ignored
[16:54] <bcoudurier> from what I've seen everything gets integrated faster in ffmpeg
[16:55] <bcoudurier> fenrir security fixes, features, codecs
[16:57] <kriegerod> i found that native RTMP implementation generates SIGPIPE sometimes. Could we add sth like setsockopt(sd, SOL_SOCKET, SO_NOSIGPIPE) as temporary measure?
[16:58] <funman> bcoudurier: fenrir said he prefered FFmpeg though
[16:59] <bcoudurier> :)
[16:59] <JEEBsv> I certainly got my AAC decoding hacks done faster in ffmpeg :)
[17:00] <j-b> I prefer the one with better porn playback
[17:00] <JEEBsv> :D
[17:00] <JEEBsv> WhatReallyMatters
[17:08] <j-b> bcoudurier: https://trac.videolan.org/vlc/ticket/6151 "By the way, using FFMBC I can transcode this to something visible to the decode"
[17:36] <funman> follow up ... http://mailman.videolan.org/pipermail/vlc-devel/2012-March/087500.html
[18:18] <CIA-17> ffmpeg: 03Clément BSsch 07master * r4fe5448790 10ffmpeg/libavfilter/vf_drawtext.c: 
[18:18] <CIA-17> ffmpeg: lavfi/drawtext: add 24 hours wrapping option for timecode.
[18:18] <CIA-17> ffmpeg: Fixes ticket #1044.
[18:23] <j-b> nevcairiel: only submitted to ffmpeg, though, right?
[18:23] <nevcairiel> i lost track which patches i submitted where
[18:24] <nevcairiel> those two where sitting on the ML for weeks
[18:24] <j-b> confused too?
[18:25] <nevcairiel> i also implemented interlaced vc1 dxva2 decoding, but i doubt it would pass any review because it completely ignores the fact that there are also other hwaccels. :)
[18:26] <Compn> oo, carl could review that
[18:26] <Compn> he was talking about interlaced vc1 
[18:26] <Compn> hes up on vdpau stuff too
[18:26] <Compn> not sure if he has time to review big pieces of code tho
[18:27] <nevcairiel> i also implemented Header and MB level parsing for interlaced vc-1 b-frames, however i have no idea how to do all the MV/MC code, so its sitting here unfinished
[18:27] <nevcairiel> maybe i should RFC is to the mailing list
[18:31] <Compn> kostya is wmv master tho :)
[18:31] <nevcairiel> i tried enlisting help from the libav people, but they didn't seem interested =p
[18:32] <Compn> well it cant hurt to ping them again , dont want to lose patches , no matter where they are
[18:33] <nevcairiel> i could probably spent a weekend and try to figure out how its supposed to work
[18:33] <nevcairiel> but time is short, etc
[20:28] <michaelni> btw, does anyone have the vc1 reference bitstreams ? (they would be usefull to test any vc1 interlaced patches/changes)
[20:44] <Amnesia_> hi, anyone willing to help me with the pad filter over here?
[20:44] <Amnesia_> ( I did read the documentation )
[20:53] <cbsrobot> Amnesia_:  I guess that's an #ffmpeg question
[20:53] <Amnesia_> already asked it:)
[20:53] <Amnesia_> no success yet
[20:54] <Amnesia_> btw I also spoke to you concerning this very issue last weekend:)
[20:55] <cbsrobot> Amnesia_: that's why I reply to your question on the other channel now ....
[21:15] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rb0a0e83aee 10ffmpeg/libavformat/flvenc.c: (log message trimmed)
[21:15] <CIA-17> ffmpeg: Revert "flvenc: Remove apparently unneeded AAC check."
[21:15] <CIA-17> ffmpeg: According to video_file_format_spec_v10_1.pdf flv stores AAC RAW
[21:15] <CIA-17> ffmpeg: thanks to Baptiste Coudurier for pointing that out
[21:15] <CIA-17> ffmpeg: thanks to Aex Converse for explaining:
[21:15] <CIA-17> ffmpeg:  This can't be at the start of a non-ADTS payload. 111 is the
[21:15] <CIA-17> ffmpeg:  EndOfFrame syntax element.
[21:25] <Daemon404> j-b, any pings back?
[22:10] <CIA-17> ffmpeg: 03Andrey Utkin 07master * ra2eecc110b 10ffmpeg/libavformat/udp.c: 
[22:10] <CIA-17> ffmpeg: udp: remove dead code line (unused result)
[22:10] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:10] <CIA-17> ffmpeg: 03Andrey Utkin 07master * r3069e70f62 10ffmpeg/libavformat/udp.c: 
[22:10] <CIA-17> ffmpeg: udp: Add option overrun_nonfatal
[22:10] <CIA-17> ffmpeg: Optionize fail/survive on circular buffer overrun
[22:10] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:10] <CIA-17> ffmpeg: 03Nico Weber 07master * r599888a480 10ffmpeg/libavcodec/x86/fft_mmx.asm: (log message trimmed)
[22:10] <CIA-17> ffmpeg: Move struc FFTContext below SECTION_RODATA
[22:10] <CIA-17> ffmpeg: Yasm creates an implicit unaligned text section if "struc" is used
[22:10] <CIA-17> ffmpeg: outside of any section:
[22:10] <CIA-17> ffmpeg: http://tortall.lighthouseapp.com/projects/78676-yasm/tickets/247
[22:10] <CIA-17> ffmpeg: Since yasm only honors the "align" annotation on the first declaration
[22:10] <CIA-17> ffmpeg: of a section, this implicit text section causes all text section
[23:02] Action: Daemon404 gives j-b an extra poke for good luck and makes some coffee
[00:00] --- Tue Mar 13 2012