[Ffmpeg-devel-irc] ffmpeg-devel.log.20130209

[burek]

[00:31] Action: Compn wonders why someone is encoding to qtrle in 720p
[00:36] <kierank> why not
[00:39] <Compn> whats qtrle good for ? seems like an old codec with poor compression ?
[00:39] <Compn> maybe i dont know how good it is
[00:40] <Compn> but most codecs from the year 2004 arent that great for hd.
[00:40] <Compn> probably earlier, but going by copyright year :P
[00:42] <kierank> it has alpha channel
[00:42] <kierank> that's why its useful
[01:02] <llogan> saste: any changes to "sort demuxers" other than moving the text?
[01:19] <cone-778> ffmpeg.git 03Michael Niedermayer 07master:a92816c4eb9d: get_se_golomb: fix decoding of very large values
[04:06] <cone-778> ffmpeg.git 03James Almer 07master:53b2083f4433: sparc: Revert some parameter name changes in dsputil_vis.c
[04:13] <cone-778> ffmpeg.git 03rogerdpack 07master:aa715cdcf5c6: dshow: attempt to build shared as well as static
[04:13] <cone-778> ffmpeg.git 03Michael Niedermayer 07master:a35f5c22339c: Merge remote-tracking branch 'rdp/dshow_shared'
[05:31] <llogan> haven't seen a gcc seg fault in a while...
[05:31] <llogan> i mean until now
[06:17] <Compn> llogan : not an ICE then ?
[11:26] <durandal_1707> michaelni: the il filter have option to (de)interleave chroma and luma separately
[11:26] <durandal_1707> and alpha is completely forgotten
[11:26] <durandal_1707> so i will fix this by making it more generic
[11:27] <durandal_1707> keepeing old syntax though does not make sense
[11:49] <cone-273> ffmpeg.git 03Diego Biurrun 07master:0b016eb99d38: dsputil: Move ff_block_permute to mpegvideo_enc
[11:49] <cone-273> ffmpeg.git 03Michael Niedermayer 07master:55151aa1013b: Merge commit '0b016eb99d38738e2c53e36549a4732a0f863b2e'
[12:10] <cone-273> ffmpeg.git 03Diego Biurrun 07master:b5b7b75e91e2: dsputil: Move get_penalty_factor() to the only place it is used.
[12:10] <cone-273> ffmpeg.git 03Michael Niedermayer 07master:bcb15e66f7c4: Merge commit 'b5b7b75e91e291c810ebd01868a42b53f267a87c'
[12:42] <cone-273> ffmpeg.git 03Mans Rullgard 07master:ed62e6e3c3ca: mjpegdec: use put_pixels instead of copy_block8
[12:42] <cone-273> ffmpeg.git 03Mans Rullgard 07master:b9ee5f2cab3f: indeo3: replace use of copy_block4 with put_pixels
[12:42] <cone-273> ffmpeg.git 03Michael Niedermayer 07master:e5212354eadd: Merge commit 'b9ee5f2cab3ffe1c962e542346b1ed61394864ec'
[12:45] <durandal_1707> michaelni: do not push hidden stuff into merge, like this snowenc change
[12:56] <cone-273> ffmpeg.git 03Paul B Mahol 07master:29a92c011480: histogram filter
[13:13] <michaelni> durandal_1707, the code doesnt compile without that change
[13:14] <durandal_1707> michaelni: yes, but hidden commits are worse
[13:15] <michaelni> my commit messages suck i know, i should have explained it in the merge commit
[13:15] <michaelni> its on my todo to see if they should be merged or
[13:15] <michaelni> if we can gain something from seperate functions
[13:17] <durandal_1707> michaelni: just commit it before merge
[13:17] <durandal_1707> quatar removed snow so it should not matter
[13:20] <durandal_1707> michaelni: why #1280 picks gray?
[13:22] <michaelni> dunno
[13:22] <michaelni> i could look into it later 
[13:22] <michaelni> libavfilter/vf_histogram.c:208:13: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement]
[13:24] <durandal_1707> when we will finally upgrade from ISO C90?
[13:24] <burek> ..A.L. ilbc                 iLBC (Internet Low Bitrate Codec)
[13:24] <burek> it doesnt say if it is encoder/decoder
[13:25] <JEEB> not better than opus
[13:25] <durandal_1707> burek: because you do not have either
[13:25] <burek> if i do not have, why is it listed?
[13:26] <burek> it*
[13:26] <durandal_1707> others are listed tooo
[13:26] <burek> even worse
[13:26] <burek> why
[13:26] <durandal_1707> to notify users of it existence
[13:26] <burek> i see
[13:26] <durandal_1707> you should still be able to demux files
[13:27] <durandal_1707> and you would know what codec is it about
[13:27] <burek> well, then maybe they should be sorted/groupped somehow at the end/beginning, to make it clear they are not available
[13:27] <durandal_1707> burek: there is no native decoder, instead you enable wrapper...
[13:27] <burek> yes, i understand that
[13:27] <burek> using external library
[13:28] <michaelni> durandal_1707, C90, when all compilers we support support that
[13:28] <durandal_1707> michaelni: translation: never
[13:30] <cone-273> ffmpeg.git 03Paul B Mahol 07master:e5670cb8a1c9: lavfi/histogram: make it C90 compliant
[13:30] <iive> imho, declarations in any location in the code is misfeature. it solves a problem caused by another misfeature
[13:31] <durandal_1707> iive: not really....
[13:37] <iive> i'm waiting for your explanation.
[13:42] <someone-noone> Hello! I'm trying to pass "magic cookie"(extradata) data to mac os Core Audio framework for decoding AAC format. File is in m4a format. For this I'm using extradata from my AVCodecContext. But extradata_size is 7, however if I read "magic cookie" using OS X API, it's size is 44 for same file. Looks like magic cookie != extradata in ffmpeg (but should be). Any help how can I extract this data from ffmpeg? 
[13:44] <cone-273> ffmpeg.git 03Diego Biurrun 07master:6a4430407404: dsputil: Move ff_h264_idct function declarations to a separate header
[13:44] <cone-273> ffmpeg.git 03Michael Niedermayer 07master:0af07495eccc: Merge commit '6a4430407404bfe55581e66a86a496c9adf6fb44'
[13:51] <cone-273> ffmpeg.git 03Diego Biurrun 07master:75d5156ac152: dsputil: Move ff_svq3 function declarations to a separate header
[13:51] <cone-273> ffmpeg.git 03Diego Biurrun 07master:35685a3c2a1e: dsputil: Move ff_shrink* function declarations to separate header
[13:51] <cone-273> ffmpeg.git 03Michael Niedermayer 07master:ffbe19fc84fe: Merge commit '35685a3c2a1ec09f3c62dcfc4368fe9e92bcddf6'
[13:52] <burek> does anyone have any idea what to use (in ffmpeg) to produce a low latency speech audio (for use in p2p audio chat)?
[13:53] <burek> whatever i try, it doesnt work, except aac/mpegts (over udp) which has a significant delay :(
[13:54] <someone-noone> burek, first of all you need low latency codec. The most talkable codec is OPUS for now.
[13:56] <burek> well i tried opus, like: ... -acodec opus -f mpegts udp://localhost:5555
[13:56] <burek> but ffplay wont play it back
[13:57] <nevcairiel> i dont think there is a defined way to store opus in mpegts
[13:57] <someone-noone> mpegts doesn't support opus
[13:57] <JEEB> did the "ogg container" have it?
[13:57] <burek> no
[13:57] <JEEB> then there's that matroska mapping that differs from the one that mkvmerge implemented
[13:57] <JEEB> which might or might not have been removed from ffmpeg
[13:57] <burek> well is there a standard way to stream voice/speech with ffpmeg (preferably over udp)
[13:58] <nevcairiel> ogg is about the only thing that properly supports opus today
[13:58] <JEEB> yeah
[13:58] <nevcairiel> short of raw opus files, which may just be ogg, which i never checked =p
[13:58] <JEEB> :D
[13:58] <JEEB> I tried have him find 'opus' from the output of -formats
[13:58] <JEEB> and there was none there
[13:59] <wm4> is proper opus in mkv defined yet?
[13:59] <JEEB> nope
[13:59] <nevcairiel> no
[13:59] <wm4> why?
[13:59] <JEEB> so both mkvmerge and ffmpeg IIRC now have conflicting experimental ways of putting it there
[13:59] <wm4> lol
[13:59] <burek> well i might use speex i guess, i just need the hint for the output format used with udp
[13:59] <nevcairiel> opus is problematic because it requires pre-roll data to start decoding
[13:59] <JEEB> ^
[13:59] <cone-273> ffmpeg.git 03Diego Biurrun 07master:b9c2408b19dc: dsputil: Remove commented-out, unused function declarations
[13:59] <cone-273> ffmpeg.git 03Michael Niedermayer 07master:a8864615e6cc: Merge commit 'b9c2408b19dc9408e9dc0ce4a099af529bbfba3b'
[14:00] <nevcairiel> and how to handle this  has caused the difficulty defining the proper standard in mkv
[14:00] <JEEB> and matroska, unlike mov/mp4, doesn't have that concept in the container until now
[14:00] <JEEB> and they want to "define it properly"
[14:00] <JEEB> somehow I feel like I'll be liking xiph next-gen container if matroska just doesn't start evolving >_>
[14:01] <burek> is it possible there is no "standard" or preferred way in ffmpeg (so far) to have it send low latency audio to another peer?
[14:01] <burek> i mean, it looks so advanced with everything
[14:01] <nevcairiel> ffmpeg is mostly a transcoding application, streaming is not its main goal
[14:01] <burek> i see
[14:02] <burek> well, ok, thanks.. at least i know where not to look then i guess..
[14:10] <someone-noone> Okay, I found difference between magic cookie and extradata for AAC. OS X meaning of magic cookie is esds ATOM in mp4 file. However ffmpeg stores only part (in the middle, only 7 bytes) of it in extradata and it looks like enough for ffmpeg to decode content in AAC. Anyway how can, probably ffmpeg stores somewhere(while decoding) this atom in AVCodecContext structure ?  
[14:13] <Paranoialmaniac> someone-noone: libavcodec doesn't store atom
[14:13] <Paranoialmaniac> atom is mp4/mov thing
[14:14] <Paranoialmaniac> aac specific info is stored in extradata as the same format between mp4 and mov
[14:15] <someone-noone> Paranoialmaniac, I see. But is there a way I can mux this data into my container that is a part of libavformat, that I could read it in another application that doesn't use libav*? 
[14:15] <nevcairiel> all the important data is stored in extradata, you just have to put all the atom data back around it
[14:15] <nevcairiel> if you need it to be an atom
[14:17] <Paranoialmaniac> someone-noone: i have no idea about libavformat muxing :P sorry
[14:17] <someone-noone> I have an idead. Look at m4a format& It's somehow playing by OS X :)
[14:21] <someone-noone> hm, looks like I can't find m4a format in libavformat folder. Doesn't it name should be mpegaudio or m4a ?
[14:21] <nevcairiel> m4a is just a normal mp4 file with a different extension
[14:21] <nevcairiel> apple just uses it for audio files, but its the same format
[14:21] <Paranoialmaniac> mov.c
[14:21] <someone-noone> didn't know that& thanks
[14:22] <Paranoialmaniac> mp4 != m4a, m4v. pure mp4 requires object descriptor and bifs for presentation :P
[14:24] <nevcairiel> its the same damn format, may just have different atoms somewhere =p
[14:24] <Paranoialmaniac> well, yes. mp4, m4a and m4v is derived from iso base media file format
[14:25] <someone-noone> well, I found mov_write_esds_tag that is doing some really crazy things :(
[14:25] <Paranoialmaniac> esds atom includes not only mpeg-4 audio specific info but also bitrate info
[14:26] <Paranoialmaniac> esds atom consists of mpeg-4 system structure
[14:26] <someone-noone> I need smth like uint8_t *esds_atom = get_atom(); OSXOpenCodec(esds_atom+4)
[14:26] <someone-noone> Really, there is no explanation in Apple's docs :(
[14:27] <cone-273> ffmpeg.git 03Diego Biurrun 07master:bf6b3ec924b4: dsputil: Move rnd_avg inline functions to a separate header
[14:27] <cone-273> ffmpeg.git 03Michael Niedermayer 07master:81a11452f8e0: Merge commit 'bf6b3ec924b4fb64d14df33077f4d4541d525dbf'
[14:27] <cone-273> ffmpeg.git 03Michael Niedermayer 07master:5fd6d85d17d5: rnd_avg: fix author attribution
[14:32] <cone-273> ffmpeg.git 03Diego Biurrun 07master:059866eb17cd: dsputil: Move WRAPPER8_16_SQ macro to the only place it is used
[14:32] <cone-273> ffmpeg.git 03Michael Niedermayer 07master:276c9c0577e9: Merge commit '059866eb17cd3c66422a61d3e9c3305f55830a50'
[14:40] <cone-273> ffmpeg.git 03Diego Biurrun 07master:218aefce4472: dsputil: Move LOCAL_ALIGNED macros to libavutil
[14:40] <cone-273> ffmpeg.git 03Michael Niedermayer 07master:6b2e65078c02: Merge commit '218aefce4472dc02ee3f12830a9a894bf7916da9'
[14:44] <durandal_1707> iive: for (int i ....)
[14:45] <durandal_1707> michaelni: i get strange internal.h macro E redefintions warnings
[14:46] <durandal_1707> #define E(x) x ????
[14:47] <nevcairiel> its for macro expansion
[14:48] <nevcairiel> some compilers need an extra expansion to properly expand all parameters
[14:48] <durandal_1707> but there is already E defined in bunch of other places for opt stuff
[14:49] <durandal_1707> In file included from libavcodec/options.c:35:
[14:49] <durandal_1707> libavcodec/options_table.h:39:9: warning: 'E' macro redefined
[14:49] <durandal_1707> #define E AV_OPT_FLAG_ENCODING_PARAM
[14:49] <durandal_1707>         ^
[14:49] <durandal_1707> ./libavutil/internal.h:63:9: note: previous definition is here
[14:49] <durandal_1707> #define E(x) x
[14:49] <durandal_1707>         ^
[14:49] <durandal_1707> libavcodec/options.c:111:5: warning: 'av_opt_set_defaults2' is deprecated [-Wdeprecated-declarations]
[14:49] <durandal_1707>     av_opt_set_defaults2(s, flags, flags);
[14:49] <durandal_1707>     ^
[14:49] <durandal_1707> 2 warnings generated.
[14:49] <nevcairiel> apparently libav doesnt have those other E defines =p
[14:50] <durandal_1707> nevcairiel: i doubt it
[14:50] <nevcairiel> or they dont include that header there
[14:50] <durandal_1707> this is the new thing recently happening
[14:52] <durandal_1707> nevcairiel: and this is introduced when moving local aligned stuff
[14:52] <durandal_1707> fairly recently
[14:52] <nevcairiel> i saw the commit that caused it
[14:53] <durandal_1707> good. i will just undefine macro
[14:53] <nevcairiel> that macro is needed
[14:54] <cone-273> ffmpeg.git 03Diego Biurrun 07master:5d3d39c72e57: dsputil: Move fdct function declarations to dct.h
[14:54] <cone-273> ffmpeg.git 03Michael Niedermayer 07master:5cfc0ae825e6: Merge remote-tracking branch 'qatar/master'
[14:54] <durandal_1707> but if you did not yet notice this breaks when using own E in opt stuff before using LOCAL....
[14:55] <nevcairiel> maybe just rename E to EXPAND or something
[14:59] <cone-273> ffmpeg.git 03Stefano Sabatini 07master:111697170e14: doc/demuxers: sort demuxers by name
[14:59] <cone-273> ffmpeg.git 03Stefano Sabatini 07master:9b9b5a475b99: doc/demuxers: fix copy&paste error in "@c man end" line
[14:59] <cone-273> ffmpeg.git 03Stefano Sabatini 07master:765dbea9fe1f: lavf/rawvideodec: fix/extend option descriptions
[14:59] <cone-273> ffmpeg.git 03Stefano Sabatini 07master:bad82243cd83: doc/demuxers: add docs for rawvideo demuxer
[14:59] <cone-273> ffmpeg.git 03Stefano Sabatini 07master:b46a8f055ee5: doc/filters: end table in histogram docs
[15:00] <durandal_1707> beastd: so native resamples is fast as sox one?
[15:01] <beastd> hi
[15:01] <durandal_1707> welcome on board
[15:02] <beastd> durandal_1707: i did not measure the speed of native vs sox resampling. i just found that new thread on HA somewhat misleading. should check if something was added
[15:03] <cone-273> ffmpeg.git 03Carl Eugen Hoyos 07master:7a259a5fcdc3: Fix libpostproc compilation on x86_32 with disable-optimizations.
[15:04] <durandal_1707> beastd: what thread, there is trollolo old few days
[15:06] <beastd> i understood from michaelni that you talked about ffmpeg resampling on HA. Or did you mean the old thread? I found this one http://www.hydrogenaudio.org/forums/index.php?showtopic=99286
[15:06] <durandal_1707> i mean whatever thread that was recently active
[15:07] <durandal_1707> yes that thread .....
[15:10] <beastd> durandal_1707: did you compare speed of native vs libsoxr resampling?
[15:10] <durandal_1707> beastd: no, i'm busy with removing mp
[15:12] <durandal_1707> and negate filter is really broken on big endian?
[15:12] <durandal_1707> why is there no fate coverage?
[15:13] <durandal_1707> saste: ping
[15:15] <saste> durandal_1707, what?
[15:16] <durandal_1707> saste: i ported il filter, did you mind if old syntax is not keept?
[15:16] <durandal_1707> it does not make sense to keep it
[15:16] <saste> durandal_1707, how hard would be to keep the old syntax?
[15:16] <saste> lemme check
[15:17] <durandal_1707> il filter does not copy alpha plane at all
[15:17] <saste> about lut/negate, I cannot test
[15:29] <cone-273> ffmpeg.git 03Paul B Mahol 07master:da726a8763d3: paf, sanm: remove unused header after b05830169d
[15:30] <saste_> michaelni, do you want to review the il filter?
[15:30] <saste_> also tell if you're ok with GPL -> LGPL
[15:31] <durandal_1707> saste: read logs, he already approved change
[15:31] <durandal_1707> it is few memcpy calls after all
[15:31] <saste> durandal_1707, ok
[15:31] <michaelni> saste, review for() memcpy ? no ;)
[15:31] <durandal_1707> :))
[15:32] <saste> that's not only that
[15:32] <saste> i'm going to review it, leave me some time
[15:32] <durandal_1707> i said, i need to fix 4 and 2 plane situation
[15:33] <durandal_1707> this means keeping old syntax is not possible at all
[15:33] <saste> durandal_1707, i don't mind about that
[15:35] <saste> durandal_1707, is mp=il broken or what?
[15:36] <durandal_1707> saste: not it is working fine
[15:36] <durandal_1707> saste: mp=il=i,mp=il=d
[15:36] <saste>  ffplay IN -vf mp=il=d
[15:37] <saste> i got chroma artefacts
[15:37] <durandal_1707> saste: mp=il=d:d
[15:37] <saste> otoh i confess i don't know what the filter really does
[15:37] <saste> yeah
[15:42] <saste> durandal_1707, what's the plan with regards to mp=fil?
[15:42] <durandal_1707> remove
[15:43] <durandal_1707> it is just il that works correctly only when width match linesize
[15:43] <saste> i mean it is possible to port such a filter (which according to docs, "does not always work" in unspecified circumstances))
[15:43] <durandal_1707> and fast when works
[15:43] <wm4> haha
[15:43] <wm4> saste: yeah that freaked me out when I saw it
[15:44] <wm4> then I removed it from my mplayer fork
[15:44] <durandal_1707> wm4: because you do not know what you are removing ....
[15:45] <wm4> well, have fun rewriting it
[15:45] <durandal_1707> saste: i could add option to il to use fil way when it will work ...
[15:45] <saste> durandal_1707, seems a good idea
[15:53] <durandal_1707> wm4: now one cant filter interlaced video with mpv and display it on tv
[15:54] <wm4> durandal_1707: how would one achieve this with "fil"?
[15:55] <durandal_1707> calling fil=d, fitlers, fil=i (note code is from 2002)
[15:57] <wm4> filtering interlaced material and then playing it on a interlaced display...?
[15:58] <durandal_1707> yes
[15:59] <wm4> well, I guess there are quite some 2002 use cases lost by removing some of these filters, so if you care about these...
[16:00] <wm4> I think extended filtering should be left to tools which are suitable for these tasks, such as avisynth (and hopefully vsynth)
[16:06] <ramiro> michaelni: I don't remember having commit rights to the git repo. I don't think I ever sent my public key (even if I did, I've changed it a few times in the past two years)
[16:22] <michaelni> ramiro, theres a key for Ramiro Polla on the server since Jan 19 2011
[17:03] <durandal_1707> saste: the nv21/nv12 works with il
[17:25] <cone-273> ffmpeg.git 03Paul B Mahol 07master:4fa9defc7fde: lavfi/histogram: make waveform mode more useful
[17:29] <cone-273> ffmpeg.git 03Paul B Mahol 07master:faac85005cad: lavu/internal: rename macro E(x) to EXPAND(x)
[18:02] <ubitux> hey beastd 
[18:02] <ubitux> how is the subtitles thing going on?
[18:02] <beastd> hi ubitux
[18:02] <ubitux> you have 3 days left
[18:02] <ubitux> ;)
[18:03] <beastd> yeah, unfortunately it is not going so well. but that is not a reason to not enjoy your remaining 3 days on another planet ;)
[18:03] <ubitux> alright, then i'm going to sleep
[18:03] <ubitux> 'night all :)
[18:03] <beastd> n8, see you
[18:16] <cone-273> ffmpeg.git 03Carl Eugen Hoyos 07master:e149127bb217: Call the common 7.1 wide channel layout "7.1(wide)".
[18:55] <durandal_1707> michaelni: what happened to dificulty tag for trac/
[19:03] <michaelni> durandal_1707, could be added but iam not sure its that usefull, open tickets with smaller numbers are harder
[19:05] <durandal_1707> michaelni: not always, but yes ...
[19:12] <durandal_1707> can lavfi open file and write to it and in another run another filter read file?
[19:21] <durandal_1707> did performance dropped with dsputil split?
[19:22] <nevcairiel> the functions just get moved to different files and different pointers, why would it get slower?
[19:29] <cone-273> ffmpeg.git 03Michael Niedermayer 07master:4401958fdc9a: motionpixels: Propagate errors in vlc table init
[19:29] <cone-273> ffmpeg.git 03Michael Niedermayer 07master:b5fc95e77f9e: motionpixels: Check that the vlc table has been fully inited
[20:05] <durandal_1707> michaelni: what kind of data rv10 encodes? major/minor/micro/ version...
[20:17] <michaelni> durandal_1707, dunno you would have to compare it with the decoder if theres no comment
[20:18] <durandal_1707> michaelni: thing is rv10 encoders does not create extradata
[20:18] <durandal_1707> and i stops here with invalid slice_count
[20:18] <durandal_1707> and nothing sets avctx->slice_count
[20:20] <durandal_1707> matroska expects extradata, and decoder too. the last one needs only last 8 bytes
[20:22] <durandal_1707> actually i do same thing as rm muxer
[20:24] <durandal_1707> which i think is hack because ignores extradata in case of -c copy i think
[20:24] <durandal_1707> ^the rm muxer
[20:35] <cone-273> ffmpeg.git 03Paul B Mahol 07master:350128b28fe1: rv10: remove check that is always true
[20:39] <durandal_1707> the rm muxers adds some extra stuff to each video packet
[20:49] <durandal_1707> i think that rv10/rv20 encoder & rm muxer are just wrong
[20:50] <durandal_1707> eg muxer does encoder job
[21:18] <cone-273> ffmpeg.git 03Nicolas George 07master:f43d09cd60a1: lavf: add tee pseudo-muxer.
[21:20] <durandal_1707> michaelni: why ./ffmpeg -f lavfi -i testsrc -vf format=gray16 /tmp/out.bmp picks monob?
[21:22] <cone-273> ffmpeg.git 03Michael Niedermayer 07master:796012af6c78: targa: Fix y check in advance_line
[21:53] <durandal_1707> saste: i fixed il to work with >8 bpp pix fmts
[21:53] <durandal_1707> also mp one crash with pal8
[21:56] <durandal_1707> so i will apply it soon and remove mp once
[21:57] <durandal_1707> *one
[22:48] <durandal_1707> michaelni: fine to remove ilpack?
[22:51] <michaelni> i could only state my oppinon, you still should probably ask on the ML
[22:52] <durandal_1707> michaelni: i asked for you opinion?
[22:52] <michaelni> i suspect its ok but iam not sure
[22:53] <durandal_1707> it is some hybrid scaler and deinterlacer
[22:53] <durandal_1707> *reinterlacer
[22:56] <cone-273> ffmpeg.git 03Paul B Mahol 07master:6e9abce5e2d2: lavfi: port MP il filter
[22:56] <cone-273> ffmpeg.git 03Paul B Mahol 07master:c0a8918f2af7: lavfi: remove mp=il filter
[23:00] <cone-273> ffmpeg.git 03Paul B Mahol 07master:969039eb6e10: lavfi/il: fix typo in Copyright
[00:00] --- Sun Feb 10 2013