[Ffmpeg-devel-irc] ffmpeg-devel.log.20120705

[burek]

[00:03] <saste> anyone want to mentor socis 2012?
[00:03] <saste> i'll ping the mail tomorrow, hopefully will be also able to fill the wiki, but fell free to beat me on the time
[00:10] <ubitux> not me, sorry
[00:13] <philipl> ubitux: Thanks for the review. I've replied.
[00:14] <philipl> Do you specifically want me to change any of the naming? I've explained why it's ended up the way it is in my reply.
[00:39] <durandal_1707> where I can look at fate coverage?
[00:52] <durandal_1707> what ffmpeg/libav/x264 have to do with space?
[00:54] <iive> durandal_1707: i guess nasa likes to work on image compression methods.
[00:55] <durandal_1707> and what happened with 2011 work?
[00:56] <saste> durandal_1707: do you mean socis 2011?
[00:56] <durandal_1707> yea, esa summer of code in space
[00:57] <saste> https://github.com/kuehnelth/libav/commits/geotiff
[00:57] <saste> the last three patches
[00:57] <saste> the avprobe one was partially integrated
[00:57] <saste> the big tiff was reviewed by me and approved on libav-devel, iirc
[00:58] <saste> now it will need to be rebased
[00:58] <saste> as for the metadata in AVFrame that's pretty interesting
[00:58] <durandal_1707> and what could be made for this year (anything that have something to do with space ...)
[00:58] <saste> if i'll find the time this weekend i should send an updated patchset for review
[00:59] <saste> but again, i'd be happy if someone wants to take over the task
[01:00] <saste> did you read the past year proposals?
[01:00] <saste> related to space... not very much
[01:01] <saste> i believe in general improving robustness/reliability/security may be somewhat useful for critical missions
[01:01] <durandal_1707> i think if there is nothing related to space ffmpeg may not be picked
[01:02] <saste> also don't forget that esa is funded by broadcasting companies (think about telecom satellites), so they have some collateral interest in improving multimedia software
[01:02] <saste> durandal_1707: i don't mind if we won't be picked, but we can give a try
[01:02] <saste> in the best case we'll get a single slot
[01:03] <durandal_1707> well improving native exr decoder could be nice small taks
[01:04] <durandal_1707> s/taks/task
[01:07] <durandal_1707> well adding any support for FITS image may be nice task too
[01:09] <durandal_1707> like adding float pixel formats, etc....
[01:12] <durandal_1707> michaelni: do you have any new comments for paf? (I did not get any reply)
[01:16] <onexserie04> hi any here can help me to stream to a flash media server with this rtmp data ( Example ) rtmp://bla.testurl.com/multicast/streamname?fms.multicast.type=3&fms.multicast.groupspec=G%3A010121058d969eef6ecad3c29a3a629280e686cf0c3f5d5a86aff3ca12020c923adc6c92010c160e666d732e6d756c7469636173742e6578616d706c65210ead30f62ecaa35c20cd9adec44d1a95705b0bac051cd314765f126957aeb13377000715313233343536
[01:42] <CIA-41> ffmpeg: 03Paul B Mahol 07master * r3b9dd906d1 10ffmpeg/libavcodec/sanm.c: 
[01:42] <CIA-41> ffmpeg: sanm: replace conditions in for loops with more common version
[01:42] <CIA-41> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[01:42] <CIA-41> ffmpeg: 03Paul B Mahol 07master * r1f9bc3104a 10ffmpeg/libavcodec/sanm.c: 
[01:42] <CIA-41> ffmpeg: sanm: remove code duplication
[01:42] <CIA-41> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[01:42] <michaelni> durandal_1707, paf looked ok already last time except the points i raised
[01:42] <michaelni> so it should be ok now
[01:44] <durandal_1707> wow, it that working 14 bit h264 support?
[01:45] <michaelni> half of the reference files decode without vissible artifacts the other hafl well doesnt yet
[01:49] <durandal_1707> what, there are higher bit depths?
[01:50] <michaelni> some files dont decode correctly while i do not know yet why
[01:50] <Skyler> Link to the patch?  I could try to look at it.
[01:51] <Skyler> Sorry, need to resubcribe to the mailing list at some point.
[01:51] <durandal_1707> Skyler: archive do not have it already?
[01:52] Action: Skyler looks.
[01:53] <Skyler> pixeltmp could probably use a better name; hpel_intermediate maybe?
[01:53] <Skyler> (well, shorter than that I guess.)
[01:54] <Skyler> I remember 10-bit (but not 9-bit?) had some workarounds in the hpel code to offset the values to fit in 16-bit, then de-offset afterwards.  If so they should probably be turned off if the intermediate is >16-bit
[01:54] <Skyler> Maybe running IOC might find the source of the artifacts.
[01:56] <michaelni> i should run it through asan too
[01:56] <Skyler> Oh, there's 4 more patches.  I missed those, *look*
[01:57] <michaelni> and yes i wanted to throw the 10bit woraround out, i forgot :)
[01:57] <Skyler> For bit depth > 14 we should probably actually error (instead of patch welcome), since that's just invalid.
[01:57] <michaelni> workaround is gone locally
[01:58] <Skyler> Yay
[01:58] <michaelni> it patch welcomes on 11 and 13 bit too
[01:59] <Skyler> Yeah, but I mean, >14-bit should get a separate error message.
[01:59] <Skyler> Since "11 or 13 bit" is valid, but not supported, while >14-bit is invalid outright.
[01:59] <michaelni> ok ok
[01:59] <Skyler> Sorry, didn't mean to be overbearing
[02:02] <michaelni> LOL
[02:02] <michaelni> no its fine
[02:03] Action: durandal_1707 lol ..... https://www.ohloh.net/p/ffmpeg
[02:05] <Skyler> michaelni: is 12-bit okay and 14-bit not, or are both wrong sometimes?
[02:08] <michaelni> i have working and non working 14bit files
[02:10] <Skyler> is 12 always okay?
[02:15] <michaelni> it seems all the files i downloaded are 14bit or 10bit :/
[02:16] <michaelni> that was "Rec._ITU-T_H.264.1\(2010-04\)_Professional_profiles_bitstreams"
[02:19] <Skyler> Maybe try encoding with JM and see if that works?
[02:20] <Skyler> I don't see a patch that touches the loopfilter, does that need patching?
[02:21] <Skyler> Ah, that's in h264dsp.
[02:50] <durandal_1707> michaelni: adts_aac_probe looks ugly
[03:13] <durandal_1707> this F/S/B output is bogus, for example codec decoder may support slice threading and codec encoder may not....
[04:27] <CIA-41> ffmpeg: 03Carl Eugen Hoyos 07master * r73bb4e9ca6 10ffmpeg/ (7 files in 3 dirs): 
[04:27] <CIA-41> ffmpeg: nut: add 12/14 bit yuv to nut/raw
[04:27] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[04:45] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * ra32032b508 10ffmpeg/ (3 files in 2 dirs): 
[04:45] <CIA-41> ffmpeg: sws/x86: add some forgotten 12bit planar yuv cases
[04:45] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[04:45] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r9b2b57681c 10ffmpeg/libswscale/x86/swscale_mmx.c: 
[04:45] <CIA-41> ffmpeg: sws/x86: add av_assert() to ensure 16bit code is used with 16bit data
[04:45] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[04:56] <CIA-41> ffmpeg: 03Paul B Mahol 07master * r7de4a16508 10ffmpeg/ (12 files in 4 dirs): 
[04:56] <CIA-41> ffmpeg: PAF demuxer and decoder
[04:56] <CIA-41> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[06:47] <BoRiS> Hi guys
[06:48] <BoRiS> Are their any simple C++ examples of using ffmpeg to transcode one file to another in real time?
[08:26] <CIA-41> ffmpeg: 03Carl Eugen Hoyos 07master * rb5baad3a4c 10ffmpeg/libavutil/Makefile: 
[08:26] <CIA-41> ffmpeg: Install libavutil/version.h
[08:26] <CIA-41> ffmpeg: Fixes ticket #1513
[10:12] <ubitux> http://lwn.net/Articles/502183/
[11:17] <CIA-41> ffmpeg: 03Alexander Strasser 07master * r5fc74f022e 10ffmpeg/doc/fate.texi: doc/fate: Add email address for contacting the FATE admin
[11:17] <CIA-41> ffmpeg: 03Alexander Strasser 07master * r52a3d57e00 10ffmpeg/doc/fate.texi: 
[11:17] <CIA-41> ffmpeg: doc/fate: Use more commonly used terms
[11:17] <CIA-41> ffmpeg:  It might be easier to understand for some people and is a bit more
[11:17] <CIA-41> ffmpeg: in line with e.g. OpenSSH documentation. The meaning of the text stays
[11:17] <CIA-41> ffmpeg: exactly the same.
[11:17] <CIA-41> ffmpeg: 03Alexander Strasser 07master * r10fa37c573 10ffmpeg/doc/fate.texi: 
[11:17] <CIA-41> ffmpeg: doc/fate: Add missing @file command
[11:17] <CIA-41> ffmpeg:  Make the document more consistent.
[11:17] <CIA-41> ffmpeg: 03Alexander Strasser 07master * rc44cae7ec5 10ffmpeg/doc/fate.texi: doc/fate: Add hint for debugging connection to server
[11:18] <CIA-41> ffmpeg: 03Alexander Strasser 07master * re1613476ff 10ffmpeg/doc/fate.texi: 
[11:18] <CIA-41> ffmpeg: doc/fate: Rework examples to a section of its own
[11:18] <CIA-41> ffmpeg:  Make it easier to spot the example. This way it can already be
[11:18] <CIA-41> ffmpeg: discovered by scanning the table of contents.
[11:18] <CIA-41> ffmpeg: 03Alexander Strasser 07master * r585bd4922a 10ffmpeg/libavformat/file.c: lavf/file: cosmetic: Remove superfluous ternary operator
[11:18] <CIA-41> ffmpeg: 03Alexander Strasser 07master * r1fe6cb448a 10ffmpeg/libavformat/file.c: 
[11:18] <CIA-41> ffmpeg: lavf/file: cosmetic: Use "not" instead of "equals zero"
[11:18] <CIA-41> ffmpeg:  This is more similar to the "overall" FFmpeg coding style.
[12:22] <CIA-41> ffmpeg: 03Michael Bradshaw 07master * r7544ef7805 10ffmpeg/libavcodec/libopenjpegenc.c: 
[12:22] <CIA-41> ffmpeg: libopenjpegenc: add support for YUV420/422/444P12/14
[12:22] <CIA-41> ffmpeg: Signed-off-by: Michael Bradshaw <mbradshaw at sorensonmedia.com>
[12:22] <CIA-41> ffmpeg: 03Michael Bradshaw 07master * r713a7854e0 10ffmpeg/libavcodec/libopenjpegenc.c: 
[12:22] <CIA-41> ffmpeg: libopenjpegenc: cosmetics: reorder pix_fmts
[12:22] <CIA-41> ffmpeg: Signed-off-by: Michael Bradshaw <mbradshaw at sorensonmedia.com>
[12:22] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r4ea705114c 10ffmpeg/: 
[12:22] <CIA-41> ffmpeg: Merge git://github.com/mjbshaw/FFmpeg-OpenJPEG-J2K-Encoder
[12:22] <CIA-41> ffmpeg: * git://github.com/mjbshaw/FFmpeg-OpenJPEG-J2K-Encoder:
[12:22] <CIA-41> ffmpeg:  libopenjpegdec: add support for decoding YUV420/422/444P12/14
[12:22] <CIA-41> ffmpeg:  libopenjpegenc: cosmetics: reorder pix_fmts
[12:22] <CIA-41> ffmpeg:  libopenjpegenc: add support for YUV420/422/444P12/14
[12:22] <CIA-41> ffmpeg: Merged-by: Michael Niedermayer <michaelni at gmx.at>
[12:22] <CIA-41> ffmpeg: 03Michael Bradshaw 07master * rd3b84fdccc 10ffmpeg/libavcodec/libopenjpegdec.c: 
[12:22] <CIA-41> ffmpeg: libopenjpegdec: add support for decoding YUV420/422/444P12/14
[12:22] <CIA-41> ffmpeg: Signed-off-by: Michael Bradshaw <mbradshaw at sorensonmedia.com>
[12:27] <CIA-41> ffmpeg: 03Nedeljko Babic 07master * r6d74e3c6f2 10ffmpeg/libavcodec/ (8 files in 2 dirs): 
[12:27] <CIA-41> ffmpeg: Optimization of MP3 decoders for MIPS
[12:27] <CIA-41> ffmpeg: MP3 fixed and floating point decoders are optimized
[12:27] <CIA-41> ffmpeg:  for MIPS architecture.
[12:27] <CIA-41> ffmpeg: Signed-off-by: Nedeljko Babic <nbabic at mips.com>
[12:27] <CIA-41> ffmpeg: Reviewed-by: Vitor Sessak <vitor1001 at gmail.com>
[12:27] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[13:07] <CIA-41> ffmpeg: 03Georg Lippitsch 07master * r325fb246f2 10ffmpeg/ (9 files in 4 dirs): 
[13:07] <CIA-41> ffmpeg: FireWire DV/HDV input device using libiec61883
[13:07] <CIA-41> ffmpeg: Signed-off-by: Stefano Sabatini <stefasab at gmail.com>
[13:32] <DelphiWorld> hello ffmpegesters
[13:32] <DelphiWorld> michaelni: http://paste.debian.net/177808/
[13:32] <DelphiWorld> michaelni: fail to build latest FFMpeg git
[13:33] <michaelni> DelphiWorld, what configure paramaters where used ?
[13:34] <DelphiWorld> michaelni, http://paste.debian.net/
[13:35] <michaelni> DelphiWorld, no its not on the front page :) pf paste.debian.net
[13:36] <DelphiWorld> michaelni, hold on hahaha
[13:36] <DelphiWorld> michaelni: try http://paste.debian.net/
[13:37] <michaelni> saste, ping, did you just break configure/compile witht that libiec61883 stuff ?
[13:37] <DelphiWorld> LOL
[13:38] <DelphiWorld> michaelni, make on top of my existing Configure: http://paste.debian.net/177809/
[13:45] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r0348102835 10ffmpeg/configure: 
[13:45] <CIA-41> ffmpeg: configure: dont blindly enable iec61883
[13:45] <CIA-41> ffmpeg: This should fix a configure failure
[13:45] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[13:46] <michaelni> saste, DelphiWorld, should be fixed
[13:46] Action: DelphiWorld git pull on top of michaelni ...
[13:48] <DelphiWorld> is fixed michaelni thank you a lot DUDE
[13:49] <DelphiWorld> michaelni: how is austria doing? vienna:-P
[13:51] <michaelni> np
[13:52] <DelphiWorld> michaelni: did you add a configure arg to enable this lib ?
[13:59] <michaelni> ill post a patch for that in a moment
[14:01] <saste> michaelni: doh
[14:01] <saste> wait for a fix this night
[14:01] <saste> i'm not sure about which policy to use for the configure stuff
[14:02] <saste> we can do : enable libraries -> enable device
[14:02] <saste> or
[14:03] <saste> enable device -> check for libraries
[14:03] <ubitux> michaelni: did you see http://lists.mplayerhq.hu/pipermail/ffmpeg-user/2012-July/007715.html ?
[14:04] <ubitux> it might just be a not recognized instruction by valgrind, but it might also be the cause of the issue
[14:06] <saste> michaelni: ah just saw the patch
[14:08] <funman> who draws the logos?
[14:12] <ubitux> "victoria"
[14:12] <ubitux> Victoria Kirst
[14:12] <ubitux> https://ffmpeg.org/trac/ffmpeg/ticket/1418
[14:13] <funman> mustache sun works in french but not in italian as it's a feminine word :)
[14:14] <DelphiWorld> saste: is me that reported it:-P
[14:14] <funman> it's a nice logo ;p
[14:15] <ubitux> yeah it's nice :)
[14:15] <DelphiWorld> hi funman
[14:18] <funman> hello
[14:19] <DelphiWorld> funman: rafael, rockbox?
[14:23] <funman> yeah
[14:23] <funman> like /whois says :)
[14:23] <DelphiWorld> funman: what's rockbowx ?
[14:23] <funman> www.rockbox.org
[14:28] <DelphiWorld> funman: i never heard about those player's except of iPod
[14:29] <funman> Apple must have a strong marketing then
[14:30] <DelphiWorld> funman: ;)
[14:38] <DelphiWorld> funman: silly question please, what codec (A|V) do you recomand for best Streaming experiance on 3G/EDGE?
[14:39] <funman> silly answer: i recommand uncompressed bitmap and pcm
[14:39] <DelphiWorld> please;, funman !
[14:41] <funman> cinepak and sun's .au
[14:45] <DelphiWorld> funman: no idea about those
[15:28] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rc88e60af76 10ffmpeg/libswresample/x86/ (audio_convert.asm swresample_x86.c): 
[15:28] <CIA-41> ffmpeg: swr/x86: 10l, missed some SSE2 instructions in code marked as SSE.
[15:28] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[15:43] <Tjoppen> michaelni: I suggested adding av_calloc() to libav's lavu. they seem to be OK with it as long as it's inline. thoughts?
[15:44] <Tjoppen> (it'd hopefully make mxfdec more similar between the projects)
[15:45] <ubitux> michaelni: thanks for the fix
[15:47] <ubitux> michaelni: btw, what means the 'u' and 'a' in the ASM?
[15:47] <ubitux> the CONV ... [ua] ...
[15:51] <michaelni> ubitux, unaligned and aligned
[15:52] <ubitux> oh, ok, thx
[15:55] <iive> what is the difference between av_calloc and av_mallocz ?
[15:56] <ubitux> overflow check?
[15:56] <Tjoppen> calloc = malloc X of Y
[15:57] <ubitux> + memset 0
[15:57] <Tjoppen> right
[15:57] <Tjoppen> most important being that X*Y doesn't overflow
[15:58] <iive> imho, this check should be done as separate function.
[15:59] <michaelni> i think so too, but iam open to technical arguments
[15:59] <iive> i'm sure there is at least one place where you need (X*Y+W*H).
[16:00] <Tjoppen> in a single alloc?
[16:01] <michaelni> what ive meant is that av_calloc() should not be inlined unless there is a technical reason to inline it
[16:01] <michaelni> not sure what iive meant
[16:01] <ubitux> why do they want to inline it Tjoppen? oO
[16:02] <michaelni> one technical reason may be that the code ends up smaller ...
[16:02] <Tjoppen> ubitux: faster, doesn't clutter up the ABI (I suppose)
[16:02] <ubitux> the call overhead is bigger than the if (size ...) check?
[16:02] <Tjoppen> I don't really care
[16:03] <Tjoppen> it's only called during init
[16:03] <ubitux> av_calloc() is used in various places
[16:03] <Tjoppen> they might be reluctant to do a minor bump. I'm just guessing though
[16:03] <ubitux> av_freep() is not an inline, i don't see any reason why av_calloc() should be one
[16:04] <ubitux> Tjoppen: isn't it because they just want to make things differently?
[16:04] <Tjoppen> they changed all the allocations to x>=INT_MAX/sizeof(foo) + av_malloc(x*sizeof(foo)) instead of av_calloc(x,sizeof(foo)) for some reason
[16:04] <Tjoppen> IIRC because calloc is "useless"
[16:05] <ubitux> :/
[16:05] <ubitux> i like calloc for doing that check
[16:06] <Tjoppen> I do too. also because it makes it easier to impose some upper allocation limit other than INT_MAX
[16:06] <Tjoppen> without having to change every demuxer
[16:06] <iive> there is some 10 old win version with calloc that doesn't check for overflow.
[16:06] <ubitux> the av_calloc() does the check
[16:07] <ubitux> so there are two reasons for having av_calloc() inline? call overhead bigger than if() one, and performance because of successive call?
[16:07] <ubitux> do we really use alloc() function is performances bottlenecks?
[16:08] <ubitux> functions*
[16:08] <ubitux> in*
[16:08] <Tjoppen> I'll run just having it as a normal function past the libav list. maybe it sparkles
[16:08] <Tjoppen> that way the two lavu version become more similar
[16:09] <Tjoppen> (of course saying that would be flamebait)
[16:13] <Tjoppen> hm, av_size_mult()
[16:21] <Tjoppen> proper angle found: smaller code
[16:23] <iive> Tjoppen: is there already av_size_mult ?
[16:24] <Tjoppen> yes, but it isn't used anyway. libav also doesn't have
[16:24] <Tjoppen> it
[16:24] <Tjoppen> s/anyway/anywhere but mem.c/
[16:43] <DelphiWorld> can ffmpeg read dvb streams ?
[16:44] <onexserie04>  hi any here can help me how i can stream to a rtmp server with this option ? rtmp://blabla.com/live/livestream?fms.multicast.type=1&fms.multicast.groupspec=G%3A01012105274123fdc56981c0201cc42d089effb4eb3484b218e7768d60ca801027015625010c160e666d732e6d756c7469636173742e6578616d706c65011b00070ae00000fe75300b1578636173646173646661&fms.multicast.address=224.0.0.254%3A30000
[16:50] <iive> DelphiWorld: these are mpeg-ts. did you mean capture from dvb device?
[16:50] <DelphiWorld> iive, right ;-)
[16:51] <CIA-41> ffmpeg: 03Carl Eugen Hoyos 07master * r50d0f6cee9 10ffmpeg/cmdutils.c: 
[16:51] <CIA-41> ffmpeg: Extend -codecs output.
[16:51] <CIA-41> ffmpeg: Allow to show different supported multi-threading methods for an
[16:51] <CIA-41> ffmpeg: encoder/decoder pair.
[16:51] <CIA-41> ffmpeg: Fixes ticket #1512
[17:55] <ubitux> Author:    Diego Biurrun <diego at biurrun.de>
[17:55] <ubitux> Committer: Diego Biurrun <diego at biurrun.de>
[17:55] <ubitux> x86: h264_intrapred: use newly introduced SPLAT* and PSHUFLW macros
[17:56] <ubitux> @_@
[17:56] <ubitux> git fail?
[18:05] <kierank> what
[18:05] <ubitux> git fail confirmed.
[18:07] <iive> ubitux: what exactly is the fail?
[18:08] <ubitux> do you think Diego is the author of the ASM code?
[18:08] <ubitux> he removed authorship while rebasing
[18:15] <DelphiWorld> iive: answer? ;)
[18:16] <iive> i got the answer I asked for.
[18:21] <DelphiWorld> iive: i didn't see it:-P
[18:22] <iive> man git-fail ;)
[18:22] <DelphiWorld> ha ha ha iive what you're talking about;-)
[18:22] <iive> ubitux> he removed authorship while rebasing
[18:25] <DelphiWorld> hahahaha!*
[18:25] Action: DelphiWorld slaps ubitux around a bit with a large trout
[18:26] <ubitux> look at the fork commits
[18:30] <DelphiWorld> iive: he killed the entire repo ?
[18:32] <iive> DelphiWorld: are you sure you don't have ubitux on ignore? You seem to miss everything he writes.
[18:32] <DelphiWorld> iive: no, i has bean out ;)
[18:32] <DelphiWorld> iive: also i use text to speech, DUDE i am blind:-P
[18:34] <iive> I see. I know there are Braille screens, but I guess they are expensive.
[18:35] <iive> diego seems to have committed a patch in the libav fork repo. he is marked as author of patch, while he haven't written it.
[18:37] <iive> the name seems to have gotten replaced when he rebased the patch.
[18:38] <DelphiWorld> :P
[18:38] <DelphiWorld> iive: brail screen is very silly
[18:41] <iive> you should know better than me.
[18:46] <DelphiWorld> :)
[19:21] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * racb52d3228 10ffmpeg/libavcodec/h264_ps.c: 
[19:21] <CIA-41> ffmpeg: h264: print error on unsupported seperate color planes
[19:21] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:41] Action: ubitux is wondering if adding a threads+valgrind box would make sense
[20:44] <CIA-41> ffmpeg: 03Matthieu Bouron 07master * r135e684311 10ffmpeg/libavformat/mxfenc.c: 
[20:44] <CIA-41> ffmpeg: mxfenc: support smpte dv codec
[20:44] <CIA-41> ffmpeg: After some internal talks it seems the code is similar to what is in FFmbc
[20:44] <CIA-41> ffmpeg: by Baptiste Coudurier; Baptiste accepted to relicense the similiar chunks
[20:44] <CIA-41> ffmpeg: from GPL to LGPL.
[20:44] <CIA-41> ffmpeg: 03Matthieu Bouron 07master * rd99001cab3 10ffmpeg/libavformat/mxfenc.c: mxfenc: support dnxhd codec
[20:44] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r92c7ef1e30 10ffmpeg/: 
[20:44] <CIA-41> ffmpeg: Merge branch 'mxfenc_improvements' of https://github.com/mbouron/FFmpeg
[20:45] <CIA-41> ffmpeg: * 'mxfenc_improvements' of https://github.com/mbouron/FFmpeg:
[20:45] <CIA-41> ffmpeg:  mxfenc: support dnxhd codec
[20:45] <CIA-41> ffmpeg:  mxfenc: support smpte dv codec
[20:45] <CIA-41> ffmpeg: Merged-by: Michael Niedermayer <michaelni at gmx.at>
[21:58] <CIA-41> ffmpeg: 03Martin Storsjö 07master * r07eeeb1d4f 10ffmpeg/libavcodec/x86/vp8dsp-init.c: 
[21:58] <CIA-41> ffmpeg: vp8: Add ifdef guards around the sse2 loopfilter in the sse2slow branch too
[21:58] <CIA-41> ffmpeg: This was missed in the the previous commit in 70a1c800.
[21:58] <CIA-41> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[21:58] <CIA-41> ffmpeg: 03Martin Storsjö 07master * r715129cdc4 10ffmpeg/avconv.c: (log message trimmed)
[21:58] <CIA-41> ffmpeg: avconv: Set audio filter time base to the sample rate
[21:58] <CIA-41> ffmpeg: If the output frame size is smaller than the input sample rate,
[21:58] <CIA-41> ffmpeg: and the input stream time base corresponds exactly to the input
[21:58] <CIA-41> ffmpeg: frame size (getting input packet timestamps like 0, 1, 2, 3, 4 etc),
[21:58] <CIA-41> ffmpeg: the output timestamps from the filter will be like
[21:58] <CIA-41> ffmpeg: 0, 1, 2, 3, 4, 4, 5 ..., leadning to non-monotone timestamps later.
[21:58] <CIA-41> ffmpeg: 03Luca Barbato 07master * r669bbedfa8 10ffmpeg/libavutil/blowfish.c: 
[21:58] <CIA-41> ffmpeg: blowfish: invert branch and loop precedence
[21:58] <CIA-41> ffmpeg: Should slightly improve performance depending on the compiler used.
[21:58] <CIA-41> ffmpeg: 03Luca Barbato 07master * r5b54a90c8b 10ffmpeg/libavformat/flvdec.c: 
[21:58] <CIA-41> ffmpeg: compilation of h264.c and reduces the code size without any
[21:58] <CIA-41> (21 lines omitted)
[21:59] <CIA-41> ffmpeg: 03Samuel Pitoiset 07master * re4a7fb3da3 10ffmpeg/libavutil/blowfish.c: 
[21:59] <CIA-41> ffmpeg: blowfish: Make the count parameter match the documentation
[21:59] <CIA-41> ffmpeg: Previously it was interpreted as number of bytes, while the
[21:59] <CIA-41> ffmpeg: documentation stated that it was the number of 8 byte blocks.
[21:59] <CIA-41> ffmpeg: This makes it behave similarly to the existing AES code.
[21:59] <CIA-41> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[21:59] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r24823a761c 10ffmpeg/: (log message trimmed)
[21:59] <CIA-41> ffmpeg: Merge remote-tracking branch 'qatar/master'
[21:59] <CIA-41> ffmpeg: * qatar/master:
[21:59] <CIA-41> ffmpeg:  qdm2: remove broken and disabled dump_context() debug function
[21:59] <CIA-41> ffmpeg:  x86: h264_intrapred: use newly introduced SPLAT* and PSHUFLW macros
[21:59] <CIA-41> ffmpeg:  x86inc: add SPLATB_LOAD, SPLATB_REG, PSHUFLW macros
[21:59] <CIA-41> ffmpeg:  x86inc: modify ALIGN to not generate long nops on i586
[22:22] <michaelni> ubitux, dunno, maybe
[22:22] <michaelni> if it turns out useless we can remove it later
[22:23] <michaelni> also depends on how much cpu cycles you have left
[22:26] <ubitux> enough, that box doesn't seem to go down randomly
[22:26] <ubitux> it seems german ppl are smarter than frenchies ;)
[22:27] <ubitux> i'll add one
[23:02] <ubitux> saste: ping
[23:05] <ubitux> :)
[23:05] <ubitux> what are we going to do about the init callback for the filters?
[23:05] <ubitux> do we keep the opaque field or we drop it?
[23:05] <ubitux> there are warning in all our filters
[23:06] <ubitux> (mismatiching prototypes for the init callback)
[00:00] --- Fri Jul  6 2012