[Ffmpeg-devel-irc] ffmpeg-devel.log.20140710

[burek]

[00:06] <cone-730> ffmpeg.git 03Vittorio Giovara 07master:a54f03bf07da: display: add matrix flip api
[00:06] <cone-730> ffmpeg.git 03Michael Niedermayer 07master:c67690f85445: Merge commit 'a54f03bf07da964a1b04b03b85bc39deba76efa4'
[00:12] <cone-730> ffmpeg.git 03Vittorio Giovara 07master:18e3d61e9e3b: h264: parse display orientation SEI message
[00:12] <cone-730> ffmpeg.git 03Michael Niedermayer 07master:cd0dc88751db: Merge commit '18e3d61e9e3b52c177aa7a1f2a054a8a753e1b09'
[00:25] <cone-730> ffmpeg.git 03Vittorio Giovara 07master:0569a7e0bd20: hevc: parse display orientation SEI message
[00:25] <cone-730> ffmpeg.git 03Michael Niedermayer 07master:efc66d69bfb6: Merge commit '0569a7e0bd2006d9a5248d17a1f4bf3ca654ae50'
[00:25] <cone-730> ffmpeg.git 03Michael Niedermayer 07master:92c29914de3d: avcodec/hevc_sei: factor return 1 out of if/else chain
[00:31] <cone-730> ffmpeg.git 03Nicolas Martyanoff 07master:6cc1fec41263: avformat/hlsenc: correctly compute target duration
[00:52] <j-b> michaelni: FYI, libav has a regression with decoding of this sample: http://root.suumitsu.eu/wiki/upload/test2-quantum-ripples-vp9-opus.mkv
[00:52] <j-b> michaelni: I have not tested yet with FFmpeg
[01:15] <michaelni> j-b, regression ? does it not play at all ? play with artifacts ?
[01:15] <michaelni> iam asking because i dont know how its supposed to sound
[01:19] <j-b> plays with artifacts after a few seconds
[01:21] <michaelni> do you have some refernce on how its supposed to sound / look ?
[01:22] <michaelni> because i cant tell if its artifacts or supposed to sound that way ;)
[01:26] <j-b> michaelni: it should work fine until sec 3
[01:26] <j-b> then, it's totally garbled
[01:26] <j-b> if you see women and alse after second 3, then you're good
[01:28] <michaelni> vissually id say there are no (major) artifacts,
[01:29] <michaelni> in case of the audio its quite possibly its supposed to sound like it dose, i cant tell
[01:30] <michaelni> as i dont know the series / havt heard it before
[01:32] <cone-730> ffmpeg.git 03Diego Biurrun 07master:f46bb608d9d7: dsputil: Split off pixel block routines into their own context
[01:32] <cone-730> ffmpeg.git 03Michael Niedermayer 07master:2d5e9451de3c: Merge commit 'f46bb608d9d76c543e4929dc8cffe36b84bd789e'
[02:05] <cone-730> ffmpeg.git 03Michael Niedermayer 07master:c38b5939040c: configure: add new dependencies for spp & mpdecimate
[02:21] <BBB> doesnt libav have a ten-year old copy of our vp9 decoder?
[02:21] <BBB> with their own refactoring, of course
[02:21] <BBB> very important refactoring
[02:23] Action: Compn checks
[02:24] <Compn> http://git.libav.org/?p=libav.git;a=history;f=libavcodec/vp9.c;h=2a6a138aa1c537cd6d0dc01d48c6065e0bc4582e;hb=HEAD
[02:25] <BBB> yeah weve had some bugfixes since then
[02:25] <BBB> I guess they didnt
[02:25] <Compn> you arent going to submit a patch there ? :P
[02:26] <Compn> whoa our vp9.c is much larger
[02:26] <Compn> ah it was split up
[02:28] <j-b> michaelni: BBB: I expected it to work for you. I was just mentionning it, just in case.
[02:28] <j-b> it works for me with an old ffmpeg
[02:28] <j-b> (2.2.4)
[02:28] <BBB> hmk
[02:28] Action: BBB goes back to sobbing
[02:28] <kierank> "But I consider my real name piracy."
[02:28] <kierank> wtf
[02:28] <j-b> BBB: sobbing?
[02:29] <BBB> we_lost :(
[02:29] <j-b> lol
[02:29] <j-b> BananaGuy was too strong
[02:33] <BBB> sad :(
[02:40] <cone-730> ffmpeg.git 03Lukasz Marek 07master:b7cfd5c29b6a: gitignore: add examples/transcoding
[02:40] <cone-730> ffmpeg.git 03Michael Niedermayer 07master:81538470c834: configure: fix select vs. deps
[02:42] <jamrial> who lost what?
[02:45] <kierank> aren't you from argentina?
[02:46] <jamrial> oh, that
[02:46] <jamrial> i thought BBB was from the USA
[02:46] Action: kierank can't tell if jamrial is trolling BBB and other dutchies
[02:50] <j-b> then meta-troll is successful
[02:51] <Compn> soccer.
[02:51] Action: Compn from usa
[02:52] <Compn> or did you guys give up being trolled by the word soccer ? that was fun for a while
[02:52] <BBB> jamrial: I live in USA, but I am originally dutvh
[02:52] <BBB> I support dutch when they play football
[02:53] <jamrial> ah, i see
[02:53] <jamrial> i remember you mentioning the former but never the latter
[02:53] <jamrial> wasn't my intention to troll or anything
[02:54] <BBB> its ok
[02:58] <Compn> maybe Timothy_Gu should put what country everyone is from on this ffmpeg dox  sheet of his
[02:58] <Compn> the other day i called a guy russian, turned out he was from romania :(
[03:18] <Timothy_Gu> Compn: I'll take that into account :) BTW which state are you in? (If you are willing to answer)
[03:21] <Timothy_Gu> ubitux: michaelni: personally I consider release note a must-have, and it is an aspect Libav is doing a much better job (sad siretart isn't with us anymore). For pkg-config stuff, I say we delay this change until after this release, then we can do a vote or something
[04:06] <cone-730> ffmpeg.git 03James Almer 07master:2571e8fe372b: alpha/pixblockdsp: move code out of dsputil
[04:07] <michaelni> Timothy_Gu, release notes would be nice to have, yes, do you volunteer to write some ?
[05:52] <Compn> Argument '1.avi' provided as input filename, but '1.avi' was already specified.
[05:53] <Compn> michaelni : how come we cant use the same input file twice on a command line ?
[05:53] <Compn> what if i want to stitch the same file together in a tile / mosiac ?
[05:54] <michaelni> Compn, just tried "./ffmpeg -i matrixbench_mpeg2.mpg  -i matrixbench_mpeg2.mpg" i see no error
[05:57] <jamrial> Compn: you can use the -map function for that as well
[05:58] <Compn> oh i'm doing something wrong here
[05:59] <Compn> michaelni : with ffplay ?
[05:59] Action: Compn has old ffplay
[05:59] Action: Compn updates
[06:01] <michaelni> ffplay supports only 1 file atm, i suggest you troll cus about that ;)
[06:02] <michaelni> you could inject more input files throug the filter graph and source fiters but that would be a hack
[06:02] <Compn> yes
[06:02] <Compn> i'm trying to use the ffmpeg mosiac filter instructions with ffplay :)
[06:03] <Compn> https://trac.ffmpeg.org/wiki/Create%20a%20mosaic%20out%20of%20several%20input%20videos
[06:03] <Compn> i see you said the other source filter . i've seen that workaround somewhere else hmm
[06:06] Action: Compn digs
[06:09] Action: michaelni falls asleep
[06:10] Action: Compn gone
[06:10] <Compn> night
[06:11] <Compn> https://www.ffmpeg.org/libavfilter.html
[06:12] <Compn> i still dont understand what good this page is , or pages like it ^
[06:12] <Compn> no useful info at all
[06:14] <Compn> shows up at the top of my google searches :\
[06:19] <Compn> never understood why ffmpeg and ffplay engines were seperate in the first place :D
[06:19] <Compn> like mencoder and mplayer i guess
[07:15] <Timothy_Gu> michaelni: probably this weekend. I'm super busy these 3 days.
[08:16] <wm4> michaelni: when is feature freeze?
[08:31] <cone-793> ffmpeg.git 03Carl Eugen Hoyos 07master:a1e3ea475ac3: Fix compilation with --disable-everything.
[10:16] <ubitux> http://mazwai.com/ some free stock videos, might interest some ppl here
[11:52] <cone-401> ffmpeg.git 03Dave Rice 07master:a15c5c44e262: doc/filters: add a signalstats example
[11:55] <ubitux> Compn: ffmpeg-filters link entry is what you are looking for and is present before this link.
[11:55] <ubitux> Compn: we can probably add "You are probably looking for ffmpeg-filters." in the doc, patch welcome
[11:56] <ubitux> ffmpeg & ffplay "engines" are separate because it's not the same usage at all; you can do a "playback" with ffmpeg with output devices but that's not intuitive, and a/v sync & friends are not done properly afaik
[12:21] <anshul> I encontered a mp4/mov video where single symbol contain multiple frame, do in need to devide delta(stts) in both frame?
[12:53] <michaelni> wm4, what would be the point of a "feature freeze" ? (a new feature like decode/demuxer/filter should not be able to break anything)
[12:59] <michaelni> i think simply a "please dont break the release by pushing high risk changes" makes more sense but even that, i think its common sense, no need to say
[13:02] <funman> feature freeze sounds like some formal process
[13:03] <funman> VLC feature freezes tend to not get respected anyway because of common sense
[13:37] <cone-401> ffmpeg.git 03Alexander V. Lukyanov 07master:5adcef9c1bf7: mpegts: pass MpegTSContext ptr explicitly
[13:37] <cone-401> ffmpeg.git 03Michael Niedermayer 07master:5682a93841be: Merge commit '5adcef9c1bf701ba2dd43363ae983ba6d74bdb9a'
[13:43] Action: BBB with michealni on this one
[13:43] <BBB> feature freezes are crazy bureaucratic monsters
[13:49] <Daemon404> fork release branch -> profit
[13:49] <Daemon404> er
[13:49] <Daemon404> not fork
[13:49] <Daemon404> create*
[14:21] <ubitux> lol @ testsrc on our twitter
[14:21] <ubitux> never realized that
[14:26] <J_Darnley> ffmpeg has one of those?
[14:26] <ubitux> one of what?
[14:27] <J_Darnley> a twitter account
[14:27] <ubitux> twitter.com/ffmpeg
[14:29] <cone-401> ffmpeg.git 03Yu Xiaolei 07master:64fdcf24289e: build: rawvideo decoder depends on bswapdsp
[14:37] <cone-401> ffmpeg.git 03Christophe Gisquet 07master:2e471e4703ed: hevc: simplify rounding
[14:47] <cone-401> ffmpeg.git 03Christophe Gisquet 07master:eca1957c4cbe: hevc: reorder loops
[15:32] <cone-401> ffmpeg.git 03Clément BSsch 07master:7e7168b82d0c: Fix 2 coeffecient typo
[15:56] <cone-401> ffmpeg.git 03Christophe Gisquet 07master:bbeaae96eb6f: hevc: derive partially amvp list
[17:12] <j-b> good morning
[17:28] <michaelni> good evening
[17:28] <michaelni> or afternoon rather
[17:38] <cone-401> ffmpeg.git 03Christophe Gisquet 07master:4576eff05d94: hevc: derive partial merge list
[20:09] <kierank> kurosu_: nice patches
[21:12] <cone-401> ffmpeg.git 03James Almer 07master:ddea3b7106b1: x86/yadif-10: remove duplicate ABS macro
[22:32] <mraulet> kurosu_, michaelni patchsets 1 to 4 https://ffmpeg.org/pipermail/ffmpeg-devel/2014-July/159660.html will not work with thread_type=slice on bitstreams having WPP or tiles features
[23:44] <ubitux> http://db0.galo.pe/ffmpeg-web/ so this is the new website from db0 
[23:45] <ubitux> the documentation still needs its integration, but the other pages are done
[23:53] <jamrial> if we're going to feature a big idiot-proof "Download" button, i think it would be better if it doesn't link to a source code snapshot
[23:54] <db0> I didn't know what was the best option
[23:54] <db0> maybe find out the user's os using js and change the link to the corresponding package
[23:55] <ubitux> i guess latest stable would be ok
[23:55] <ubitux> latest stable tarball
[23:55] <jamrial> db0: that's a good idea, yes
[23:55] <jamrial> assuming we host windows binaries, that is
[23:56] <db0> but then what happens if they use android or ios
[23:56] <ubitux> redirect them to disney.com
[23:56] <jamrial> lol
[23:56] <db0> anyway I like ubitux option, it's easier for me
[23:57] <J_Darnley> You're using the wrong logo on that site
[23:57] <db0> the previous one obviously
[23:57] <db0> I know :/ but the 3d one just doesn't fit at all
[23:58] <J_Darnley> Too bad for your design then.  It is trademarked and it went to Libav.
[23:58] <db0> I know
[23:58] <db0> ubitux told me to make a new one
[23:58] <J_Darnley> okay
[23:58] <ubitux> and fast enough so you don't get sued
[00:00] --- Fri Jul 11 2014