[Ffmpeg-devel-irc] ffmpeg-devel.log.20150305

[burek]

[00:26] <cone-531> ffmpeg 03wm4 07master:cca44764b9de: avformat: add avformat_flush()
[00:26] <cone-531> ffmpeg 03Reimar Döffinger 07master:7c2e35d1d0f1: avformat/avformat: improve the text about headerless formats in avformat_flush()
[01:20] <cone-531> ffmpeg 03Martin Storsjö 07master:15621cbbe4ed: rtpdec_asf: Don't set RTP_FLAG_KEY
[01:20] <cone-531> ffmpeg 03Michael Niedermayer 07master:9575d8a652c0: Merge commit '15621cbbe4ed88409ef40257afaf239487964e98'
[01:39] <cone-531> ffmpeg 03Martin Storsjö 07master:c13ab42a0a39: rtpdec_qt: Use a local variable instead of RTP_FLAG_KEY
[01:39] <cone-531> ffmpeg 03Michael Niedermayer 07master:2eb890aab22d: Merge commit 'c13ab42a0a3979b2c292b2315487f5f5e14ba26b'
[01:49] <selsta_> does this look right? http://sprunge.us/BIEU (test url with sample-aes encryption: http://demo.unified-streaming.com/video/caminandes/caminandes-sample-aes.ism/caminandes-sample-aes.m3u8)
[01:50] <selsta_> without the patch it would download garbage, with the patch it writes an error message and exits
[02:17] <cone-531> ffmpeg 03Martin Storsjö 07master:b72b212a4c43: rdt: Use a separate variable to clarify the different value namespaces
[02:17] <cone-531> ffmpeg 03Michael Niedermayer 07master:4574f11e4dd3: Merge commit 'b72b212a4c43563f1b9fc3ce9a5ff91f89b857ac'
[02:47] <cone-531> ffmpeg 03Himangi Saraogi 07master:f3b74d944bde: pcxenc: Return more meaningful error codes
[02:47] <cone-531> ffmpeg 03Michael Niedermayer 07master:dcf7924f4df6: Merge commit 'f3b74d944bdef9bbe81f7690f8d837d843ec13a6'
[03:21] <llogan> kierank: BL never replied. I guess that's it.
[08:02] <diginet> does ffmpeg have any support for "exotic" color spaces? (i.e. things besides RGB or YCbCr)
[08:04] <diginet> I'm trying to experiment with a custom codec, but I want to use YCgCo instead
[08:13] <funman> diginet: try ffmpeg -pix_fmts
[08:19] <diginet> hmm, apparently there have been a few patches to support it
[08:20] <diginet> however, it looks like I'd have to specifically add support for my version I guess
[10:23] <himangi> hey! When running : ffmpeg  -i in.mp4 -vf telecine=top out.mp4, I get the error "The encoder 'aac' is experimental but experimental codecs are not enabled, add '-strict -2' if you want to use it."
[10:24] <himangi> Adding -strict -2 does not help
[10:28] <cone-620> ffmpeg 03Thomas Volkert 07master:9c09fbd80532: rtpdec: experimental VP9 depacketizer (draft 0)
[10:28] <cone-620> ffmpeg 03Michael Niedermayer 07master:63ea3a86b7ce: Merge commit '9c09fbd8053292267cdc4b542913ceced43ea3a9'
[10:44] <ubitux> JEEBsv: http://permalink.gmane.org/gmane.comp.internationalization.fribidi/572
[10:44] <ubitux> that's you?
[10:44] <JEEB> yes
[10:44] <ubitux> JEEB: did you find a solution? :P
[10:45] <JEEB> got told to make an issue on their tracker instead, and never did that
[10:45] <JEEB> the dev responded at least
[10:45] <ubitux> got exactly the same issue, it's fucking annoying
[10:45] <ubitux> gonna get the tarball and apply the 3 patches..
[10:46] <wm4> fribidi's build system is fucked
[10:46] <wm4> shame because it's such a small library
[10:46] <JEEB> http://permalink.gmane.org/gmane.comp.internationalization.fribidi/576
[10:46] <wm4> also, wow, newsgroups still exist?
[10:47] <JEEB> yeah
[10:47] <ubitux> yeah it's completely fucked
[10:47] <JEEB> also ugh, from bar to flight
[10:47] <ubitux> it doesn't even support jobbed make
[10:47] <JEEB> this is going to end sooo goood
[10:47] <cone-620> ffmpeg 03Timo Rothenpieler 07master:f5668e614bdc: avformat/dashenc: Update extradata for mov muxer
[10:47] <ubitux> i don't understand how much build shit there is in this project compared to the 3 random sources files
[11:01] <kurosu> ah, the joy of fribidi... has development picked up steam or is it mostly Behdad Esfahbod by himself ?
[11:02] <JEEB> same
[11:50] <ubitux> JEEB: fribidi features are not going to be merged into freetype/harfbuzz stack?
[11:50] <ubitux> (btw, i noticed there is a cross dep between harfbuzz and freetype...)
[11:52] Action: ubitux wonders if libav is going to NIH the pulse output
[11:53] <nevcairiel> probably
[11:53] <nevcairiel> we dont exist, you know
[11:55] <cone-620> ffmpeg 03Vilius Grigaliknas 07master:2af260e3a85e: avformat/mxfdec: Detect XYZ pixel format for digital cinema files
[12:14] <cone-620> ffmpeg 03Stefano Sabatini 07master:6db20926c32e: doc/protocols/tcp: fix units of listen_timeout option value, from microseconds to milliseconds
[12:27] <cone-620> ffmpeg 03Kevin Wheatley 07master:86b59e6a50c8: avutil: Extract gamma determination from PNG encoder for future use. Adds private avpriv_get_gamma_from_trc() function to libavutil.
[12:27] <cone-620> ffmpeg 03Kevin Wheatley 07master:45555a20c6f4: avformat/movenc: Add support for writing 'gama' atom to QuickTime .mov files.
[12:50] <cone-620> ffmpeg 03Nedeljko Babic 07master:dfa920807494: mips/float_dsp: fix a bug in vector_fmul_window_mips
[13:43] <cone-620> ffmpeg 03Michael Niedermayer 07master:d48c8540a2c9: fate-aac-aref-encode: increase FUZZ to 5, needed for ppc64be
[13:57] <cone-620> ffmpeg 03Michael Niedermayer 07master:6d7e57e88b75: avutil/opt: Do not print inf in selftest
[16:02] <cone-620> ffmpeg 03Michael Niedermayer 07master:7ccc5848248b: avfilter/vf_paletteuse: Use int where AVERROR can be returned
[16:07] <cone-620> ffmpeg 03Clément BSsch 07master:1cded1f8d7d2: avfilter/vf_paletteuse: indent fix after 7ccc5848
[17:30] <kierank> excellent use of sexagesimal 
[17:45] <wm4> lol an apng encoder
[17:46] <wm4> because the world needs it
[17:46] <wm4> (mozilla-only mostly-internal format)
[17:54] <michaelni> wm4, if you disagree with some task, best to discuss it privately and suggest a better task which you want to mentor. trolling on IRC just disencourages people, thats not good
[18:09] <selsta> should i resubmit my patch to ffmpeg-devel or is a github pull request fine?
[18:09] <Daemon404> use ffmpeg-devel
[18:09] <poste9> is there a proper format to follow for submiting patch ?
[18:10] <Daemon404> yes
[18:10] <Daemon404> it is all documented
[18:10] <Daemon404> http://ffmpeg.org/developer.html#Submitting-patches-1
[18:10] <selsta> okay i'll try ffmpeg-devel, never used a mailing list before
[18:15] <Daemon404> kierank, but what if it goes over 99 hours
[18:15] Action: Daemon404 runs
[18:17] <wm4> michaelni: I just don't get why of all the things that are to do, something like apng gets exposure
[18:18] <Daemon404> its easy
[18:18] <Daemon404> and doable
[18:18] <Daemon404> is wy
[18:18] <Daemon404> why
[18:19] <aetasx> I don't think I've ever seen an actual apng in the wild
[18:20] <ubitux> there is one on wikipedia
[18:20] <aetasx> is it on the wikipedia page for apng?
[18:21] <aetasx> ;)
[18:21] <wm4> insert overly sarcastic statement
[18:21] <Daemon404> dont forget animated webp
[18:21] <wm4> still has more future than apng...
[18:22] <aetasx> dunno how it is now but all the reviews I've read on webp put it as not offering any measurable improvements, maybe thats changed
[18:22] <Daemon404> it offers some
[18:22] <Daemon404> but no double-digit improvement will supplant jpeg.
[18:23] <aetasx> think one of mplayer's own devs wrote one of those negative reviews I saw actually
[18:23] <Daemon404> mplayer doesnt have developers
[18:23] <aetasx> well there was an mplayer listed somewhere in the title near their name
[18:24] <wm4> remaining mplayer devs who work on ffmpeg too: reimar, iive, nicolas, cehoyos, michaelni
[18:24] <aetasx> oh, no it was an x264 dev
[18:24] <wm4> mplayer is still pretty dead
[18:25] <wm4> aetasx: that was years ago
[18:25] <ubitux> wm4: still maintained though
[18:25] <aetasx> Im aware, like I saw it was one of the reviews I saw
[18:28] <Daemon404> no x264 dev wrote about webp
[18:28] <Daemon404> youre thinking og vp8.
[18:28] <Daemon404> of*
[18:29] <iive> afaik mplayer2 is completely dead, aka no commits of any sort since 2013.
[18:30] <wm4> yes it's dead
[18:30] <Daemon404> well it was one (angry) person
[18:30] <wm4> mplayer2? never seen uau been angry
[18:31] <Daemon404> ok let me rephrase
[18:31] <Daemon404> he was an ashole of the highest order.
[18:31] <Daemon404> asshole even
[18:31] <wm4> just some kind of communication issue
[18:31] Action: ubitux always found uau a pretty awesome and decent developer
[18:31] <iive> it's more an issue of core principles.
[18:32] <ubitux> (awesome technically, decent humanly)
[18:32] <selsta> okay i hope i've done everything right
[18:47] <cone-620> ffmpeg 03Michael Niedermayer 07master:8a048fe6f8bf: avformat/avidec: initialize header array
[18:48] <aetasx> Daemon404: http://x264dev.multimedia.cx/archives/541
[18:49] <Daemon404> oh indeed i forgot
[18:49] <Daemon404> though some of the bits of it are no longer accurate
[18:50] <aetasx> yeah, I mostly just scimmed through it because of its age and went onto the other ones
[18:51] <wm4> that blog was interesting
[18:52] <Daemon404> its mostly true though
[18:52] <Daemon404> its not worth the hassle
[18:52] <aetasx> theres another one by a mozilla dev from 2013
[18:52] <Daemon404> jpeg is good enough
[18:52] <aetasx> https://blog.mozilla.org/research/2013/10/17/studying-lossy-image-compression-efficiency/ this one
[18:52] <wm4> mp3 is good enough too?
[18:52] <wm4> and divx
[18:53] <aetasx> that one is pretty comprehensive as far as testing is concerned
[18:54] <Daemon404> wm4, the only reason mp3 was supplanted was apple
[18:54] <Daemon404> IMP
[18:54] <Daemon404> IMO*
[18:54] <aetasx> supplanted by what?  I rarely see anything outside of mp3s for music these days still
[18:54] <Daemon404> aac
[18:55] <Daemon404> all of itunes is aac.
[18:55] <Daemon404> and all videos obline
[18:55] <Daemon404> online*
[18:55] <aetasx> ah, yeah never bothered with their itunes store
[18:56] <aetasx> anything they seem to do usually has a relatively good chance to annoy me
[18:56] <aetasx> though at least they're consistent with it
[18:56] <jamrial> afaics, outside of itunes and online video, mp3 is still king
[18:56] <jamrial> it's "good enough" and suported by everything under the sun. it's not going anywhere
[18:57] <aetasx> only time I've ever seen an .aac file outside of a movie container is when its being transcoded and then deleted right afterwards
[18:57] <nevcairiel> thats because it usually comes in m4a and not raw aac
[18:58] <aetasx> same for m4a ;)
[18:58] <aetasx> don't typically come across them
[19:00] <aetasx> meant to ask you guys, how is openh264?
[19:02] <kepstin-laptop> aetasx: piece of junk, use x264 if you can.
[19:03] <kepstin-laptop> if you need something covered by cisco's patent license, and don't mind that it can't do high profile, then use it i guess.
[19:03] <aetasx> yeah, that's what I guessed
[19:16] <wm4> yeah, both for encoding and decoding, the only reason to use this cisco thing is apparently that the license is covered
[19:18] <Daemon404> wm4, and only if you download the binary
[19:18] <Daemon404> from their sever
[19:18] <Daemon404> server
[19:18] <wm4> right
[19:18] <wm4> can't compile it yourself
[19:18] <wm4> and combine this with no stable API/ABI (wtf...)
[19:18] <Daemon404> that said, you likely dont need it on any platform
[19:19] <Daemon404> everything has its own decoders
[19:19] <Daemon404> hw even.
[19:19] <Daemon404> maybe some chinese android phones from 2010
[19:39] <cone-620> ffmpeg 03Michael Niedermayer 07master:8c013a9e55d9: avformat/avidec: Check avio_read() return before using the results for meta-data
[19:46] <Compn> aetasx : yes, itunes nonsense requiring the itunes program to use the itunes site and other annoyances is why i never got into itunes either
[19:47] <Compn> so my music is 98% mp3, 1% ogg vorbis (look, it was a bad idea ok?) and maybe a few .mp4 aac i picked up at random
[19:52] <cone-620> ffmpeg 03Andreas Cadhalpun 07master:66624ed6319c: rv10: check size of s->mb_width * s->mb_height
[19:52] <cone-620> ffmpeg 03Michael Niedermayer 07master:31559723ab45: Merge commit '66624ed6319cb1a959256fe1a717fec5b748fbfa'
[19:54] <lionaneesh> durandal_1707: ping.
[19:54] <lionaneesh> Oh, he's not online.
[19:54] <lionaneesh> michaelni: ping.
[19:55] <lionaneesh> Can you please assign me the APNG qualification task?
[19:57] <michaelni> lionaneesh, the listed quali task is already taken by myra see: https://trac.ffmpeg.org/wiki/SponsoringPrograms/GSoC/2015-Qualis
[19:59] <lionaneesh> michaelni: So only one guy is allowed to take it?
[19:59] <lionaneesh> Hey BBB 
[19:59] <BBB> hello
[20:00] <lionaneesh> BBB: long time! :)
[20:00] <BBB> indeed
[20:00] <BBB> gsoc-interested I spose? :)
[20:00] <lionaneesh> BBB: Yeah! :)
[20:01] <lionaneesh> BBB: libav isn't applying, is it?
[20:03] <BBB> I dont know, I dont follow that half anymore tbh
[20:03] <BBB> too much RealLife going on :)
[20:04] <michaelni> lionaneesh, many people can take a task but ideally each student should work on a seperate qualification task
[20:06] <michaelni> lionaneesh, do you have a idea for a qualifiation task for APNG ?
[20:08] <cone-620> ffmpeg 03Andreas Cadhalpun 07master:4f2ee9daeede: webp: validate the distance prefix code
[20:08] <cone-620> ffmpeg 03Michael Niedermayer 07master:6af5bf604c97: Merge commit '4f2ee9daeedec636fb4fe953a1e0990a7102d5eb'
[20:32] <cone-620> ffmpeg 03Andreas Cadhalpun 07master:bb4edddd9389: rmenc: limit packet size
[20:32] <cone-620> ffmpeg 03Michael Niedermayer 07master:be07270bcaae: Merge commit 'bb4edddd9389cc1601db618ed3c1375b62628d04'
[20:45] <poste9> can u tell me where the groups variable of ffmpeg.c is filled?
[20:45] <cone-620> ffmpeg 03Andreas Cadhalpun 07master:c73c3c34fe72: avformat/adxdec: set avctx->channels in adx_read_header
[20:45] <cone-620> ffmpeg 03Michael Niedermayer 07master:777179edb5ad: Merge commit 'c73c3c34fe72bb06966450207917ce902146b172'
[20:46] <poste9> actually its ffmpeg_opt.c
[20:47] <poste9> nvm, found it
[21:04] <cone-620> ffmpeg 03Anton Khirnov 07master:ad786dd450f2: h264: remove a now redundant error check
[21:04] <cone-620> ffmpeg 03Michael Niedermayer 07master:22ad0546c8ab: Merge commit 'ad786dd450f26ecfbd35bb26e8b149664ecde049'
[21:47] <Compn> michaelni : wm4 was wondering if we can make hevc asm optimizations as qualifying tasks. assuming thats not too much work for a student to try ?
[21:47] <Compn> or if the right student comes along, at least that would be a possibility
[21:47] <wm4> Compn: I never suggested it as qualification task btw.
[21:48] <wm4> and from what I can see, the mentor is missing
[21:48] <Daemon404> jamrial or kurosu maybe? they might be too busy though.
[21:48] <Compn> can make it a full project if it gets a mentor and student that can do it ? :)
[22:10] <michaelni> Compn, if you find a mentor and a student then everything is possible ...
[22:25] <cone-620> ffmpeg 03Clément BSsch 07master:d75a73b5fc13: Drop outdated RELEASE_NOTES
[22:38] <ubitux> please poke me ASAP if you want me to add some stuff into the release notes
[22:38] <ubitux> since the release is coming soon
[23:03] <klaxa> hi, i want to apply for the gsoc mentored project for basic servers, how do i formally apply for that?
[23:05] <michaelni> klaxa, contact the listed mentor for a qualification task and register at google once student registrartions are open (they are not yet open)
[23:06] <klaxa> okay, will do, thanks!
[23:33] <cone-620> ffmpeg 03selsta 07master:6c8a158989b4: hls: detect SAMPLE-AES encryption