[Ffmpeg-devel-irc] ffmpeg-devel.log.20150226

burek burek021 at gmail.com
Fri Feb 27 02:05:02 CET 2015


[01:03] <cone-666> ffmpeg.git 03Michael Niedermayer 07master:8064265a38a3: avformat/dss: correct sample rate
[01:13] <cone-666> ffmpeg.git 03Michael Niedermayer 07master:5634ca6d378c: avformat/dss: set packet duration
[02:31] <cone-666> ffmpeg.git 03Michael Niedermayer 07master:626904b77797: avformat/dss: set bitrate
[02:32] <cone-666> ffmpeg.git 03Michael Niedermayer 07master:111d79aa8886: avformat/dss: implement seeking
[11:37] <ubitux> michaelni: i changed the qsort in paletteuse and explicted bgra as well for the output; do you want to test anyway before i push?
[11:51] <michaelni> ubitux, yes
[11:51] <ubitux> ok, just a moment then
[11:52] <ubitux> michaelni: pick the colorquant branch on my github
[11:53] <ubitux> https://github.com/ubitux/FFmpeg/compare/colorquant
[11:58] <ubitux> mmh my fate instances aren't reporting anything anymore?
[11:59] <ubitux> ah they are stalled in threads-misc, nuce
[11:59] <ubitux> nice.
[12:07] <michaelni> ubitux,  seems working
[12:22] <ubitux> michaelni: cool, will push soon then, thanks
[12:37] <cone-657> ffmpeg.git 03Andreas Cadhalpun 07master:64ea4a0598e7: avformat/adxdec: set avctx->channels in adx_read_header
[12:37] <cone-657> ffmpeg.git 03Andreas Cadhalpun 07master:7faa40af9829: avformat/adxdec: check avctx->channels for invalid values
[14:20] <cone-657> ffmpeg.git 03Clément BSsch 07master:ab093bdeefad: avfilter/palettegen: export color quantization ratio
[14:20] <cone-657> ffmpeg.git 03Clément BSsch 07master:4216968f854b: avfilter/palettegen: use a logging context for the dupped color warning
[14:20] <cone-657> ffmpeg.git 03Clément BSsch 07master:30f4e9d3d90b: avfilter/palettegen: consistently use the same frame for dimensions
[14:20] <cone-657> ffmpeg.git 03Clément BSsch 07master:321de0349280: avfilter/palettegen: use AV_QSORT()
[14:20] <cone-657> ffmpeg.git 03Clément BSsch 07master:4629993d99f7: avfilter/paletteuse: use AV_QSORT()
[14:20] <cone-657> ffmpeg.git 03Clément BSsch 07master:e0f03de8f558: tests: add palettegen test
[14:20] <cone-657> ffmpeg.git 03Clément BSsch 07master:a6a2da2f3202: tests: add paletteuse test
[14:22] <Daemon404> can anyone tell me what the different av_asserts are for
[14:22] <Daemon404> trying to figure the correct one to use
[14:22] <ubitux> speed trade off
[14:22] <Daemon404> that doesnt tell me shit
[14:22] <ubitux> av_assert0 all the time
[14:22] <Daemon404> an assert is an assert
[14:22] <ubitux> av_assert1 relatively speed relevant
[14:22] <compn> different sizes of assert checks ?
[14:22] <ubitux> av_assert2 very speed relevant
[14:22] <Daemon404> thats... very vague
[14:22] <Daemon404> [13:21] <@Daemon404> an assert is an assert
[14:22] <ubitux> av_assert0 is always compiled
[14:23] <ubitux> av_assert1 is compiled at --assert-level=1 and --assert-level=2
[14:23] <ubitux> av_assert2 is compiled at --assert-level=2
[14:23] <Daemon404> heh
[14:23] <Daemon404> riced asserts
[14:26] <ubitux> ASAN complains
[14:26] <ubitux> on h264 stuff
[14:36] <ubitux> sorry h265
[14:43] <nevcairiel> the big thing is that av_assert0 is compiled even in release builds, so having it in speed critical paths is a very bad idea
[14:44] <wm4> good thing C's assert() does the right thing here
[14:44] <Daemon404> yeah
[14:44] <Daemon404> but im putting it in a call-once patch
[14:44] <Daemon404> path
[14:44] <Daemon404> which is fine
[14:44] <Daemon404> wm4, ;)
[14:51] <ubitux> wow, mips 64
[14:58] <wm4> the cpu you'd use for multimedia
[14:59] <j-b> I would
[15:02] Action: Daemon404 recalls vlc has patches in-tree for thinks like gmp on mips
[15:05] <ubitux> wm4: what if you want to re-encode vp8 stream into h264 on the fly on your router?
[15:05] <ubitux> @_@
[15:06] <wm4> enlightenment
[15:06] <wm4> Daemon404: what does vlc use gmp for?
[15:07] <Daemon404> i dunno but
[15:07] <Daemon404> https://git.videolan.org/?p=vlc.git;a=tree;f=contrib/src/gmp;h=53a174b0c605607b6c458c04873bdafeb0edf6d5;hb=HEAD
[15:07] <Daemon404> probably a dep of a dep of a dep
[16:12] <cone-657> ffmpeg.git 03James Cowgill 07master:213e823dc6bc: mips/mathops: remove 64-bit code
[16:12] <cone-657> ffmpeg.git 03James Cowgill 07master:390771481ab7: mips/aacpsdsp: fix definition of ps_decorrelate_mips
[16:12] <cone-657> ffmpeg.git 03James Cowgill 07master:d6f66aec508f: mips/fft: remove some useless assembly
[16:34] <cone-657> ffmpeg.git 03James Cowgill 07master:83589e691876: mips/sbrdsp: remove sbr_neg_odd_64_mips
[16:34] <cone-657> ffmpeg.git 03James Cowgill 07master:f6bf745c5c2a: mips/aacdec: refactor out duplicated assembly code
[17:21] <Daemon404> hey ubitux did you help the FOSDEM guys use avfilter?
[17:25] <cone-657> ffmpeg.git 03James Cowgill 07master:b9de1303a641: mips/acelp_filters: fix incorrect register constraint
[17:26] <kierank> Daemon404: yes
[17:26] <Daemon404> i see
[17:26] <Daemon404> because they got teh aspect ratio wrong and fucked up the colors on the slides
[17:26] <Daemon404> slides that were white are now grey
[17:28] <kierank> I don't think they care about the AR
[17:39] <ubitux> Daemon404: i think so
[17:39] <ubitux> but just here publicly
[17:40] <Daemon404> i see
[17:40] <ubitux> they forgot to +faststart btw
[17:48] <Daemon404> ubitux, ;)
[19:27] <cone-657> ffmpeg.git 03Vittorio Giovara 07master:43e5e3c077bc: colormatrix: Accept bt601 aliases as input
[19:35] <Daemon404> michaelni, [18:34] < Paranoialmaniac> Daemon404: lgtm
[19:35] <Daemon404> i double checked with yusuke
[19:35] <Daemon404> about the colr patch
[19:35] <Daemon404> seems ok.
[19:38] <michaelni> there are multiple colr patches
[19:38] <Daemon404> michaelni, the one to write proper isom colr
[19:38] <Daemon404> you LGTM'd it too
[19:38] <michaelni> ahh ok sure
[19:39] <michaelni> will you apply it or should i ?
[19:39] <Daemon404> wait 2 sec
[19:39] <Daemon404> he came up with some comment
[19:40] <michaelni> ok no hurry
[19:40] <jamrial> http://fate.ffmpeg.org/report.cgi?time=20150226181447&slot=x86_64-archlinux-gcc-valgrindundef all seem to point to avpriv_mpa_decode_header2()
[19:43] <cone-657> ffmpeg.git 03Luca Barbato 07master:82b6e451161f: configure: Move the cross_prefix setting after the toolchain one
[19:43] <cone-657> ffmpeg.git 03Michael Niedermayer 07master:51b7ab2ee70f: Merge commit '82b6e451161f19ab90bfb9565b710021caf5dcbd'
[19:45] <jamrial> most likely 8ca098f4
[20:11] <cone-657> ffmpeg.git 03Michael Niedermayer 07master:6ad42b3e1547: avformat/mp3dec: properly allocate dummy AVCodecContext
[20:13] <ubitux> heh, yeah valgrind is back and not happy :p
[20:44] <cone-657> ffmpeg.git 03Michael Niedermayer 07master:b851bc20c693: avformat/mp3dec: Check for avcodec_alloc_context3() failure
[20:44] <cone-657> ffmpeg.git 03Dyami Caliri 07master:50833c9f7b4e: Fix buffer_size argument to init_put_bits() in multiple encoders.
[23:16] <cone-657> ffmpeg.git 03Nicolas George 07master:a92193f24781: lavd/alsa: set frame_size field.
[23:16] <cone-657> ffmpeg.git 03Nicolas George 07master:508d6a23b453: ffmpeg: allow to set the thread message queue size.
[23:16] <cone-657> ffmpeg.git 03Nicolas George 07master:d92c6d82c03b: ffmpeg: notify when the thread message queue blocks.
[23:16] <cone-657> ffmpeg.git 03Michael Niedermayer 07master:19dbe89f03d0: Merge remote-tracking branch 'cigaes/master'
[23:27] <cone-657> ffmpeg.git 03Andreas Cadhalpun 07master:eeda2c3de8a8: avformat/bit: check that pkt->size is 10 in write_packet
[23:27] <cone-657> ffmpeg.git 03Andreas Cadhalpun 07master:d0b8640f75ff: avformat/bit: only accept the g729 codec and 1 channel
[00:00] --- Fri Feb 27 2015


More information about the Ffmpeg-devel-irc mailing list