[Ffmpeg-devel-irc] ffmpeg-devel.log.20190102

burek burek021 at gmail.com
Thu Jan 3 03:05:04 EET 2019


[00:24:41 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:07607a1db879: avcodec/4xm: Fix returned error codes
[00:24:41 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:fe17f9b95683: swscale/yuv2rgb: Return a more specific error code from ff_yuv2rgb_c_init_tables()
[00:24:41 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:5d9a69167cb1: avcodec/aacps: Return more specific error codes.
[00:24:41 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:057328b29d74: avcodec/bitstream: Return specific error codes when building vlc tables
[00:24:41 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:9520d51e21f9: avcodec/avpacket: Avoid unspecific return -1 for av_grow_packet()
[00:24:41 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:6747a6339759: avcodec/asvenc: Replace bitstream space check by assert
[00:24:41 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:34f942c6c41f: avcodec/assenc: Return more specific error codes for ass_encode_frame()
[00:24:42 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:c6db2e97dcd2: avcodec/ac3enc: Use specific error code in validate_float_option()
[00:24:43 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:8e1b5ba53839: libavcodec/ac3dec.c: Avoid unspecific -1 error code
[00:24:44 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:5079eca22a16: avcodec/alac: Avoid unspecific error codes and forward error codes
[00:24:45 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:2d9fe6321e32: avcodec/adpcm: use assert for codec id instead or error return
[00:24:45 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:bb2d8708f7bc: avcodec/binkaudio: Use assert instead of return for use of compile time disabled codecs
[00:24:47 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:0505b35a38c5: avcodec/dct: Propagate error code from ff_rdft_init()
[00:24:48 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:8b1099c288ab: avcodec/dcaenc: Return specific error code from subband_bufer_alloc()
[00:24:49 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:eb49d259f420: avcodec/eatqi: Return more specific error code from tqi_decode_mb()
[00:24:50 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:3ee323583993: avcodec/eacmv: Return more specific error codes and forward error codes
[00:24:51 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:b33de5574704: doc/indevs: fix upto typo
[00:24:52 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:47c3a10b16f2: avfilter/vf_tonemap_opencl: Make static tables const
[00:24:53 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:796807022c93: avfilter/ebur128: Simplify by using log10()
[00:24:54 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:5785e0e9c401: avfilter/af_crossfeed: Use ff_exp10()
[00:24:55 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:1b6695354d9e: avfilter/af_biquads: minor simplification by using ff_exp10()
[00:24:56 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:044be48faead: avcodec/acelp_pitch_delay: Minor simplification by using ff_exp10()
[00:24:56 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:33d1fb06279c: avcodec/aacdec_fixed: Provide context to av_log()
[00:24:58 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:052ad4d25540: avcodec/imc: Provide non NULL context to av_log()
[00:24:59 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:9446c5887f99: avcodec/mss2: Provide non NULL context to av_log()
[00:25:00 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:b89fc3d01b8e: avcodec/bink: Provide non NULL context to av_log()
[00:25:00 CET] <cone-364> ffmpeg 03Michael Niedermayer 07master:1dcb5b7dca2b: avcodec/bfi: Provide non NULL context to av_log()
[15:15:58 CET] <ubitux> why is ffmpeg.c mux_timebase set to the encoder timebase instead of the output AVStream timebase set by the muxer itself?
[15:16:07 CET] <ubitux> i'm confused.
[15:17:49 CET] <Compn> in case codec and container timebase differ ?
[15:18:18 CET] <Compn> maybe hardware encoders ?
[15:32:16 CET] <ubitux> the pkt timestamps are rebased from the encoder timebase to the muxer timebase before being sent to the muxer
[15:32:24 CET] <ubitux> but in ffmpeg it seems kind of different
[15:32:40 CET] <ubitux> it's like, it's converted from encoder time base, to a muxer timebase and then to the ost tb
[15:32:59 CET] <ubitux> there is this intermediate mux timebase which isn't actually about the muxer
[15:33:16 CET] <ubitux> since the muxer expect it in ost timebase
[15:33:25 CET] <ubitux> (defined by itself)
[15:36:12 CET] <cone-314> ffmpeg 03Paul B Mahol 07master:af5b6b45b4d8: avfilter/vf_detelecine: fix obvious frame memory leaks
[16:15:46 CET] <cone-314> ffmpeg 03Paul B Mahol 07master:25c7aa9974af: aformat/bfi: use ff_get_extradata()
[16:35:22 CET] <cone-314> ffmpeg 03Paul B Mahol 07master:1139887ace7f: avformat/acm: use ff_get_extradata()
[16:35:23 CET] <cone-314> ffmpeg 03Paul B Mahol 07master:e8d36ba47753: avformat/yop: use ff_get_extradata()
[16:35:24 CET] <cone-314> ffmpeg 03Paul B Mahol 07master:37af319086a6: avformat/anm: use ff_get_extradata()
[16:51:48 CET] <cone-314> ffmpeg 03Derek Buitenhuis 07master:8f585c3c550c: mov: Remove duration-of-last-frame heuristic hack
[17:05:20 CET] <atomnuker> kierank: https://www.scotsman.com/news/uk-government-fighting-eu-plans-to-end-daylight-saving-time-1-4849258
[17:05:50 CET] <thardin> haha what
[17:05:58 CET] <nevcairiel> what does the UK care, dont you want to be gone by then
[17:05:59 CET] <atomnuker> are they just doing it because they can and it won't matter soon?
[17:06:26 CET] <thardin> they're out by what, march this year?
[17:06:32 CET] <thardin> april maybe
[17:08:48 CET] <atomnuker> nevcairiel: wish I knew where to go, it seems to me its all bad everywhere
[20:08:21 CET] <atomnuker> it'd be nice if taxes weren't inversely proportional to internet speeds, >= 40% is insane
[20:28:33 CET] <durandal_1707> using avx should be faster that pure sse3 by what amount?
[20:28:47 CET] <nevcairiel> anything from 0 to 100%
[20:29:08 CET] <durandal_1707> why?
[20:29:41 CET] <nevcairiel> if its entirely compute bound and stuff scales perfectly with bigger registers, you can get 100% increase (well, almost, anyway)
[20:29:54 CET] <nevcairiel> if  other factors are limiting it, then it'll be lower then that
[20:29:57 CET] <nevcairiel> there is no one answer
[20:33:56 CET] <durandal_1707> what other factors?
[20:37:44 CET] <kurosu> amd
[20:38:50 CET] <nevcairiel> memory access may be a concern, and there is always the lane business to consider depending on  how the data is structured and has to be moved
[20:40:25 CET] <nevcairiel> pretty short algorithms may also cost a bit more due to avx clock changes, although that factor is disappearing more and more with newer gen cpus
[20:40:58 CET] <nevcairiel> and of course AMD only has  128-bit ALUs, so they dont benefit from avx much, if at all. presumably thats changing with zen2 though, so yay for that, i guess
[22:55:50 CET] <cone-256> ffmpeg 03Gyan Doshi 07master:54109b1d1472: ffmpeg: skip disabled streams
[00:00:00 CET] --- Thu Jan  3 2019


More information about the Ffmpeg-devel-irc mailing list