[Ffmpeg-devel-irc] ffmpeg-devel.log.20150116

burek burek021 at gmail.com
Sat Jan 17 02:05:02 CET 2015


[01:14] <cone-975> ffmpeg.git 03raxon1s 07master:4184d344af4a: ffmpeg: Add bitrate value under -progress action
[02:59] <cone-975> ffmpeg.git 03Michael Niedermayer 07master:ba97cf2c4562: avdevice: Use av_format_get_control_message_cb()
[02:59] <cone-975> ffmpeg.git 03Anshul Maheshwari 07master:a47c9331ef13: avformat: Improve data streams/codec support
[02:59] <cone-975> ffmpeg.git 03Anshul Maheshwari 07master:bb0344dc5582: ffmpeg: improve data stream support
[03:02] <aetasx> hey guys, when Im doing a static build with avx disabled, there are still the runtime tables that get build and make their way into the static libs.  Is there something different about these that keeps them from being skipped like the others?
[03:09] <Timothy_Gu> aetasx: what kind of runtime tables are you talking about?
[03:11] <aetasx> Timothy_Gu: they're used in the h264 codecs.  these ones ff_hevc_v_loop_filter_luma_8_avx.weakfilter
[03:11] <aetasx> ff_vp9_ipred_h_16x16_avx
[03:37] <kierank> those look like functions
[03:39] <aetasx> they're the table generated functions related to the hardcoded-tables configure flag
[06:47] <fffan> why av_frame_get_best_effort_timestamp does not work for wma audio for wmv?
[13:09] <cone-530> ffmpeg.git 03Vittorio Giovara 07master:89df3fd49e99: swscale: fix warning about incompatible function pointer type
[13:09] <cone-530> ffmpeg.git 03Michael Niedermayer 07master:a0aa3eaafc79: Merge commit '89df3fd49e9992441f680326902b4912d79f514f'
[14:20] <cone-530> ffmpeg.git 03Vittorio Giovara 07master:2df7277711cb: swscale: fix gbrap to gbrap alpha scaling
[14:20] <cone-530> ffmpeg.git 03Michael Niedermayer 07master:aee51ee8b9b8: Merge commit '2df7277711cbe3ca5415ff8e31228dcffa309211'
[14:47] <cone-530> ffmpeg.git 03Nidhi Makhijani 07master:3941df546276: aea: Return proper error code on invalid header
[14:47] <cone-530> ffmpeg.git 03Michael Niedermayer 07master:2d0440f93f93: Merge commit '3941df546276b190cc9362fd093e6721e8e52f50'
[20:06] <cone-168> ffmpeg.git 03Michael Niedermayer 07master:261e8ccbd1e3: ffmpeg: drop redundant frame_pts value
[20:06] <cone-168> ffmpeg.git 03Michael Niedermayer 07master:7b32856011d1: ffmpeg: Maintain higher precision pts between video filter and fps code
[20:41] <cone-168> ffmpeg.git 03Michael Niedermayer 07master:d3d9a00bcdfd: ffmpeg: Use filter graph output frame rate also for frame duration estimation
[20:45] <akira4> ubitux,do you have a minute?
[21:14] <michaelni> argh, next time akira is here someone please tell her to ask her question as someone else than ubitux might know the awnser ...
[21:14] <Daemon404> also also stay for more than 20 mins
[21:14] <Daemon404> or use email
[21:14] <michaelni> yes
[21:41] <wm4> wow why is the flac parser so complex
[21:43] <kierank> lovely http://git.videolan.org/?p=ffmpeg.git;a=commitdiff;h=7b32856011d1859a974b884f76bd17b7a7fdab76
[21:43] <kierank> ffmpeg.c is such a masterpiece
[21:44] <wm4> kierank: at least strict reviews ensure future contributions will be of high quality?
[21:45] <kierank> wm4: yep so strict that the review was done in secret
[21:45] <kierank> and by one person on his own code
[21:49] <Daemon404> [20:41] < wm4> wow why is the flac parser so complex <-- because the flac format is shite?
[21:49] <Daemon404> bitpack all the things
[21:50] <Daemon404> byte alignment is for chumps
[21:50] <Daemon404> etc
[22:00] <BtbN> This sample_aspect_ratio thing confuses me.
[22:00] <BtbN> The -aspect cli option does not seem to mean the sample_aspect_ratio option.
[22:01] <nevcairiel> there is two of those, one in AVStream for the container and one in AVCodecContext for the codec
[22:03] <BtbN> And how would i get the display aspect ratio, so 16:9, 4:3 or whatever it can be, from within an encoder?
[22:11] <cone-168> ffmpeg.git 03Michael Niedermayer 07master:a46a23d30fea: avfilter/vf_framepack: Check and update frame_rate
[22:13] <aetasx> you guys familiar with the runtime generated prediction tables code?
[22:27] <michaelni> kierank, I hadnt posted it to devel because i thought noone wants to maintain it, does understand and is interrested in that code except me. If thats wrong, please do review it now, iam happy to make any corrections needed if reviewes find issues
[22:34] <aetasx> do you guys know why the runtime predicted table functions like ff_hevc_h_loop_filter_chroma_10_avx still get generated and included in the static libs even if you disable avx?  wasn't sure if there was a technical reason for it
[22:35] <ubitux> michaelni: precision* ;)
[22:36] <ubitux> (akira4 sent me a mail)
[22:46] <michaelni> also if someone wants to review patches to the fps code theres still one on ffmpeg-devel "[FFmpeg-devel] [PATCH 1/2] ffmpeg: last frame duplication heuristic"
[22:49] <halfie> hi! I have a raw bitstream which i suspect is encoded AAC audio samples. is there a way to prove that samples are really AAC encoded?
[22:50] <cone-168> ffmpeg.git 03Michael Niedermayer 07master:f96148913b31: Fix precission typos
[22:56] <ubitux> thx :)
[00:00] --- Sat Jan 17 2015


More information about the Ffmpeg-devel-irc mailing list