[Ffmpeg-devel-irc] ffmpeg-devel.log.20161020

[burek]

[01:15:25 CEST] <Matador> Compn : yes I am using months ago ffmpeg for QSV
[01:22:40 CEST] <Matador> just wish it was sorted so could use an updated ffmpeg w/ qsv
[08:21:31 CEST] <beauty> why av_frame_get_channels(frame) return 0, but frame->channels is 2?
[09:13:20 CEST] <cone-146> ffmpeg 03Muhammad Faiz 07master:56689a46b304: avfilter/showcqt: add cscheme option
[09:13:20 CEST] <cone-146> ffmpeg 03Muhammad Faiz 07master:258c49d61451: avfilter/firequalizer: use zero phase kernel
[09:56:59 CEST] <cone-146> ffmpeg 03Carl Eugen Hoyos 07master:c0e2846dcdc1: lavc/sheervideo: Increase av_get_codec_tag_string() input buffer size.
[18:09:28 CEST] <cone-507> ffmpeg 03Steven Liu 07master:5702416c57af: avformat/flvenc: do not attempt to write duration and filesize when not seekable
[18:11:39 CEST] <atana> michaelni: I installed ffmpeg from github repo but I don't see ffplay executable file
[18:11:57 CEST] <atana> after make install
[18:12:07 CEST] <michaelni> atana, you need libsdl
[18:12:35 CEST] <atana> for testing the filter I need ffplay right?
[18:13:09 CEST] <michaelni> you dont "need", its one way to test filters
[18:14:28 CEST] <atana> what's the other way to test using command line. For ffplay do I need to install libsdl?
[18:15:29 CEST] <michaelni> you could use ffmpeg 
[18:15:37 CEST] <michaelni> ffplay needs sdl yes
[18:16:19 CEST] <atana> ffmpeg someaudio.mp3 myfilter=filterargs  is this the format?
[19:02:34 CEST] <Chloe> note that it needs sdl2
[19:06:41 CEST] <RiCON> nevcairiel: can you confirm dxva2-copy is bugged/broken since http://git.videolan.org/?p=ffmpeg.git;a=commitdiff;h=9b462a0b9 ?
[19:24:09 CEST] <CFS-MP3> michaelni can you comment on this? http://ffmpeg.org/pipermail/ffmpeg-devel/2016-October/201358.html 
[19:30:18 CEST] <nevcairiel> RiCON: it probably needs more bracers if you compile both dxva2 and d3d11va
[19:30:55 CEST] <nevcairiel> let me push a fix
[19:37:33 CEST] <nevcairiel> RiCON: hopefully that solves it
[19:37:34 CEST] <cone-507> ffmpeg 03Hendrik Leppkes 07master:8bd38ec5bde1: dxva2: fix surface selection when compiled with both d3d11va and dxva2
[19:46:08 CEST] <michaelni> atana, ffmpeg -i  someaudio.mp3 -af ... outfile.wav
[19:55:54 CEST] <cone-507> ffmpeg 03Andreas Cadhalpun 07master:ee5f0f1d355f: rsd: limit number of channels
[20:00:42 CEST] <RiCON> nevcairiel: fixed, thanks
[20:09:45 CEST] <Chloe> Any of the OPs use matrix.org?
[20:11:22 CEST] <BtbN> Nah, I still use classic phone lines to log out.
[20:13:08 CEST] <llogan> kind of looks like Slack
[20:13:44 CEST] <BtbN> Which looks suspiciously like IRC.
[20:15:02 CEST] <Chloe> Yeah, it's pretty much a federated slack. There's a nice matrix<->(freenode)irc bridge, but it requires setup from a channel OP afaik
[20:16:02 CEST] <llogan> Most of my clients request that I use Slack. None have wanted to use traditional IRC.
[20:17:49 CEST] <Chloe> err, by setup, I just mean it needs to be enabled, so matrix users can join. 
[20:23:50 CEST] <CFS-MP3> Thanks for taking a look at the SCTE-35 stuff, michaelni ... now, what's the next step to actually have this patchset merged? Who actually does the commit?
[20:27:00 CEST] <BtbN> So it's also an over-complicated IRC client
[20:27:35 CEST] <Chloe[m]> yes :) Looks like there's no setup needed though
[20:28:02 CEST] <Chloe[m]> it's only if you want to link it to a channel on the matrix network rather than using the normal bridge alias
[21:50:03 CEST] <CFS-MP3> michaelni - Can you merge this? https://github.com/FFmpeg/FFmpeg/pull/238  (or point me in the right direction if you're not the right person to ask)
[21:54:40 CEST] <BtbN> we don't do github PRs.
[21:54:43 CEST] <Chloe[m]> CFS-MP3: ffmpeg doesn't use github.
[21:55:28 CEST] <Chloe[m]> CFS-MP3: it will get merged by a maintainer, assuming there is no longer anyone who still has an issue with it
[22:08:11 CEST] <CFS-MP3> BtbN what's the process then?
[22:08:50 CEST] <Chloe> CFS-MP3: I just said
[22:09:40 CEST] <CFS-MP3> Ah sorry, missed that Chloe
[22:09:45 CEST] <CFS-MP3> OK, so I'll just wait then
[22:33:43 CEST] <atomnuker> michaelni: what do the scale and bits arguments do in ff_mdct_init()?
[22:38:12 CEST] <atomnuker> (putting the scale to 1.0 should (?) make the mdct output normalized values, but what does the bits do, define operation precision?)
[22:45:43 CEST] <peloverde> nbits is log2(number of points)
[22:46:31 CEST] <atomnuker> duh
[22:49:18 CEST] <atomnuker> can't use it then, my input's infuriatingly not a power of two
[22:50:28 CEST] <nevcairiel> we also have the opus 960 mdct if that helps
[22:52:59 CEST] <atomnuker> they're only inverse
[22:53:38 CEST] <nevcairiel> hm i guess so
[00:00:00 CEST] --- Fri Oct 21 2016