[Ffmpeg-devel-irc] ffmpeg-devel.log.20181122

burek burek021 at gmail.com
Fri Nov 23 03:05:05 EET 2018

[02:40:50 CET] <cone-184> ffmpeg 03James Almer 07master:3913d6f73484: x86/scene_sad: fix link errors when HAVE_X86ASM is not defined
[07:37:01 CET] <cone-648> ffmpeg 03kjeyapal at akamai.com 07master:4903662a1da9: avformat/dashenc : Handled error from ff_http_do_new_request() cleanly
[07:37:01 CET] <cone-648> ffmpeg 03kjeyapal at akamai.com 07master:ce0a75363405: avformat/http : Added check for valid URL context before calling shutdown
[10:02:04 CET] <durandal_1707> nobody reviewed my patches? :( time to push them!
[10:10:34 CET] <JEEB> w/25
[10:17:24 CET] <cone-648> ffmpeg 03Paul B Mahol 07master:7d70f0930784: avutil/avsscanf: do not use unsupported p suffix for hex variable
[10:20:30 CET] <cone-648> ffmpeg 03Paul B Mahol 07master:027f032bbce9: avformat/movenc: fix size calculation in mov_write_eac3_tag()
[13:06:21 CET] <j-b> 'morning
[13:11:46 CET] <pross> libavlibc any day now
[13:24:40 CET] <durandal_1707> pross: libavc
[15:31:59 CET] <jya> j-b: hi "Sir" :)
[17:53:19 CET] <nevcairiel> BBB: that blog says dav1d is "feature complete" in 8/10bit, does that include 422/444, or only profile 0?
[17:55:05 CET] <BtbN> I still wonder how hwaccel for AV1 should work
[17:56:10 CET] <BBB> nevcairiel: 422/444/400 work also
[17:56:19 CET] <nevcairiel> BBB: cool
[17:56:32 CET] <jamrial> BtbN: seeing hardware people are waiting for the new spec revision which will apparently be incompatible with the current one, it's not something we should worry about for the next two years :p
[17:56:47 CET] <JEEB> lol
[17:56:55 CET] <nevcairiel> they are breaking the bitstream already?
[17:57:01 CET] <jamrial> looks like
[17:57:03 CET] <atomnuker> yeah
[17:57:13 CET] <nevcairiel> way to fail early
[17:58:19 CET] <jamrial> it's not ideal, but better do it now before hardware is out in the wild than after that
[18:11:10 CET] <BtbN> philipl, made a new round of coded-header releases. Now watch a new patch come up on the ML within a few hours :D
[18:13:25 CET] <nevcairiel> personally I wouldnt even add every function some random person requests, if its not in use by ffmpeg =p
[18:34:05 CET] <philipl> BtbN: 
[18:34:10 CET] <philipl> thanks
[18:36:34 CET] <BBB> nevcairiel: it does everything except 12bit and ext-tile
[18:36:45 CET] <BBB> nevcairiel: I don't know what ext-tile is or how to make samples using it
[18:36:48 CET] <BBB> and 12-bit will be done
[18:37:13 CET] <BBB> more important is to get 10bit avx2 and 8bit earlier optimizations, and finish the tiny pieces of avx2 missing (cfl_ac, z1/2/3, some super_res/scalable_mc stuff)
[18:37:32 CET] <jamrial> isn't that what the tile list OBU is for?
[18:37:40 CET] <nevcairiel> 12-bit is probably a bit niche, as always
[18:38:40 CET] <jamrial> afaik, tile list obu stuff is not required for conformance
[18:44:48 CET] <BBB> nevcairiel: well, I'd like to support it at some point
[18:44:57 CET] <BBB> but not immediately, yes
[18:45:02 CET] <BBB> it's not the most important thing
[21:05:51 CET] <durandal_1707> looks like copy-trac3074 needs to be updated
[21:11:17 CET] <jamrial> durandal_1707: if the changes from 027f032bbc are correct, then update the test
[21:21:11 CET] <durandal_1707> well, it should actually use number of written bits instead of hacks
[21:33:30 CET] <cone-664> ffmpeg 03Paul B Mahol 07master:97d1ee437bbf: avformat/movenc: get number of written bytes from bitstream writer
[21:40:25 CET] <durandal_1707> any more comments for truehd_core_bsf patch?
[22:14:03 CET] <jamrial> durandal_1707: seems to work with the sample i tried. 29mb, 45 seconds long truehd+atmos sample into 20mb, and it decodes fine (same checksum since our decoder just ignores atmos anyway)
[22:15:20 CET] <nevcairiel> i dont suppose you can test if hardware still likes it? maybe bitstreaming the resulting file to some receiver?
[22:15:58 CET] <jamrial> i don't have that kind of hardware :(
[22:19:05 CET] <atomnuker> why is thilo wanting to stir shit now?
[22:20:35 CET] <durandal_1707> we need more bureaucracy
[22:21:28 CET] <jamrial> atomnuker: what?
[22:21:40 CET] <atomnuker> durandal_1707: lets bring the worst of the real world online
[22:23:28 CET] <atomnuker> jamrial: he'll spam the email addresses of every subscriber pointlessly asking to participate in a survey
[22:25:23 CET] <durandal_1707> hey, that is not spamming, its information gathering
[22:34:47 CET] <atomnuker> >every email survey ever sent
[00:00:00 CET] --- Fri Nov 23 2018

More information about the Ffmpeg-devel-irc mailing list