[Ffmpeg-devel-irc] ffmpeg-devel.log.20160807

[burek]

[04:41:39 CEST] <cone-642> ffmpeg 03Chris Moeller 07master:7387506b0693: avformat: read id3v2 comment tags
[05:46:34 CEST] <Timothy_Gu> atomnuker: you mentioned the fast coder has been tested a lot, but the patch hasn't been merged into master yet, right?
[05:47:22 CEST] <Timothy_Gu> Also if you could add the 6-7 times info into the commit message it'd be great
[05:47:49 CEST] <Timothy_Gu> and nice work :)
[08:39:02 CEST] <bp0> Can I push that hdcd analyze_mode patch? https://patchwork.ffmpeg.org/patch/110/
[08:39:21 CEST] <bp0> What is the normal time to wait for comments?
[08:49:59 CEST] <Timothy_Gu> bp0: the hdcd_ana_mode_t typedef isn't really needed right? Either way, the _t suffix is supposed to be reserved, so an int could be used instead
[08:51:01 CEST] <bp0> its not anymore, I guess, it was an enum
[08:51:12 CEST] <bp0> but i changed to use #defines
[08:51:46 CEST] <bp0> I will change to int
[08:52:51 CEST] <Timothy_Gu> !!(ctlret == HDCD_TG_MISMATCH) the double !! isn't needed
[08:54:24 CEST] <bp0> !! removed
[08:56:08 CEST] <bp0> next? :)
[11:21:13 CEST] <cone-654> ffmpeg 03Vitaly Ostrosablin 07master:90ddcaaf2de7: avfilter: add crystalizer audio filter
[11:23:46 CEST] <cone-654> ffmpeg 03Paul B Mahol 07master:81fcd9126544: doc/filters: fix typo in extrastereo docs
[14:27:59 CEST] <BtbN> Can someone get rid of that broken chatbot/troll over in #ffmpeg?
[14:32:38 CEST] <durandal_1707> das bot?
[15:09:32 CEST] <iive> BtbN: which one?
[15:11:51 CEST] <BtbN> jookiyaya. Kind of too late now though. Also hard to tell if it's just a troll. Seriously suspected a chatbot for a while.
[15:12:18 CEST] <iive> more like a troll, he poasted same link in few other channels
[15:12:24 CEST] <iive> but got ingored there
[15:54:54 CEST] <cone-654> ffmpeg 03Paul B Mahol 07master:1562a4c4f3eb: doc/filters: reorganize some entries, move scale_npp to right section
[16:10:00 CEST] <durandal_1707> 2 filters left for 300th one
[16:49:41 CEST] <kierank> shame  the api is useless
[16:50:14 CEST] <kierank> still willing to fund a decent api
[16:54:16 CEST] <Chloe> What's useless about it?
[16:54:42 CEST] <nevcairiel> its only useless for his usecase
[17:01:17 CEST] <durandal_1707> kierank: I will look at that
[17:01:33 CEST] <kierank> nevcairiel: it's not
[17:01:49 CEST] <kierank> you can count the number of api users of lavfi on one hand
[17:02:26 CEST] <nevcairiel> and i seriously doubt they have the same reasoning as you
[17:02:32 CEST] <kierank> any realtime playback applicaton (vlc) can't use lavfi
[17:02:36 CEST] <kierank> because it buffers arbitrarily
[17:03:18 CEST] <kierank> it's too "designed" around ffmpeg.c and files
[17:03:44 CEST] <nevcairiel> it doesnt guarantee any certain buffering, but thats not the same as buffering arbitrarily, because in reality its not going to just buffer all sorts of random shit unless you use something dumb like the loop or reverse filters
[17:04:01 CEST] <kierank> in my email i gave a perfectly reasonable use-case where it buffered arbitrarily
[17:04:17 CEST] <kierank> you have no idea what the latency is and whether it's fixed or variabble
[17:05:01 CEST] <kierank> again not a problem if you are a file
[17:05:08 CEST] <durandal_1707> its for overlay two frames...
[17:05:36 CEST] <kierank> overlaying subtitles on a video is not a hard demand
[17:05:42 CEST] <nevcairiel> there is only so many options you have when merging two streams, either you buffer or you drop
[17:06:35 CEST] <kierank> the next problem is undefined latency in a filter chain overall
[17:07:38 CEST] <nevcairiel> a new API isnt going to fix that, not witout making the filtergraph internally extremely inflexible
[17:10:57 CEST] <durandal_1707> well filtergraph as concept in lavfi is very limited
[17:11:50 CEST] <kierank> i don't particularly care what's going inside the filtergraph just so long as after n milliseconds I'll have output
[17:12:47 CEST] <kierank> non-file playback is not a hard problem
[17:15:20 CEST] <kierank> maybe we should have a session on it at vdd
[17:16:06 CEST] <bencoh> that's not just "kierank's usecase", that's actually pretty much every "broadcast" (live) usecase, afaict
[17:16:17 CEST] <kierank> it's any live playback case
[17:16:26 CEST] <kierank> vlc would have the same problem
[17:16:38 CEST] <bencoh> (and one of the reason I didn't use it as well - not the only one, though)
[17:16:58 CEST] <kierank> bencoh: what was the other reason?
[17:18:07 CEST] <bencoh> avframe handling/referencing but that was quite some time ago (and maybe it'd have been worked out given more time)
[17:18:26 CEST] <kierank> I think this is sorted now
[17:20:10 CEST] <durandal_1707> it was nightmare before..
[17:20:55 CEST] <bencoh> durandal_1707: that's what it looked like to me, yeah. haven't had a look at all that stuff for over a year now
[17:36:54 CEST] <cone-654> ffmpeg 03Michael Niedermayer 07master:bbec14de3126: avcodec/raw: Fix decoding of ilacetest.mov
[19:19:48 CEST] <cone-654> ffmpeg 03Yuli Khodorkovskiy 07master:f0adb99d068e: avcodec/qsvdec: Fix null dereferences in the qsv decoder
[20:21:59 CEST] <cone-654> ffmpeg 03Paul B Mahol 07master:571f33de0aa8: avfilter/avf_showspectrum: improve dB scale description
[20:42:49 CEST] <Compn> thats great though that edit list patch is posted :)
[20:44:55 CEST] <Compn> only a couple hundred lines too
[20:45:08 CEST] <Compn> impressive for such a feature people requested for years
[23:08:45 CEST] <Prelude2004c> hey guys.. nobody really talking in ffmpeg .. question that might be simple
[23:08:46 CEST] <Prelude2004c> hey guys.. good day ... i am looking for some help here with something silly that i can't figure out... so i am splitting an input like " -filter_complex '[0:v]yadif,format=nv12,split=3[vout1][vout2][vout3]" and the output can't keep up with the input and its running at less than 1x... ok so to debug i did -filter_complex '[0:v]yadif,format=nv12,split=1[out1]   and it did the same thing..only a single output . same problem. now 
[23:08:46 CEST] <Prelude2004c> when i took out the ,split=1[out1] -map '[out1]', everything went ok and it was 1x .. any ideas ?
[23:56:15 CEST] <cone-654> ffmpeg 03Paul B Mahol 07master:22d13e4290c8: avfilter/af_dynaudnorm: use better check for infinite loop
[00:00:00 CEST] --- Mon Aug  8 2016