[Ffmpeg-devel-irc] ffmpeg-devel.log.20170313

[burek]

[01:09:54 CET] <ubitux> jkqxz: thanks a lot for the qsv merges :)
[02:46:09 CET] <atomnuker> bofh_: does fate pass?
[02:51:54 CET] <stevenliu> Hi guy,can concatdec.c add live streaming support? or i saw concatdec have only file, is there have some problem to  support live stream and file complex in list?
[03:24:21 CET] <Compn> where steven go
[03:24:30 CET] <Compn> i thought we figured a way for concat:http to work...
[04:49:15 CET] <cone-224> ffmpeg 03Michael Niedermayer 07master:a720b854b0d3: avcodec/mpeg12dec: Fix runtime error: left shift of negative value -1
[04:49:15 CET] <cone-224> ffmpeg 03Michael Niedermayer 07master:a66c6e28b543: avcodec/rv34: Fix runtime error: signed integer overflow: 36880 * 66288 cannot be represented in type 'int'
[04:49:15 CET] <cone-224> ffmpeg 03Michael Niedermayer 07master:967feea5ebb7: avcodec/vp6: clear dimensions on failed resolution change in vp6_parse_header()
[04:59:23 CET] <cone-224> ffmpeg 03Steven Liu 07master:33e997d992cc: avformat/hlsenc: second_levels flags process function extract
[07:55:34 CET] <andrey_turkin> Did anybody experienced issues with cuvid+nvenc transcoding with recent ffmpeg? It stopped working for me since lazy filters initialization patch came in
[09:23:28 CET] <stevenliu> Hi Carl here?
[09:23:57 CET] <ubitux> rarely
[09:24:05 CET] <stevenliu> This is a duplicate of other h264-related tickets.        Which ticket?
[09:25:29 CET] <stevenliu> if the pkt->duration is 0, the hlsenc will has wrong duration at the last segment or discontinuity start segment
[09:48:49 CET] <nevcairiel> like I wrote in the ticket, its not the demuxers job to try to guess the packet duration
[09:49:07 CET] <nevcairiel> it works for s ome si mple codecs with fixed packet sizes
[09:49:15 CET] <nevcairiel> because its just hard-coded somewhere
[11:08:00 CET] <BtbN> andrey_turkin, worked for me when I tested that merge.
[11:08:17 CET] <BtbN> Haven't tested after the recent batch of merges though
[11:30:28 CET] <ubitux> http://b.pkh.me/ffmpeg.jpg so i got bored
[13:57:45 CET] <BBB> michaelni: I dont disagree that he should be on ffmpeg-devel, but well& he isnt& ;-)
[13:58:36 CET] <durandal_1707> who?
[13:59:21 CET] <durandal_1707> vittorio? he just prefer less mails in inbox
[14:00:43 CET] <BBB> yeah vittorio
[14:01:13 CET] <BBB> Im getting a bunch of students that want to do vmaf now
[14:01:18 CET] <BBB> I need more qualification tasks...
[14:01:42 CET] <BBB> maybe Ill make turn netflix code into library a qualification task& although that might be quite a bit of work& too much?
[14:03:28 CET] <wm4> isn't it conceptually pretty simple?
[14:03:37 CET] <wm4> isn't it just basically a function
[14:12:14 CET] <BBB> what is?
[14:12:31 CET] <BBB> vmaf? sure, yes, but their code isnt organized quite like that :)
[14:33:01 CET] Action: kierank hopes the next step isn't an IMF demuxer
[15:31:30 CET] <cone-307> ffmpeg 03Paul B Mahol 07master:62c8dc46429b: avfilter/vf_lut3d: actually skip lines when encountering DOMAIN_ string
[15:45:54 CET] <Compn> BBB : that realplayer ivr format?
[15:46:07 CET] <BBB> kierank: imf=mxf
[15:46:36 CET] <BBB> kierank: ffmpeg already supports it
[15:47:10 CET] <Compn> http://samples.ffmpeg.org/real/ivr/
[15:47:35 CET] <Compn> guess it hasnt been REd by us, wonder if theres spec elsewhere, https://wiki.multimedia.cx/index.php/IVR
[15:47:48 CET] <JEEB> no, IMF is much more complex
[15:47:50 CET] <JEEB> than MXF
[15:47:58 CET] <JEEB> it's multiple MXF files and XML
[15:48:25 CET] <JEEB> also virtual timelines etc defined in the XML if I recall correctly
[15:48:50 CET] <JEEB> IMF is a horrible, horrible format and I hope it dies in a fire
[17:14:04 CET] <pps> is any one online
[18:59:14 CET] <atomnuker> bofh_: ping
[19:33:21 CET] <BBB> JEEB: omg :D well the mxf demuxer can read imf files for me
[19:33:37 CET] <BBB> JEEB: but thats all I know, Ive never looked at these files and it doesnt sound like the type of thing Id care much about
[20:42:03 CET] <cone-334> ffmpeg 03Paul B Mahol 07master:fbfbd97be25c: avcodec/xpmdec: there are XPM files with dos line endings
[21:42:47 CET] <durandal_1707> michaelni: do you know why scpr doesnt work with vlc direct rendering?
[21:57:44 CET] <JEEB> BBB: yea, the separate media files in an IMF package is something that lavf can handle
[21:57:47 CET] <JEEB> https://media.xiph.org/video/derf/meridian/MERIDIAN_SHR_C_EN-XX_US-NR_51_LTRT_UHD_20160909_OV/
[21:57:51 CET] <JEEB> this is an IMF sample
[21:58:00 CET] <JEEB> (and a simple one at that)
[21:58:31 CET] <JEEB> N levels of XML because muxing things together is way overrated :D
[21:59:02 CET] <nevcairiel> its probably designed for use in mastering houses and whatnot, where replacing or adding streams is a common workflow?
[21:59:17 CET] <JEEB> yea
[21:59:35 CET] <michaelni> durandal_1707, no i dont know that
[21:59:48 CET] <JEEB> although the idea is that IMF is what the mastering people export for final encodes
[22:00:23 CET] <durandal_1707> michaelni: then who will know?
[22:12:28 CET] <michaelni> durandal_1707, whats different between scpr and others which work ?
[22:13:50 CET] <durandal_1707> michaelni: it needs to keep prev decoded untouched, for mv stuff and others
[22:14:14 CET] <durandal_1707> using reget buffer doesnt work
[22:14:48 CET] <michaelni> sounds like mpeg P frames
[22:15:03 CET] <kierank> JEEB: yeah it's like a dvd i guess
[22:15:13 CET] <kierank> all these multiformat things don't really fit into lavf
[22:15:20 CET] <kierank> ffmpeg -i imf:// 
[22:15:21 CET] <kierank> eugh
[22:16:26 CET] <JEEB> kierank: don't call it out, it might actually appear ;)
[22:16:45 CET] <kierank> in java
[22:17:01 CET] <JEEB> iä iä
[23:07:09 CET] <nevcairiel> i nominate JEEB to work on it
[23:11:27 CET] <JEEB> after I finally get ATRAC3+ parsing and earthsoft dv upstreamed
[23:11:41 CET] <nevcairiel> havent you been talking about those things for a while now
[23:12:53 CET] <JEEB> that's the idea 8)
[23:13:03 CET] <JEEB> earthsoft dv demuxing I had done
[23:13:30 CET] <JEEB> although then I hit the whole thing about bitstream'd audio
[00:00:00 CET] --- Tue Mar 14 2017