[Ffmpeg-devel-irc] ffmpeg-devel.log.20190407

[burek]

[02:01:17 CEST] <cone-429> ffmpeg 03Carl Eugen Hoyos 07master:d234ed76330a: lavf/Makefile: Fix kux demuxer standalone compilation.
[04:27:35 CEST] <cone-429> ffmpeg 03Jun Zhao 07master:7c1875143db4: doc/examples/metadata: fix the example can't dump FLV metadata
[08:20:57 CEST] <cone-973> ffmpeg 03Lauri Kasanen 07master:50e672bc5437: swscale/ppc: VSX-optimize non-full-chroma yuv2rgb_1
[08:20:58 CEST] <cone-973> ffmpeg 03Lauri Kasanen 07master:3256e949be7b: swscale/ppc: VSX-optimize yuv2rgb_full_2
[08:20:59 CEST] <cone-973> ffmpeg 03Lauri Kasanen 07master:8607e29fa3d4: swscale/ppc: VSX-optimize yuv2rgb_full_X
[11:22:31 CEST] <superware> AVPacket holds pts/dts, where can I find the corresponding time_base relating to the entire packet?
[11:23:53 CEST] <nevcairiel> depends where you got the packet from
[11:24:48 CEST] <nevcairiel> when demuxing, its time_base in the corresponding AVStream
[11:25:08 CEST] <nevcairiel> when encoding its the time_base of the encoding context
[11:25:20 CEST] <nevcairiel> thats probably all major sources of AVPackets
[11:26:28 CEST] <superware> I'm demuxing using av_read_frame, there are multiple streams inside an AVPacket
[11:26:43 CEST] <nevcairiel> no, every packet always corresponds to one stream
[11:26:53 CEST] <superware> hmm
[11:27:15 CEST] <nevcairiel> identified by stream_index
[13:04:53 CEST] <superware> nevcairiel: is there a way to have ffmpeg return the next frame (av_read_frame) from a file at playback speed?
[13:05:16 CEST] <nevcairiel> no
[13:11:47 CEST] <superware> I'll really appreciate any help: I have an app which has a MediaInput class that has a thread spitting out packets, and a MediaPlayer class which has a thread filling a queue of frames (avcodec_receive_frame) and another thread which dequeue and handles synchronization and rendering, this works well for network streams, but not for files since MediaInput has no synchronization. is this a valid use case?
[13:12:34 CEST] <nevcairiel> typically you would synchronize on rendering, ensuring that audio a nd video are rendered at the appropriate times, and with limited queue space it would limit everything else  automatically
[13:13:15 CEST] <superware> so you basically say the MediaPlayer queue is redundant
[13:13:56 CEST] <superware> ?
[14:02:34 CEST] <superware> nevcairiel: thanks
[16:06:59 CEST] <cone-083> ffmpeg 03Nikolas Bowe via ffmpeg-devel 07master:4c8e3725d9ff: avfilter/af_asetnsamples: fix sample queuing.
[16:30:20 CEST] <j-b> Oh, come on
[16:30:25 CEST] <j-b> "Nikolas Bowe via ffmpeg-devel
[16:30:30 CEST] <j-b> How is that a name?
[16:30:42 CEST] <nevcairiel> thats just email being terrible
[16:30:57 CEST] <nevcairiel> and the commiter being lazy or forgetting to fix it
[16:30:58 CEST] <j-b> yet, it stays forever in the git history
[17:07:43 CEST] <durandal_1707> force push it
[17:07:50 CEST] <JEEB> we already have a few of those :P
[17:07:59 CEST] <JEEB> just means that whomever is pushing should double-check author/e-mail
[17:30:51 CEST] <michaelni> or automate it with a script that works better than mine, which failed to catch this one as it was just looking for a name which this one had
[17:35:01 CEST] <JEEB> michaelni: yea. I guess checking for "via ffmpeg-devel" or something would have caught this
[17:35:27 CEST] <JEEB> also checking for the e-mail
[17:35:38 CEST] <JEEB> XXX via ffmpeg-devel <ffmpeg-devel at ffmpeg.org>
[17:35:40 CEST] <JEEB> seems to be the pattern
[17:35:50 CEST] <jamrial> looking for @ffmpeg.org email as author seems easier
[17:36:20 CEST] <JEEB> yea
[17:53:50 CEST] <BtbN> How does it determine the commiter Name/Email? Just the user.name and user.email of the one who ran git commit?
[17:53:54 CEST] <BtbN> Or push, rather?
[17:55:16 CEST] <JEEB> git commit, yea
[17:55:56 CEST] <JEEB> git push just pushes those info nodes onto a remote server
[18:12:25 CEST] <durandal_1707> vel0city: do not forget to finalize proposal, without it you can not work at all...
[18:32:27 CEST] <durandal_1707> michaelni: now when i have both codes, symbols and bit-lengths how can i dechipher this algo?
[18:34:43 CEST] <michaelni> look for patterns
[19:15:17 CEST] <durandal_1707> michaelni: what patterns? anything specific?
[19:22:32 CEST] <durandal_1707> what a waste, libreoffice dec2bin is limited to -512-512
[21:11:30 CEST] <durandal_1707> michaelni: ugh asetnsamples change is wrong
[21:12:00 CEST] <durandal_1707> warning is issued here
[21:16:52 CEST] <BradleyS> $(.+) via (.+)^
[21:17:23 CEST] <cone-851> ffmpeg 03Paul B Mahol 07master:3a2adeedaff3: avformat/riffdec: pass correct pointer to av_log
[21:17:23 CEST] <cone-851> ffmpeg 03Paul B Mahol 07master:ecdaa4b4fa52: avfilter/af_asetnsamples: use correct function
[21:17:40 CEST] <BradleyS> ^$ obviously
[22:05:52 CEST] <durandal_1707> apparently mpc-be guys ripped my dhav code
[22:08:52 CEST] <jamrial> for what purpose?
[22:10:13 CEST] <durandal_1707> for demuxing in dvrsplitter
[00:00:00 CEST] --- Mon Apr  8 2019