[Ffmpeg-devel-irc] ffmpeg-devel.log.20170723

[burek]

[08:23:35 CEST] <ZeroWalker> so, if i always check for if there's something encoded after pushing something to be encoded (avcodec_send_frame), should i just use the older API functions?
[08:28:21 CEST] <atomnuker> no, they're deprecated
[08:57:37 CEST] <ZeroWalker> roger
[08:58:40 CEST] <ZeroWalker> how does one set bitrate if it's VFR input. i can set it, but it doesn't really stay anywhere near it
[10:32:59 CEST] <KGB> [13FFV1] 15retokromer opened pull request #80: tiny typo (06master...06patch-1) 02https://git.io/v7ICy
[11:53:34 CEST] <ZeroWalker> how does one calculate the pts timestamp based on samples + samplerate?
[11:53:59 CEST] <ZeroWalker> i thought it was just to turn it into ms, but that didn't seem to go that well (samples / (samplerate / 1000))
[11:58:13 CEST] <thebombzen> depends on the audio frame size
[11:58:21 CEST] <thebombzen> PTS is the timestamp of the frame in timebase units
[11:58:42 CEST] <ZeroWalker> oh
[11:59:18 CEST] <ZeroWalker> i always fed it the milliseconds based on the cpu clock from the time of capture
[11:59:49 CEST] <ZeroWalker> but so,  if the framesize is 960
[11:59:57 CEST] <ZeroWalker> does that mean, 960 = 1 pts:s?
[12:02:31 CEST] <Blubberbub> at the moment i just do pts += pkt->nb_samples
[12:03:23 CEST] <Blubberbub> so to turn that into ms i would do pts / samplerate
[12:05:29 CEST] <ZeroWalker> packet, isn't it set at frame?
[12:06:47 CEST] <ZeroWalker> http://sprunge.us/iRLF
[12:12:24 CEST] <Blubberbub> packet also has a pts, i think
[12:12:36 CEST] <Blubberbub> (and i'm writing a demuxer and not a decoder)
[12:13:31 CEST] <ZeroWalker> yeah it does, but i think that's just what the frame had, might be wrong
[12:13:54 CEST] <ZeroWalker> ah, well i am encoding a live capture, so it's, messy:(
[12:33:08 CEST] <Blubberbub> does ffmpeg do some magic with stdout? Because i'm missing some output from a script i'm calling from within ffmpeg. 
[13:09:12 CEST] <Blubberbub> ah - found the problem. Its not line buffered, so i need to flush it
[14:59:33 CEST] <cone-621> ffmpeg 03Vodyannikov Aleksandr 07master:47c93657249f: avcodec/cfhd: Fix decoding regression due to height check
[14:59:34 CEST] <cone-621> ffmpeg 03Brice Waegeneire 07master:6a6eec485d23: doc/filters: typo in frei0r
[15:35:18 CEST] <cone-621> ffmpeg 03Muhammad Faiz 07master:7e9f5500039a: avfilter/w3fdif: do not write to line before start line
[18:53:53 CEST] <atomnuker> michaelni: would the patch be okay if I changed length to an int in the function definition?
[22:13:35 CEST] <michaelni> atomnuker, i think so, yes
[00:00:00 CEST] --- Mon Jul 24 2017