[Ffmpeg-devel-irc] ffmpeg.log.20190621

[burek]

[00:00:00 CEST] --- Fri Jun 21 2019
[03:17:18 CEST] <^Neo> If I want to stream some MPEG-TS stream within a single host machine, is the preferred method using RTP?
[03:27:48 CEST] <kevinnn> hi all, I am having some trouble with artifacting right now in my stream
[03:27:54 CEST] <kevinnn> increasing i_slice_max_size to 10000
[03:28:00 CEST] <kevinnn> resolves like 99% of the artifacting
[03:28:06 CEST] <kevinnn> but increases the bandwidth to a number that is not acceptible
[03:28:13 CEST] <kevinnn> are there any other solutions?
[03:28:20 CEST] <kevinnn> I don't mind giving up some CPU power or stream quality
[03:28:27 CEST] <kevinnn> just no artifacting
[03:28:33 CEST] <kevinnn> also I can't give up on latency either
[03:28:36 CEST] <kevinnn> thanks!
[03:34:20 CEST] <DHE> kevinnn: you're gonna need to specify a lot more detail. plus I can't even find that option in ffmpeg -h full on my system
[03:36:39 CEST] <kevinnn> DHE: I am using x264.h directly via c++
[03:37:04 CEST] <DHE> oh... well this is ffmpeg
[03:37:22 CEST] <kevinnn> here is my full x264 config
[03:37:23 CEST] <kevinnn> https://pastebin.com/W99KbK1v
[14:54:57 CEST] <grosso> is av_packet.side_data automatically deallocated at av_packet_unref?
[14:56:09 CEST] <DHE> yes
[14:56:13 CEST] <JEEB> > av_packet_free_side_data(pkt);
[14:56:17 CEST] <JEEB> in av_packet_unref
[14:56:18 CEST] <JEEB> so yes
[14:59:11 CEST] <grosso> thanks
[15:00:13 CEST] <DHE> tfw you run ffmpeg under valgrind to try and find a memory leak
[15:00:15 CEST] <DHE> :/
[15:00:19 CEST] <DHE> (yes I've done that)
[15:00:23 CEST] <JEEB> yes
[15:00:35 CEST] <JEEB> same here, and generally I've tested my patches so that there's no memleaks
[15:01:12 CEST] <DHE> #5799  found one in the release version of ffmpeg
[17:58:30 CEST] <sklv> hi i have this script to convert wmvs to mp4 and i'm trying to make it decode the progress informatino so i can calculate an eta http://ix.io/1MpA
[17:59:20 CEST] <sklv> the issue is it captures all the lines *except* the ones with the progress information - any idea why? #python says it could be because ffmpeg detects if its running in a pty
[18:06:03 CEST] <DHE> sklv: that's possible. you should be able to force it with -stats 1 though
[18:10:14 CEST] <sklv> that gives an error but just -stats on its own runs but doesn't make a difference
[18:10:30 CEST] <sklv> I think i will pipe to a log file and tail the log file
[00:00:00 CEST] --- Sat Jun 22 2019