[Ffmpeg-devel-irc] ffmpeg-devel.log.20191203

[burek]

[00:02:08 CET] <BenLubar> is there some kind of skeleton demuxer/decoder or am I better off grabbing one of the simple ones and dissecting it?
[00:05:55 CET] <JEEB> since I don't see ffmpeg.org results on DDG, I think while there have been some changes regarding registration of formats, this probably is a pretty good overview https://wiki.libav.org/Internals/DemuxerHowTo
[00:06:44 CET] <BenLubar> thanks
[00:09:10 CET] <JEEB> the registration is now handled by the addition of the thing to libavformat/allformats.c it seems
[00:09:18 CET] <JEEB> makefile changes are still needed, of course
[00:11:04 CET] <JEEB> while the read_header is awfully long, libavformat/yuv4mpegdec.c might be OK to look at?
[00:11:18 CET] <JEEB> since it is a relatively simple raw video container and thus doesn't really have much in there
[00:11:55 CET] <JEEB> basically you implement the callbacks you require
[01:41:58 CET] <cone-533> ffmpeg 03Ross Nicholson 07release/4.2:289838b7bd20: libavformat/rtsp: return error if rtsp_hd_out is null instead of crash
[05:53:56 CET] <BenLubar> I don't know the correct terminology for this, but CMV has an 8 bit index into a texture map and 4+3 bits of color information per pixel. Is there anything in ffmpeg that decodes more than one "pixel per pixel"?
[07:21:08 CET] <BenLubar> alright, I figured out what seems to be a pretty good way to do it. not sure how to do audio properly, but there are only two cmv files in existence with audio as far as I know
[07:58:43 CET] <thardin> "Byte X+(Columns*Y). An index into IBM PC Code Page 437"  wut
[07:59:22 CET] <thardin> this sounds like the format used in that ibm 5050 demo
[07:59:55 CET] <thardin> https://wiki.multimedia.cx/index.php?title=CMV  this is also funny
[08:00:02 CET] <thardin> MJPEG, but worse
[08:06:34 CET] <thardin> 8088 corruption
[08:27:26 CET] <thardin> oh they left
[09:29:17 CET] <cehoyos> I am currently working on a new presentation for the archivists conference in Budapest this week. If anybody has ideas about what to include in the presentation, please say so here and forward the question to Paul! I have to go now but will leave the irc window open, thank you!
[09:55:55 CET] <cone-539> ffmpeg 03Andreas Rheinhardt 07master:710ab136931f: avfilter/vf_unsharp: Don't dereference NULL
[09:55:55 CET] <cone-539> ffmpeg 03Linjie Fu 07master:8fc8bdddbf73: libavformat/utils: Fix code indentation
[09:55:55 CET] <cone-539> ffmpeg 03Guo, Yejun 07master:b864af033d41: MAINTAINERS: add myself to libavfilter/dnn
[11:10:13 CET] <cone-539> ffmpeg 03Marton Balint 07master:f5b83d5419a3: avformat/mpegtsenc: allow any sensible PID for elementary and PMT PIDs
[11:10:14 CET] <cone-539> ffmpeg 03Marton Balint 07master:565dc3e451c7: avformat/mpegtsenc: set priority flag for AC3 codecs if writing BluRay
[11:10:15 CET] <cone-539> ffmpeg 03Marton Balint 07master:db63db3977bb: avformat/mpegtsenc: move around setting m2ts_mode
[11:10:16 CET] <cone-539> ffmpeg 03Marton Balint 07master:998906a0a408: avformat/mpegtsenc: factorize writing packet
[11:10:17 CET] <cone-539> ffmpeg 03Marton Balint 07master:1e0ea369456b: avformat/mpegtsenc: add padding to m2ts streams
[12:43:10 CET] <cehoyos> durandal_1707: I am currently working on a new presentation for the archivists conference in Budapest this week. If anybody has ideas about what to include in the presentation, please say so here
[12:45:19 CET] <durandal_1707> cehoyos: cfhd
[12:59:30 CET] <cehoyos> I will mention it but better would be developments that happened in the last two years
[13:04:40 CET] <JEEB> the last time I had looked into what archivists wnated it was mostly lossless + checksums
[13:04:51 CET] <JEEB> most likely not the only thing, but a theme at least
[13:05:10 CET] <JEEB> (which is why they're standardizing matroska+ffv1 for that apparently)
[13:20:09 CET] <cehoyos> The goal is not to tell the archivists what they want to hear but what we developed over the last two years.
[13:21:00 CET] <JEEB> sure
[13:21:20 CET] <JEEB> I was mostly thinking in the sense of what their interest areas would be :P
[13:21:52 CET] <cehoyos> So basically what was developed since 3.4 was released
[13:22:43 CET] <cehoyos> 50% of those attending use FFmpeg because of FFv1, 50% don't know what FFmpeg is, so my goal is to explain to all 100% why they need FFmpeg independently of FFv1
[17:49:25 CET] <cone-669> ffmpeg 03Michael Niedermayer 07master:3ae87bb3c1cb: tools/target_dec_fuzzer: Support fuzzing error detection
[17:49:25 CET] <cone-669> ffmpeg 03Michael Niedermayer 07master:5ac8675cb1a4: tools/target_dec_fuzzer: Support setting AV_CODEC_FLAG2_FAST
[19:42:40 CET] <cehoyos> kierank: I believe you should tell us what (and who) you would like to hear...
[19:46:03 CET] <kierank> cehoyos: well as I mentioned at vdd they are scared of us
[19:50:51 CET] <cehoyos> Fosdem...
[19:51:28 CET] <cehoyos> (As said, I claim to know the archivists as good as you, I am just curious if anybody wants me to talk about something he - the FFmpeg developer - thinks is important)
[19:53:03 CET] <cehoyos> (As said, I claim to know the archivists as good as you, I am just curious if anybody wants me to talk to the archivists about something he - the FFmpeg developer - thinks is important)
[19:55:27 CET] <kierank> Ok
[19:58:56 CET] <cehoyos> kierank: I believe you should tell us what (and who) you would like to hear at fosdem
[19:59:25 CET] <kierank> Anyone and their project
[21:14:25 CET] <cone-253> ffmpeg 03Kusanagi Kouichi 07master:12bbfc4ccaa1: avdevice/xcbgrab: Handle reply and error properly
[22:32:35 CET] <BenLubar> I'm not sure how to implement the audio for https://patchwork.ffmpeg.org/patch/16572/ - can a container format tell ffmpeg to pull in a different file at a specific timestamp?
[22:40:47 CET] <cehoyos> BenLubar: It may be better to start with video only.
[22:40:52 CET] <cehoyos> The demuxer file is missing a license header
[00:00:00 CET] --- Wed Dec  4 2019