[Ffmpeg-devel-irc] ffmpeg-devel.log.20181213

burek burek021 at gmail.com
Fri Dec 14 03:05:04 EET 2018


[06:11:37 CET] <Compn> how does duckduckgo make its money ?
[06:12:03 CET] <Compn> and alternatively, why would anyone pollute google results with other , lesser results
[06:16:41 CET] <talih0> Hello, how long does it usually take for a patch to appear on mailing list? I sent mine in ~20 minutes, but it still hasn't been forwarded
[06:19:35 CET] <Compn> probably you are stuck in the queue talih0
[06:19:37 CET] <Compn> let me check
[06:20:55 CET] <Compn> i dont see it in the queue talih0
[06:21:02 CET] <Compn> where did  you send your mail to ?
[06:21:52 CET] <talih0> I sent it to ffmpeg-devel at ffmpeg.org 
[06:22:54 CET] <talih0> from email: andriy.gelman at gmail.com
[06:26:13 CET] <Compn> http://ffmpeg.org/pipermail/ffmpeg-devel/2018-December/237567.html
[06:26:15 CET] <Compn> that one ?
[06:27:01 CET] <talih0> yes, that's the one
[06:27:09 CET] <Compn> it reached the list then :)
[06:28:10 CET] <talih0> thanks, Compn :)
[06:28:19 CET] <Compn> no problemo
[10:23:46 CET] <cone-886> ffmpeg 03Paul B Mahol 07master:f5aad350d369: avformat/wavdec: add support for 'id3 ' chunk
[10:23:46 CET] <cone-886> ffmpeg 03Paul B Mahol 07master:102c11745e2d: avformat/utils: add support for reading ID3 tags at start of wav
[12:01:43 CET] <durandal_1707> what was function in avformat to read lines from file with no limit on length of line?
[14:24:01 CET] <JEEB> so is it valid to make MPEG-TS demuxer ignore non-trustworthy packets until PCR is received?
[14:24:19 CET] <JEEB> because for such streams the demuxer seems to adjust the timestamps IFF PCR is available
[14:24:28 CET] <JEEB> problem is if you receive such a packet before PCR is received
[14:25:30 CET] <JEEB> as in, would anyone here be against making that the default behavior, and then AVOption if people really want to keep such packets that were received before PCR was received
[14:25:56 CET] <JEEB> in theory before PCR is received you don't know if any timestamp is valid, so I think VLC f.ex. just keeps reading until PCR is received
[14:53:42 CET] <durandal_1707> everybody is against it
[14:53:50 CET] <JEEB> yes, thank you
[14:55:18 CET] <durandal_1707> write patch, wait for michael to reply with no, start everything again.
[14:58:38 CET] <JEEB> well yea, I'm expecting someone to cry out "BUT WHAT ABOUT MY PRECIOUS PACKETS?!"
[14:58:59 CET] <JEEB> which is why I would have it have an AVOption if you know your input PTS and DTS are trustable
[14:59:30 CET] <JEEB> another option would be to buffer those untrustworthy streams' packets until PCR
[14:59:33 CET] <JEEB> buut
[14:59:34 CET] <JEEB> not sure I want to go there :P
[15:00:02 CET] <JEEB> simpler to discard them if no PCR
[15:19:13 CET] <michaelni> durandal_1707, iam not stoping you from commiting your original h264 patch. I just think theres an issue and i pointed that out. I understand quite well that my comment is annoying as this is complex code and finding a different solution that works is not easy. Its certainly not my intend to cause you extra work. So i wont complain if you push it as it was. 
[15:22:16 CET] <durandal_1707> michaelni: it is not perfect solution, it gives worse output with sample that have single IDR? frame
[19:32:32 CET] <cone-203> ffmpeg 03Paul B Mahol 07master:faca28c264be: avcodec: rewrite gif muxing and encoding
[19:32:32 CET] <cone-203> ffmpeg 03Paul B Mahol 07master:3f8cdd4ee3fa: avcodec/gif: add support for alpha
[19:32:32 CET] <cone-203> ffmpeg 03Paul B Mahol 07master:80f0f8660c8f: avformat/gifdec: export sample aspect ratio
[19:32:32 CET] <cone-203> ffmpeg 03Paul B Mahol 07master:bb0984cc29e9: avcodec/gif: enable encoding single gif image per frame
[19:32:32 CET] <cone-203> ffmpeg 03Paul B Mahol 07master:280b43246434: avformat/gif: extend description
[19:32:32 CET] <cone-203> ffmpeg 03Paul B Mahol 07master:f2664a306fb5: avcodec/codec_desc: extend gif description
[19:32:33 CET] <cone-203> ffmpeg 03Paul B Mahol 07master:0aa5a7b2e98e: avformat/gifdec: export duration, nb_frames and comment
[19:32:34 CET] <cone-203> ffmpeg 03Paul B Mahol 07master:ec8502f9cb96: avcodec/gif: use avctx->frame_number
[19:32:35 CET] <cone-203> ffmpeg 03Paul B Mahol 07master:0ed678a97abd: bump micro after recent gif changes
[20:43:34 CET] <durandal_1707> shouldn't there be -sidedata:[v,a] option to drop global stream sidedata?
[20:59:55 CET] <durandal_1707> this community is full of clueless people
[21:00:18 CET] <JEEB> thank you
[21:00:47 CET] <durandal_1707> JEEB: not you, you have not designed this mess like side data
[21:01:30 CET] <JEEB> I'm just happy that closed caption side data is going through filter_complex
[21:02:08 CET] <JEEB> which means that if you do lavfi filter chains it won't butcher AVFrame side data
[21:02:12 CET] <durandal_1707> replaygain side data is extracted from global metadata and stored in stream side data, resulting that in some cases it remain in output even when disabled
[21:02:22 CET] <JEEB> ah
[21:02:53 CET] <JEEB> yea I bet a lot of use cases haven't fully been implemented
[21:03:01 CET] <JEEB> be it stream or global or frame side data
[21:11:40 CET] <atomnuker> yeah, I came across that when I tried to remove replaygain side data when remuxing, it was annoying but I had to leave it as-is
[21:55:31 CET] <atomnuker> do wave files support cover art?
[21:55:55 CET] <atomnuker> I have a file here which mediainfo claims to have embedded coverart
[21:56:29 CET] <JEEB> time to check the RIFF parts?
[21:56:33 CET] <JEEB> since IIRC WAV was RIFF?
[21:57:38 CET] <atomnuker> I'm seeing a JFIF sequence in the file followed by some entropy coded looking data
[21:57:48 CET] <durandal_1707> atomnuker: complete uncut ffmpeg console output missing!
[22:08:39 CET] <durandal_1707> atomnuker: does wav file have id3/ID3 chunk ? or it starts with ID3 ?
[22:11:48 CET] <atomnuker> yep, and your patch from today exports it fine so nevermind
[22:28:05 CET] <atomnuker> "wav: Discarding ID3 tags because more suitable tags were found." <- eh, aren't id3 tags more reliable than wav's native?
[22:29:35 CET] <durandal_1707> that happens with ID3 at beginning of wavs
[22:29:49 CET] <durandal_1707> such files are hack anyway
[22:30:06 CET] <durandal_1707> like you could insert ID3 to any file
[22:46:53 CET] <atomnuker> yeah, the files seem sketchy as they have incorrectly signalled tag encoding causing everything to appear as ??????
[22:50:07 CET] <durandal_1707> atomnuker: from where you got it?
[22:53:08 CET] <atomnuker> baidu -_-
[22:56:13 CET] <durandal_1707> michaelni: nut does not take into consideration duration of last frame to return real duration
[23:05:24 CET] <durandal_1707> michaelni: so duration is always smaller by small amount
[23:05:40 CET] <durandal_1707> michaelni: looks like design failure
[23:47:06 CET] <cone-112> ffmpeg 03Paul B Mahol 07master:8c9fff618382: avcodec/xfacedec: fix order of operations
[23:53:16 CET] <cone-112> ffmpeg 03Paul B Mahol 07master:e5a0013c4a62: avformat/mxfdec: fix error check in macro
[00:00:00 CET] --- Fri Dec 14 2018


More information about the Ffmpeg-devel-irc mailing list