[Ffmpeg-devel-irc] ffmpeg-devel.log.20161203

[burek]

[00:13:57 CET] <cone-811> ffmpeg 03James Almer 07master:b5a0b8572178: fate: rename fate-mov-init-nonkeyframe ref file
[00:13:57 CET] <cone-811> ffmpeg 03James Almer 07master:f88c8e0dc3a6: compat/atomics: rename header guards
[00:19:16 CET] <durandal_170> ubitux:  i want that music to try it on our scopes
[01:20:13 CET] <nevcairiel> the music isnt free, but you can buy it from the artist that created it if you really care
[01:22:06 CET] <nevcairiel> costs only ¬5
[01:33:16 CET] <J_Darnley> Impressive visuals
[01:46:02 CET] <Compn> nice
[01:46:06 CET] <Compn> great now im hooked
[01:46:07 CET] <Compn> https://www.youtube.com/watch?v=XziuEdpVUe0
[01:48:36 CET] <J_Darnley> Things like this is (a small) part of why my winamp vis port is such slow work.
[01:49:24 CET] <J_Darnley> I make a change, compile, then watch the track play.
[01:53:59 CET] <wm4> I thought everyone on winows now uses fb2k
[01:58:31 CET] <wm4> I wonder is it intentional that ffh264 doesn't log any errors if broken decoded frames are discarded?
[02:00:57 CET] <Compn> people hate messages
[04:10:26 CET] <cone-423> ffmpeg 03Michael Niedermayer 07master:90da187f1d33: avformat/utils: Check start/end before computing duration in update_stream_timings()
[04:10:26 CET] <cone-423> ffmpeg 03Michael Niedermayer 07master:afcf15b0dbb4: avformat/oggparsespeex: Check frames_per_packet and packet_size
[04:15:52 CET] <Compn> going through cds from 1999, so many vivos
[07:57:36 CET] <Zeranoe> Getting a build error with zimg 2.3 https://paste.ubuntu.com/23571355/
[07:59:21 CET] <durandal_170> Zeranoe: you need latest master because of zimg bugs
[09:39:42 CET] <nevcairiel> durandal_170: you should make configure check for a working version if its required
[12:56:31 CET] <cone-471> ffmpeg 03Carl Eugen Hoyos 07master:6c059bb2afc0: configure: Fix standalone compilation of the rtp muxer.
[14:06:42 CET] <cone-471> ffmpeg 03Michael Niedermayer 07master:5cd3ea1cc080: avformat/ffmdec: Remove some st->codec uses which set encoder parameters
[14:06:43 CET] <cone-471> ffmpeg 03Michael Niedermayer 07master:591c03242164: avformat/ffmdec: replace most codec uses by codecpar
[14:06:44 CET] <cone-471> ffmpeg 03Michael Niedermayer 07master:e706e2e77573: avformat/ffmdec: Check media type for chunks
[14:06:45 CET] <cone-471> ffmpeg 03Michael Niedermayer 07master:418a2b247818: avformat/ffmdec: Drop flags2, debug and codec->time_base setting
[14:06:46 CET] <cone-471> ffmpeg 03Michael Niedermayer 07master:6c15618b90e4: ffserver: remove one avcodec_context_copy()
[15:42:30 CET] <cone-471> ffmpeg 03Carl Eugen Hoyos 07master:a8cb567e559a: lavc: Remove CR/LF from avpriv_request_sample() calls.
[15:42:31 CET] <cone-471> ffmpeg 03Carl Eugen Hoyos 07master:5a237c33b86a: lavf/asfdec_f: Remove CR/LF from avpriv_request_sample() call.
[17:32:21 CET] <cone-471> ffmpeg 03Michael Niedermayer 07master:c143e7939f27: avformat/ffmdec: remove last use of st->codec
[17:32:22 CET] <cone-471> ffmpeg 03Michael Niedermayer 07master:acc163c6ab52: avcodec/flacdsp_template: Fix undefined shift in flac_decorrelate_indep_c
[17:32:23 CET] <cone-471> ffmpeg 03Michael Niedermayer 07master:83a75bf6c31b: avcodec/flacdec: Fix signed integer overflow in decode_subframe_fixed()
[17:33:52 CET] <wm4> so does ffserver not use any deprecated API anymore now?
[19:35:21 CET] <cone-471> ffmpeg 03Muhammad Faiz 07master:01ebb57c03ab: swresample/resample: do not rebuild filter when sample_delta is zero
[19:47:45 CET] <cone-471> ffmpeg 03Michael Niedermayer 07master:fdf588d7cf68: avformat/ffmdec: Forward error code from ffm_read_header()
[19:47:46 CET] <cone-471> ffmpeg 03Michael Niedermayer 07master:b98510de5c2a: avformat/ffmdec: cleanup on extradata memory allocation failure
[19:47:47 CET] <cone-471> ffmpeg 03Michael Niedermayer 07master:9859f54db3d2: avformat/ffmdec: Silence "may be used uninitialized in this function" warnings
[20:09:04 CET] <RiCON> BtbN, philipl: -hwaccel cuvid doesn't work but -c:v *_cuvid does?
[20:09:20 CET] <BtbN> hm?
[20:09:44 CET] <BtbN> Worked fine last time I tested it.
[20:10:04 CET] <BtbN> It was broken on master for a few days, but that should be long fixed
[20:10:23 CET] <RiCON> using N-82699
[20:10:57 CET] <RiCON> both with "-f null -" and "-c:v *_nvenc"
[20:11:17 CET] <BtbN> N-82722-gf31730f works.
[20:12:35 CET] <BBB> michaelni: I really dont think it helps the vote to open discussions on unrelated topics in that very thread
[20:12:48 CET] <BBB> michaelni: Id recommend to open a discussion into these questions in a new email thread
[20:13:33 CET] <BBB> michaelni: and then clearly specify your vote in your email, right now youre saying something about treating applications equally, I think I know what you mean but in a legal sense, your vote would be declared invalid if this was a regular parliamentary election
[20:14:10 CET] <BBB> just say Id like to keep ffserver (which is what I think you mean, but its very unclear), or Id prefer to get rid of ffserver
[20:14:28 CET] <JEEB> in general it's just good manners to make it clear if you're voting what you're voting for
[20:14:50 CET] <JEEB> so that there can be no free interpretations of it by the people taking those in
[20:26:30 CET] <BtbN> RiCON, just tested with latest master. Still works as expected.
[20:28:03 CET] <RiCON> weird, still broken with 82755
[20:28:10 CET] <RiCON> "CUVID hwaccel requested, but impossible to achieve."
[20:31:00 CET] <RiCON> working fine in mpv too, with both cuda and cuda-copy
[20:31:34 CET] <BtbN> you requested an impossible decoder/filter/encoder chain then, as some segment of it does not support cuda input and/or output
[20:34:37 CET] <RiCON> i tried with 8-bit h264/hevc/hevc main10, with "-f null -" and even transcoding with nvenc
[20:35:05 CET] <BtbN> even transcoding with nvenc? That's like the only thing it's useful for.
[20:39:12 CET] <RiCON> -hwaccel cuvid: https://i.fsbn.eu/FoOj.txt / -c:v h264_cuvid: https://i.fsbn.eu/n6fn.txt
[20:41:31 CET] <BtbN> you're trying to use cuvid hwaccel with the h264 software decoder. That's why it fails.
[20:41:42 CET] <RiCON> oh
[20:41:51 CET] <RiCON> so one needs both -hwaccel and -c:v?
[20:42:18 CET] <BtbN> yes, of course. cuvid is not a classic hwaccel like vdpau or vaapi
[20:42:50 CET] <RiCON> hm, so qsv happened to work without it
[20:43:32 CET] <wm4> -hwaccel enables ffmpeg_qsv.c
[20:43:43 CET] <wm4> which I guess does device creation and opaque pixfmt output
[20:43:59 CET] <BtbN> qsv is also not a classic hwaccel
[20:44:16 CET] <BtbN> it should be under the same constraints. Unless ffmpeg_qsv does some hackery to change the default decoder.
[20:44:43 CET] <BtbN> Or it does not validate the chain like cuvid does, and you are using a software chain without realizing
[00:00:00 CET] --- Sun Dec  4 2016