[Ffmpeg-devel-irc] ffmpeg-devel.log.20170923

[burek]

[01:47:49 CEST] <atomnuker> last version of the opus psychoacoustic system sent to the ml, will push tomorrow if there are no objections
[01:48:01 CEST] <atomnuker> RiCON: ping if you want to test how it sounds
[07:20:52 CEST] <cone-030> ffmpeg 03James Almer 07master:bb4b7624d9dd: avformat/apngenc: use av_packet_alloc()
[07:20:52 CEST] <cone-030> ffmpeg 03James Almer 07master:afe674734bbe: avformat/gif: use av_packet_alloc()
[08:48:55 CEST] <cone-030> ffmpeg 03Jorge Ramirez-Ortiz 07master:1ef7752d64cb: libavcodec: v4l2: add support for v4l2 mem2mem codecs
[09:45:28 CEST] <cone-030> ffmpeg 03Rostislav Pehlivanov 07master:2ad1768c7b48: opusenc: implement a psychoacoustic system
[14:48:40 CEST] <JEEB> funky
[14:51:33 CEST] <JEEB> atomnuker: I think you broke ffserver :D (I only know because it gets built with the default config)
[14:52:57 CEST] Action: JEEB adds explicit "disable-ffserver
[15:03:54 CEST] <atomnuker> JEEB: I broke it? how?
[15:08:04 CEST] <JEEB> I dunno, it just doesn't link :D
[15:08:06 CEST] <JEEB> sec
[15:10:30 CEST] <nevcairiel> opusenc_psy.c:510: undefined reference to `ff_generate_window_func'
[15:10:32 CEST] <nevcairiel> atomnuker: ^
[15:10:40 CEST] <nevcairiel> all of fate is turning red
[15:11:21 CEST] <JEEB> yea, I just pointed out that for him :)
[15:17:44 CEST] <atomnuker> yeah, I have a fix, testing now
[15:22:50 CEST] <atomnuker> durandal_1707: patch ok? https://0x0.st/hBB.patch
[15:25:51 CEST] <durandal_1707> atomnuker: switch have one indentation level too much
[15:26:47 CEST] <atomnuker> odd, was my text editor, fixed
[15:27:45 CEST] <atomnuker> so okay now?
[15:27:51 CEST] <nevcairiel> does "simplify dependency" mean "gets rid of"? avcodec can't depend on lavfi, because lavfi already depends on avcodec, circular is trouble
[15:28:44 CEST] <atomnuker> I guess
[15:28:51 CEST] <atomnuker> I could move it to lavu
[15:29:07 CEST] <atomnuker> (since its universally useful)
[15:32:40 CEST] <durandal_1707> well whatever you prefer
[15:34:46 CEST] <atomnuker> I prefer the way its done currently, until I find another application for it not in lavfi
[15:36:13 CEST] <cone-133> ffmpeg 03Rostislav Pehlivanov 07master:039ebaa5f39e: lavfi: make window_func an inline function
[17:26:20 CEST] <JEEB> ok, I think I got some stuff backported that is useful
[17:26:58 CEST] <JEEB> for some reason our version of movenc utilizes a different time base in the muxer, but otherwise the output seems correct so I'll just update the output of the FATE tests
[17:27:31 CEST] <JEEB> -                    timescale = 30
[17:27:32 CEST] <JEEB> +                    timescale = 15360
[17:27:45 CEST] <JEEB> why? E_NO_IDEA
[17:28:06 CEST] <JEEB> (none of the timescale-related code was changed)
[17:29:12 CEST] <JEEB> which just means the +/-1 becomes +/-512 in CTS offsets
[17:31:43 CEST] <cs122> Hello!
[17:45:12 CEST] <JEEB> alright, posted the back-ports for the negative CTS offset support onto ffmpeg-devel
[18:24:45 CEST] <cone-133> ffmpeg 03James Almer 07master:b8eaecbf39a5: avcodec/opusenc_utils: add missing preprocessor guards
[18:24:46 CEST] <cone-133> ffmpeg 03James Almer 07master:e4fd7b1fea7e: avcodec/opusenc_psy: use av_clip_uintp2()
[18:40:31 CEST] <cs122> The android bionic libc defines B0 on termbits.h, and ffmpeg libavcodec uses  some variables with B0 in its name, which gets renamed by the preprocessor expansion.
[18:41:38 CEST] <cs122> This leads to build errors that can be fixed, amongst other ways, by renaming those variables  by replacing B0 with B_0 or another similar name
[18:41:52 CEST] <cs122> However I don't know if this is the way to go.
[19:22:50 CEST] <cone-133> ffmpeg 03Marton Balint 07master:7f80b065a626: avformat/mxfdec: factorize packet pts setter function
[19:22:51 CEST] <cone-133> ffmpeg 03Marton Balint 07master:01911b9b3cad: avformat/mxfdec: use the common packet pts setter function for opatom files
[21:20:57 CEST] <cone-133> ffmpeg 03Thomas Mundt 07master:58ca446672fe: avfilter/tinterlace: use drawutils for pad mode
[21:20:58 CEST] <cone-133> ffmpeg 03Thomas Mundt 07master:40bfaa190c61: avfilter/interlace: add support for 10 and 12 bit
[22:53:35 CEST] <cone-133> ffmpeg 03James Almer 07master:015f976aaee5: avcodec/frame_thread_encoder: use av_packet_alloc()
[23:14:21 CEST] <cone-133> ffmpeg 03James Almer 07master:ff7f859c2595: avcodec/Makefile: skip v4l2_m2m headers if needed
[23:14:22 CEST] <cone-133> ffmpeg 03James Almer 07master:d1e7e4fbe2b0: avcodec/v4l2_m2m: add missing header inclusions
[23:46:21 CEST] <brash> What do 'starting new cluster due to timestamp' info messages mean? Specifically, I am reading with av_read_frame into a packet, and then using av_copy_packet into a new packet using av_write_frame to write the packet.
[23:59:23 CEST] <jamrial> brash: that's a matroska mxuer log message, and it probably means something may be wrong with your packet timestamps
[00:00:00 CEST] --- Sun Sep 24 2017