[Ffmpeg-devel-irc] ffmpeg-devel.log.20150118

[burek]

[00:04] <cone-46> ffmpeg.git 03Luca Barbato 07release/2.4:23fe589e1948: prores: Evaluate all the quantizers
[00:04] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:20d6ae4626a4: hevc: always clip luma_log2_weight_denom
[00:04] <cone-46> ffmpeg.git 03Thiago Santos 07release/2.4:20f9cf744a9a: vaapi: wrap codec specific functions in appropiate #ifs
[00:05] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:73046fa19593: Merge commit '23fe589e19487bc9a40b77dae9509068da5b14b8' into release/2.4
[00:05] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:9d43e51eec49: Merge commit '20d6ae4626a42fbc6e7b9c0040bdd64397b23e11' into release/2.4
[00:05] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:0d801c0bfdec: Merge commit '20f9cf744a9a82ac4b269cb4317a5d59a8553baf' into release/2.4
[00:23] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:1fd55ec507f6: svq1dec: Unbreak the scratch buffer allocation
[00:23] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:bbb86717b303: vf_frei0r: do not increment string if it reached the end
[00:23] <cone-46> ffmpeg.git 03Luca Barbato 07release/2.4:4baee1124b90: mp3: Tweak the probe scores
[00:23] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:12d6bbf7cbcc: Merge commit '1fd55ec507f6f47b4d9fddf8e79a0df4540ef6e4' into release/2.4
[00:23] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:c8a2b1ac3c62: Merge commit 'bbb86717b303a3e4c0809d3cc6fb55580766a17e' into release/2.4
[00:23] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:85d05fdc8000: Merge commit '4baee1124b905cbd75240530e081c8ffa68fddbe' into release/2.4
[00:23] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:705e0e05131f: avformat/mpeg: do not count PES packets inside PES packets during probing
[00:37] <cone-46> ffmpeg.git 03Luca Barbato 07release/2.4:da35008c314e: latm: Do not give a score for a single instance
[00:37] <cone-46> ffmpeg.git 03Luca Barbato 07release/2.4:036f5c5420e4: rm: Use the correct codec_data_size signedness
[00:37] <cone-46> ffmpeg.git 03Michael Lynch 07release/2.4:b31bb39bdd7b: rtsp: Check a memory allocation
[00:37] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:1dc53463d9c9: Merge commit 'da35008c314eae5a10a8c070f5185b7694c0a37e' into release/2.4
[00:37] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:365e0652f94d: Merge commit '036f5c5420e4529f05fa5180f5fa28ca2c5c4065' into release/2.4
[00:37] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:aded1110a32e: avformat/rmdec: rm_read_extradata: add error message for oversized extradata
[00:37] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:564d943b2755: avformat/rmdec: Check for overflow in ff_rm_read_mdpr_codecdata()
[00:37] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:1da83d52bc0f: Merge commit 'b31bb39bdd7b5a53e0d282acc0f0f62b32b17acc' into release/2.4
[00:39] <cone-46> ffmpeg.git 03Clément BSsch 07master:ca3c7be7c68a: avcodec/tiffenc: remove double ;;
[00:39] <cone-46> ffmpeg.git 03Clément BSsch 07master:4d028bf202bd: avcodec/vp8: remove double ;;
[00:39] <cone-46> ffmpeg.git 03Clément BSsch 07master:fa6c9e54df59: avdevice/pulse_audio_dec: remove double ;;
[00:39] <cone-46> ffmpeg.git 03Clément BSsch 07master:14f3e8b821d6: avformat/sbgdec: remove double ;;
[00:47] <cone-46> ffmpeg.git 03Luca Barbato 07release/2.4:bb823e26b1cb: avformat: Make avformat_free_context handle NULL
[00:47] <cone-46> ffmpeg.git 03Luca Barbato 07release/2.4:4d4869162214: cmdutils: Use the correct guard
[00:47] <cone-46> ffmpeg.git 03Luca Barbato 07release/2.4:608e8d8dd754: vf_drawtext: Do not leak the mmapped textfile
[00:47] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:2496dbd68b29: vf_showinfo: Forward the av_image_get_linesize error
[00:47] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:679c3d210f5d: Merge commit 'bb823e26b1cbb91b475c10772cdfd867d4809a65' into release/2.4
[00:47] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:ee025a0c6137: Merge commit '4d48691622149ba5998de08a0acec85d1f4ed46a' into release/2.4
[00:47] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:6fd7456f8572: Merge commit '608e8d8dd754199b657b439f9e722e0b45f84461' into release/2.4
[00:47] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:158cf1d52df6: Merge commit '2496dbd68b29bd36fbf4753a46163d33f6dae70e' into release/2.4
[00:49] <kierank> why is it so hard to understand
[00:49] <kierank> bt601 SARs are not arithmetic
[00:49] <kierank> nor do they apply to the entire digital picture
[00:50] <Daemon404> nicholas does not live in reality
[00:50] <Daemon404> just ask wm4.
[01:01] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:daef7feb09a0: af_resample: check av_opt_set_dict return value
[01:01] <cone-46> ffmpeg.git 03Luca Barbato 07release/2.4:c246b0b4c3a3: avresample: Make sure the even check does not overflow
[01:01] <cone-46> ffmpeg.git 03Timothy Gu 07release/2.4:51f76e4e932e: oss_audio: use a macro to simplify ioctl() error checking
[01:01] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:27487944eff7: swscale: fix sign extensions in yuv planar conversion
[01:01] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:b4004124432e: Merge commit 'daef7feb09a0dde2265d56f77de8ae03f9612d47' into release/2.4
[01:01] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:adfb4661d9ab: Merge commit 'c246b0b4c3a3b02a714e99423cf23d59f8f81409' into release/2.4
[01:01] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:39f2e53a2457: Merge commit '51f76e4e932ebdce8ccf6cf0797651d632cfc3e2' into release/2.4
[01:01] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:a299d93928ae: Merge commit '27487944eff721ef8e310db1a2a52329d9377f71' into release/2.4
[01:09] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:5aceced0a0d2: avio: fix sizeof argument
[01:09] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:6cf27b550d66: aviobuf: check context before using it
[01:09] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:d1ad85fc0206: vf_format: check input validity
[01:09] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:4cd0041d3866: rmdec: check av_new_packet return value
[01:09] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:d6f4e475a107: Merge commit '5aceced0a0d235d485e88fda502019b221dbf3d2' into release/2.4
[01:10] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:153a012384b3: Merge commit '6cf27b550d66963b89b5917568b85c4b49fb18ba' into release/2.4
[01:10] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:34b28d3823af: Merge commit 'd1ad85fc020653be5ad25ed1d3d5091e89fee782' into release/2.4
[01:10] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:8120061be269: Merge commit '4cd0041d38664adcb6f4b3038e277631b85d5dc8' into release/2.4
[01:18] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:1551602b4237: nutdec: check av_new_packet return value
[01:18] <cone-46> ffmpeg.git 03Anton Khirnov 07release/2.4:556a5090f2b0: mp3dec: fix reading the Xing tag
[01:18] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:a7e79f6a43bd: avfilter: check filter link validity
[01:18] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:749fadaa651e: Merge commit '1551602b423755c4ed98c5b7b2c2d6504416726e' into release/2.4
[01:18] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:275cd55a0f4b: Merge commit '556a5090f2b0a20fd9998e1a327875f5b0c8d1d5' into release/2.4
[01:18] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:fefb152cd94c: Merge commit 'a7e79f6a43bdc7e3b164dd78b060f0d5575a0cf0' into release/2.4
[01:28] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:7c710c38f648: audiointerleave: check av_new_packet return value
[01:29] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:b9b689550e75: img2dec: check av_new_packet return value
[01:29] <cone-46> ffmpeg.git 03Luca Barbato 07release/2.4:19fc283dbb53: lavf: replace rename() with ff_rename()
[01:29] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:298f7f1fac0b: Merge commit '7c710c38f6481b892e01d3c1e4781ad160b2935e' into release/2.4
[01:29] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:324797eb83df: Merge commit 'b9b689550e7531b1a2cc893d2af623e37f266936' into release/2.4
[01:29] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:d3316c3f85c2: Merge commit '19fc283dbb53a5f7d6658fd4edcfa59b99369b58' into release/2.4
[01:39] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:8a982092cc34: mxfdec: add missing break
[01:39] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:26ba78adacd8: mov: fix assigment check
[01:39] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:bae05e532670: matroskaenc: write correct Display{Width, Height} in stereo encoding
[01:39] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:b6cd6da17329: Merge commit '8a982092cc3436c25d68ec15b27277c176ce0061' into release/2.4
[01:39] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:97a55f00d35c: Merge commit '26ba78adacd8469fca97c8c833e2e6364b13b7c8' into release/2.4
[01:39] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:6f9bb83dbb93: Merge commit 'bae05e5326703dad3bfe0ed5b31ba73ee9254515' into release/2.4
[01:59] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:242fc6394fec: mtv: improve header check and avoid division by zero
[01:59] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:34e7f70f9f49: assdec: check av_new_packet return value
[01:59] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:2a75c0b1ca16: aacsbr: change order of operation to prevent out of array read
[01:59] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:4edbb0955e04: png_parser: fix size of chunk_lenght
[01:59] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:c7831e55b517: Merge commit '242fc6394fecb403bcbd0f652920f2647d0b08ae' into release/2.4
[01:59] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:6aece62b50cc: Merge commit '34e7f70f9f493f340daab80eba4f12d005ec3e63' into release/2.4
[01:59] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:f038623d5342: Merge commit '2a75c0b1ca16b5480497de0d4c79ef122406a0b5' into release/2.4
[01:59] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:b5b36e45b328: Merge commit '4edbb0955e043d698dcc1b5073b6e714f9cc7960' into release/2.4
[02:08] <cone-46> ffmpeg.git 03Luca Barbato 07release/2.4:e7ee74485b43: hnm4: Use av_image_check_size
[02:08] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:12e1a7013a53: roqaudio: Always use the frame buffer on flush
[02:08] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:b82170336f90: tiffenc: initialize return value
[02:09] <cone-46> ffmpeg.git 03Luca Barbato 07release/2.4:484e015dc8b9: cook: Make sure there is enough extradata
[02:09] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:6cb8e40f4372: Merge commit 'e7ee74485b436c34591177c18c8643764a55d516' into release/2.4
[02:09] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:5b965a508ca3: Merge commit '12e1a7013a53ad957c4ff11a3aebc0763024d24b' into release/2.4
[02:09] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:49f477071dcd: Merge commit 'b82170336f90d06c645d8252ddeccfc92c2f9ccb' into release/2.4
[02:09] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:6f1443421887: Merge commit '484e015dc8b9983297e9269b406c65084daf4528' into release/2.4
[02:23] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:fbc20c3b85be: aacdec: avoid an out-of-bounds write
[02:23] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:cce99f72d1b4: mpegenc: prevent a NULL pointer dereference
[02:23] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:e9aeaa6441f6: qdm2: avoid integer overflow
[02:23] <cone-46> ffmpeg.git 03Luca Barbato 07release/2.4:5891fd017aa7: dvdsubdec: Do not leak on failure path
[02:23] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:4f72aef342c8: Merge commit 'fbc20c3b85be169389b6c9f8806a311d3dea91ea' into release/2.4
[02:23] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:e3ce2a87731d: Merge commit 'cce99f72d1b49d3dfee859136eeff3db32553750' into release/2.4
[02:23] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:f251409adc04: Merge commit 'e9aeaa6441f6fd18fc951d9737887dcf8a9584c0' into release/2.4
[02:23] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:c2b658c0e7b7: Merge commit '5891fd017aa7bed4c423b8511090cf8641a0afa4' into release/2.4
[02:33] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:4d11e4b47db6: libtwolame: prevent a NULL pointer dereference
[02:33] <cone-46> ffmpeg.git 03Luca Barbato 07release/2.4:32701252af65: xwma: Do not leak on failure path
[02:33] <cone-46> ffmpeg.git 03Martin Storsjö 07release/2.4:527617485914: lavc: Move the libtwolame encoder registration to the list for external libraries
[02:33] <cone-46> ffmpeg.git 03Martin Storsjö 07release/2.4:93bf4a74de99: configure: Fix enabling memalign_hack automatically
[02:33] <cone-46> ffmpeg.git 03Paul B Mahol 07release/2.4:375c1050bf74: libavcodec/libtwolame: fix null pointer dereference
[02:33] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:206ca5c7c820: Merge commit '4d11e4b47db6387077682595d53e9a17b50511cb' into release/2.4
[02:33] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:dc84cf241ccc: Merge commit '32701252af65014bb68194bb61d67ec1882ae75d' into release/2.4
[02:33] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:099e0178bf56: Merge commit '527617485914004dc8b772056322ea5ae74c800d' into release/2.4
[02:33] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:9b669f9b1413: Merge commit '93bf4a74de99300fdc0deb83f672bf12e6f1c262' into release/2.4
[02:38] <nevcairiel> is it just me, or does nicolas just show up to troll around on certain topics on the ML recently, and that being the entire extend of what he is doing =p
[02:39] <nevcairiel> without having checked, whatever the outcome, nvenc and libx264 should produce identical behaviour in any case
[02:39] <nevcairiel> if the patch in question does that, it should be applied, and we can fix bt601 things later
[02:47] <compn> nevcairiel : i think most devs just show up to troll actually :P
[02:50] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:61fdbf7ff64c: lavc: fix bitshifts amount bigger than the type
[02:50] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:29e720da76ca: librtmp: append the correct field to the string
[02:50] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:dd195c2c587f: on2avc: Fix out of array access
[02:50] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:21aeae9c6796: svq1enc: check ff_get_buffer return value
[02:50] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:4e251b23c31a: Merge commit '61fdbf7ff64c0ae1bdd6a8d573092dc6924c1dba' into release/2.4
[02:50] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:e82140b09b0f: avformat/librtmp: fix swfurl
[02:50] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:d48d2f166d65: Merge commit '29e720da76ca353dbda9f881562902b41c355b77' into release/2.4
[02:50] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:a0bad45cd8e7: Merge commit 'dd195c2c587f44dbc4be7f059ed182f7d83e6cb4' into release/2.4
[02:50] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:65b2973a5433: Merge commit '21aeae9c679657a1537d0d9127eff280bafc901a' into release/2.4
[03:20] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:39e07ac9fcaf: ansi: check ff_set_dimensions return value
[03:20] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:9e0a38d32b36: avs: check ff_set_dimensions return value
[03:20] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:796bd81835eb: g2meet: check ff_set_dimensions return value
[03:20] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:3aba00b1b331: indeo3: check ff_set_dimensions return value
[03:20] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:baef2736b07d: Merge commit '39e07ac9fcaf3d412f9a33f427072e8ded032d24' into release/2.4
[03:20] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:98ba2887370e: Merge commit '9e0a38d32b36fac7fd73bdb93e820ae0b9e03616' into release/2.4
[03:20] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:faa67e9777eb: Merge commit '796bd81835ebcb69d40bc87b33f13924a5937a1b' into release/2.4
[03:20] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:55c6a482d4fc: Merge commit '3aba00b1b331461611d32cfd56ebe8334f6d1b93' into release/2.4
[04:06] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:3f09d4f6d434: ffv1: fix out-of-bounds read
[04:06] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:88411b87b4bb: display: fix order of operands
[04:06] <cone-46> ffmpeg.git 03Vittorio Giovara 07release/2.4:431f57f04672: libopusenc: prevent an out-of-bounds read by returning early
[04:06] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:ff77fa754bf2: vp7: fix checking vp7_feature_value_size()
[04:06] <cone-46> ffmpeg.git 03Luca Barbato 07release/2.4:5f6f24122116: doc: Update the Changelog for release 11.2
[04:06] <cone-46> ffmpeg.git 03Luca Barbato 07release/2.4:4039f11ed82d: Prepare for 11.2 Release
[04:06] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:fb62580fb6cb: Merge commit '3f09d4f6d43468dbc9307bb937516a32287008dc' into release/2.4
[04:06] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:02d979d0074a: Merge commit '88411b87b4bb3c5820ec232f26ba4a284c11a7f9' into release/2.4
[04:06] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:6feb72a094de: Merge commit '431f57f0467244686ae63a3d06a8cf51f60090ed' into release/2.4
[04:06] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:f5dd1270d9dd: Merge commit 'ff77fa754bf2d6e7231d2e050babefa9a13d1fec' into release/2.4
[04:06] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:2b95572b10ec: Merge commit '5f6f2412211609e46d3f3573f4ac4e136534f1d4' into release/2.4
[04:06] <cone-46> ffmpeg.git 03Michael Niedermayer 07release/2.4:fa1ca88cb6fc: Merge commit '4039f11ed82d84009b5146f0a563f33ced33e4db' into release/2.4
[04:41] <j-b> 'morning
[04:46] <michaelni> 'night
[05:15] <cone-46> ffmpeg.git 03Michael Niedermayer 07master:741fa0493b94: avcodec/mpegvideo_parser: more robust dimension setup
[05:15] <cone-46> ffmpeg.git 03Michael Niedermayer 07master:dff6c9250ba8: avcodec/mpegvideo_parser: Check the ff_set_dimensions() return code
[10:47] <wm4> ubitux: not sure if you can use the DTS field for readorder
[10:47] <wm4> ubitux: DTS must be a valid timestamp, and I'd argue it should at least be "close enough" to the PTS
[10:48] <nevcairiel> does the readorder have to be a strictly continous stream? ie. 2 4 6 is not valid?
[10:48] <wm4> yes
[10:49] <nevcairiel> why does this field even exist anyway
[10:49] <wm4> well it's probably ok, but it will at least confuse the ffmpeg ASS muxer
[10:49] <wm4> to serialize .ass files to a muxed packet stream
[10:49] <nevcairiel> but it has timestamps
[10:49] <wm4> events in ass files don't have to be sorted by timestamps
[10:49] <wm4> and the order matters
[10:49] <wm4> for subtle reasons
[10:49] <nevcairiel> borrow my time machine, shoot whoever though random order was a good idea
[10:49] <nevcairiel> +t
[10:50] <wm4> it was created with the "total compatibility" mindset
[10:50] <wm4> before ass-in-mkv, it was an accidental implication at best
[10:51] <wm4> ubitux: also I'm noticing you use the timestamp 9:59:59.99 for subtitles that have no end time?
[12:08] <wm4> so the nvenc behavior makes "sense" after all...
[12:16] <kierank> wm4: yes
[12:17] <nevcairiel> it jsut doesnt match other encoders behaviour in ffmpeg, which is bad imho
[12:19] <kierank> that's because other encoders don't work in terms of DAR
[12:19] <kierank> working in terms of DAR is actually the correct way to do it
[12:21] <kierank> x262 has this problem as well
[12:21] <kierank> and so I had to hack the sar field in x262 to be a DAR
[12:23] <kierank> unfortunately the "SARs must be mathematical" crowd won't understand that
[12:24] <JEEB> uhh, I'm a person like that yet I'm alright with the simplified representations f.ex.
[12:25] <JEEB> like 40:33 for 16:9 480p instead of the longer mathematical one :D
[12:25] <Mavrik> I've actually met some decoders that choke on more "complex" SARs
[12:26] <Mavrik> some Windows Phones for example O.o
[12:27] <kierank> JEEB: yeah
[12:27] <kierank> https://github.com/kierank/broadcastencoder/blob/staging/filters/video/video.c#L95
[12:28] <JEEB> yup
[12:30] <wm4> makes me glad that many file format specify aspect in the container
[12:31] <JEEB> I made a bug report on one of MS's parsers of AVC streams once, and it seems like they just failed to understand some of the predefined SAR values, and only parsed if the "custom SAR" flag was set
[13:03] <kierank> heh
[14:30] <kierank> I give up
[14:42] <compn> as well you should
[14:43] <compn> spend hours fighting over some random aspect ratio
[14:43] <compn> :P
[14:43] <compn> ehe
[14:43] <JEEB> what...
[14:43] <JEEB> I should read the ML
[14:43] <compn> should have applied nvidia's patch instead
[14:43] <compn> let them deal with it :D
[14:43] Action: compn trooooolls
[14:43] <compn> this is what happen when NIH
[14:44] <compn> arguing about things and citing wikipedia
[14:49] <JEEB> I do understand the point that the encoder should encode with the aspect ratio that you set, and it should be a straightforward thing. The proper way would be to have the proper SAR/DAR set when the source is read so people wouldn't expect the f.ex. 16:9 part of 480p to be the full picture
[14:51] <JEEB> and for the nonexact aspect ratios one just has to look at the DVD and BD specs
[14:51] <JEEB> oh well
[14:51] <compn> just let nvidia change aspect, its their crap, let them fix it
[14:51] <compn> kierank : you need to reply in patches, not english
[14:51] <compn> thats the only way to solve this issue of not understanding each other
[14:52] <JEEB> well my idea is more like ffmpeg should read the aspect ratio correctly to begin with, which is not always possible of course
[14:52] <kierank> you'll all go beserk if I submitted a patch that makes it right
[14:52] <kierank> I can tell you that now
[14:52] <compn> kierank : try it
[14:52] <compn> do it!!!!
[14:53] <kierank> I would change the API to set DAR
[14:53] <kierank> and if DAR wasn't set you would read SAR
[14:53] <kierank> because SAR and DAR *don't* necessarily have a mathematical relationship
[14:54] <JEEB> but wouldn't the DAR just say "16:9"
[14:54] <JEEB> or "4:3"
[14:54] <kierank> yes
[14:54] <kierank> and you then use a LUT for the bt601 case
[14:54] <kierank> for other cases I don't really care what ffmpeg does
[14:54] <kierank> nvidia uses DAR
[14:54] <kierank> and ffmpeg assumes you can just do a multiplication to get DAR
[14:55] <kierank> that's why the thread doesn't like the "correction"
[14:55] <JEEB> I'm just thinking of how you could possibly see which of the two is set
[14:55] <JEEB> as in, is it the BT.601 case
[14:55] <kierank> if( dar.width && dar.height )
[14:55] <compn> just have it report 'par: not set'
[14:55] <JEEB> or is it the "standard" case
[14:56] <JEEB> (where the full picture is meant"
[14:56] <wm4> making api changes is hard
[14:56] <JEEB> or is it a if ( dar.width && dar.height && BT.601 ) kind of thing
[14:56] <kierank> wm4: henc
[14:56] <kierank> e
[14:56] <kierank> 1:52 PM <"kierank> you'll all go beserk if I submitted a patch that makes it right
[14:56] <compn> kierank : submit rfc about dar api change
[14:56] Action: JEEB scratches his head
[14:56] <kierank> anyway it's good - I can add a slide about how open source devs don't understand aspect ratios in my fosdem presentation
[14:57] <compn> JEEB : solution is to troll people to send patches, not complain endlessly on irc
[14:57] <JEEB> I understand the two types of aspect ratios but I'm not wholly sure how you'd automatize the correct behavior
[14:57] <kierank> compn: the nvidia behaviour is correct
[14:57] <kierank> so no patches are necessary
[14:57] <JEEB> you can see a lot of OSS shit doing it the wrong way, but things that do it correctly are usually limited to a single field
[14:57] <compn> isnt that what i said ?
[14:57] <JEEB> (like DVD players)
[14:58] <JEEB> where all of their input is of that one specific type
[14:58] <compn> let nvidia deal with it, roughly
[14:58] <kierank> but the people in that thread want things to work based on their beliefs
[14:58] <JEEB> but yeah, I would rather bring this up as the wider issue of handling the BT.601 ARs correctly
[15:01] <JEEB> like, if you read a 16:9 DVD with ffmpeg it should set the BT.601 AR, not the one that sets the full picture to be 16:9
[15:01] <JEEB> that said
[15:01] <JEEB> how do you do it properly?
[15:02] <JEEB> limit to specific resolutions, specific colorimetry and DAR being set?
[15:02] <kierank> mpeg-2 is easy
[15:02] <kierank> it just has DAR
[15:02] <kierank> done
[15:02] <JEEB> true
[15:02] <kierank> h264 is tricker
[15:02] <kierank> because you have to map the aspect_ratio_idc to a DAR
[15:03] <JEEB> didn't AVC have SAR?
[15:03] <JEEB> and at least the blu-ray specs mandate the correct ones
[15:03] <JEEB> same for PSP video
[15:04] <JEEB> s/PSP/UMD/
[15:04] <kierank> sar is used if you don't have a valid aspect_ratio_idc
[15:04] <kierank> in x264 when you use SAR it gets mapped to an aspect_ratio_idc
[15:04] <JEEB> yes
[15:04] <kierank> otherwise, an escape value is used and the SAR is coded directly
[15:04] <JEEB> and those IIRC in the specs are set to specific ones
[15:04] <JEEB> like you have a value for 40:33
[15:04] <JEEB> etc
[15:04] <kierank> yes  they are set to the bt601 values
[15:04] <kierank> and the simple HD ones
[15:04] <JEEB> yup
[15:05] <JEEB> so unless someone did some funky set-up there it should be pretty simple to get the correct one, since it should in correct cases be already set
[15:06] <Daemon404> i am surprised nobody has linked The Link yet
[15:07] <JEEB> yeah, just rechecked the spec and it seems simple enough, unless someone set the value incorrectly (each of the aspect_ratio_idc values has a SAR except for zero)
[15:07] <JEEB> although that is "not present"
[15:07] <kierank> Daemon404: which one
[15:07] <Daemon404> 1 sec
[15:07] <JEEB> but yeah, with MPEG-2 it should be simple enough
[15:08] <JEEB> might post a patch at some point and get all the hatred
[15:08] <Daemon404> hmm
[15:08] <Daemon404> http://lipas.uwasa.fi/~f76998/video/conversion/
[15:08] <Daemon404> but its gone now
[15:08] <Daemon404> :(
[15:08] <kierank> i have "The Link" for interlacing
[15:08] <JEEB> Daemon404, that's what nicholas is actually holding upon for "nv is doing the wrong thing"
[15:08] <Daemon404> heh
[15:09] <JEEB> unless it also explains the simplified versions"
[15:09] <Daemon404> whats there to explain though?
[15:09] <Daemon404> it's just "shits in a standard"
[15:09] <kierank> that's what i'm saying
[15:09] <JEEB> ^
[15:09] <kierank> but he wants things to mathematically work
[15:10] <JEEB> should I take a look at mpegdec?
[15:10] <JEEB> at how it sets the sar/dar?
[15:10] <Daemon404> [23:50] <@Daemon404> nicholas does not live in reality
[15:10] <Daemon404> [23:50] <@Daemon404> just ask wm4.
[15:10] <wm4> once nicholas thinks he's right, there's nothing you can do
[15:10] <Daemon404> pretty much
[15:11] <JEEB> and I guess you'd also have to poke the raw video formats?
[15:11] <JEEB> because when you capture shit it generally should be BT.601 if it's SD
[15:12] <JEEB> (and that would be an even bigger can of worms tbqh)
[15:14] <kierank> ok i posted the links
[15:16] <JEEB> hmm, where did mpegvideo set the sar/dar?
[15:16] <kierank> in the sequence header there is a DAR fiel
[15:16] <kierank> field somewhere
[15:16] <JEEB> mpegvideo_extract_headers doesn't seem to touch it at least
[15:16] <JEEB> where it sets width/height etc
[15:16] <kierank> loool
[15:17] <kierank> it must do
[15:17] <JEEB> https://github.com/FFmpeg/FFmpeg/blob/master/libavcodec/mpegvideo_parser.c#L85
[15:17] <JEEB> sequence extension
[15:17] <JEEB> a lot of shit gets set
[15:18] <JEEB> except for sar/dar :D
[15:18] <nevcairiel> the actual decoder sets its anyway
[15:18] <nevcairiel> thats the parser
[15:19] <JEEB> so it's in mpegvideo.c?
[15:19] <JEEB> I didn't notice it :/
[15:19] <nevcairiel> its in mpeg12dec.c
[15:19] <JEEB> ugh
[15:20] <JEEB> I don't know why I was looking only at mpegvideo
[15:20] <wm4> I thought parser and decoder code was shared to some extend, and that's why parsers have an AVCodecContext...
[15:22] <nevcairiel> mpeg2 aspect ratio info is kinda confusing in the code, there is aspect_ratio_info in the sequence header, but it also takes the pan_scan info into account when calculating the final aspect ratio .. probably because of dar/sar differences
[15:22] <wm4> ah
[15:23] <kierank> i have no idea why it cares about pan scan
[15:24] <nevcairiel> AVRational dar = av_mul_q(av_div_q(ff_mpeg2_aspect[s->aspect_ratio_info], (AVRational) { s1->pan_scan.width, s1->pan_scan.height }), (AVRational) { s->width, s->height });
[15:24] <nevcairiel> thats what it does :D
[15:24] <kierank> yeah i saw
[15:24] <JEEB> yeah, I would simplify a lot of that shit
[15:25] <JEEB> meanwhile I really wonder what should be done with MPEG-2 where the colorimetry is set to BT.709 specifically
[15:25] <JEEB> Japan airs such stuff in 720x480
[15:27] <JEEB> I should check the broadcast specs for that
[15:34] <JEEB> thanks for the ETSI doc btw kierank 
[15:35] <JEEB> reading the japanese side of things too
[15:39] <JEEB> anyways, food and then maybe something :3
[15:45] <Daemon404> kierank, fyi i mirrored the other link here: https://chromashift.org/video/
[15:45] <Daemon404> (because it is easier for me to remember)
[15:45] <kierank> I was going to do that
[15:45] <kierank> but well done
[15:46] <kierank> the notable part of that document is that it doesn't use SAR anywhere
[15:46] <kierank> SAR is a construct of ffmpeg and pals
[15:46] <kierank> i mean the way SAR is calculated is a construct
[15:58] <JEEB> wow
[15:58] <JEEB> the ARIB spec sure has nice tables
[16:08] <JEEB> kierank, how do you read tables 6-5 to 6-7 http://www.arib.or.jp/english/html/overview/doc/6-STD-B21v5_4-E1.pdf
[16:08] <JEEB> of chapter 6.1
[16:08] <kierank> s000 slow
[16:10] <JEEB> because my reading of table 6.7 (output to a progressive 720p screne)
[16:10] <JEEB> seems to hint that no cropping should be done
[16:11] <JEEB> it does use the wording "in 750p format"
[16:11] <kierank> 750 i guess because of vbi
[16:12] <JEEB> 480 * (3/2) and 720 (* (16/9))
[16:12] <JEEB> is what it says to do for interlaced 720x480
[16:12] <JEEB> as far as I can tell
[16:12] <JEEB> which hits 1280x720
[16:14] <JEEB> I'm just trying to understand if this means that they indeed use BT.709 stuff
[16:14] <JEEB> instead of BT.601 aspect ratios
[16:15] <JEEB> the colorimetry is specified to always be BT.709 at least
[19:24] <rcombs> that doesn't look like a crash to me
[20:44] <cone-798> ffmpeg.git 03Muhammad Faiz 07master:eba5cb2028f1: avfilter/showcqt: add gamma2 option
[21:07] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:8d7ce5cdb707: avformat/nsvdec: Check the return value of av_get_packet()
[21:07] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:65b83ce01b94: avformat/rmdec: Check the return value of av_get_packet()
[21:56] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:2a3af7728441: avcodec/svq1enc: Check memory allocations
[21:56] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:79888388e7c4: avcodec/svq1enc: fix error handling/cleanup in case of ff_get_buffer() or scratchbuffer alloc failure
[22:09] <cone-798> ffmpeg.git 03Mark Reid 07master:b396914c6813: libavformat/mxfdec.c: support demuxing opatom audio without index
[22:20] <ubitux> wm4: well, it felt like it was the exact purpose of dts to deal with reordering... now does dts make any sense in a text subtitle context?
[22:28] <wm4> well it would be nice if subtitles wouldn't just blatantly break assumptions about PTS which are otherwise always true
[22:29] <wm4> also, it seems PGS uses PTS for a real purpose
[22:29] <ubitux> you mean dts?
[22:32] <wm4> err
[22:32] <wm4> yes
[22:33] <ubitux> i wanted to apply that just for text subtitles
[22:33] <ubitux> but i guess i need to find another way?
[23:44] <kierank> ubitux: for closed captions it does
[23:44] <kierank> it keeps a buffer in memory and then presents them
[23:46] <nevcairiel> pgs uses them the same way
[00:00] --- Mon Jan 19 2015