[Ffmpeg-devel-irc] ffmpeg-devel.log.20180617

[burek]

[04:02:13 CEST] <cone-211> ffmpeg 03Jun Zhao 07master:24be912827f6: fftools/ffmpeg: Replace the number by macro for bprint init
[04:02:13 CEST] <cone-211> ffmpeg 03Jun Zhao 07master:df9fe79b3c32: fftools/ffmpeg_filter: Replace the number by macro for bprint init
[04:02:13 CEST] <cone-211> ffmpeg 03Jun Zhao 07master:d0a2ad241c91: lavc/dvdsubenc: Replace the number by macro for bprint init
[04:02:13 CEST] <cone-211> ffmpeg 03Jun Zhao 07master:e4d14b046ea3: lavc/pngdec: Replace the number by macro for bprint init
[04:02:13 CEST] <cone-211> ffmpeg 03Jun Zhao 07master:9809ac98bde1: lavfi/af_amerge: Replace the number by macro for bprint init
[04:02:13 CEST] <cone-211> ffmpeg 03Jun Zhao 07master:483532bc72a2: lavfi/graphdump: Replace the number by macro for bprint init
[04:02:14 CEST] <cone-211> ffmpeg 03Jun Zhao 07master:abb28d41ea8d: lavf/icecast: Replace the number by macro for bprint init
[04:02:14 CEST] <cone-211> ffmpeg 03Jun Zhao 07master:f7ffb10fb733: lavf/tedcaptionsdec: Replace the number by macro for bprint init
[04:02:15 CEST] <cone-211> ffmpeg 03Jun Zhao 07master:9ba94ac57a2c: avutil/log: Replace the number by macro for bprint init
[04:02:16 CEST] <cone-211> ffmpeg 03Jun Zhao 07master:7f5b8f0883b8: tests/bprint: Replace the number by macro for bprint init
[04:02:17 CEST] <cone-211> ffmpeg 03Jun Zhao 07master:bd47cca6a796: fftools/cmdutils: Replace the number by macro for bprint init
[06:15:52 CEST] <ramiro> jamrial: thanks. is it ok to reply all 4 patches to the existing thread or is it preferred to resend them?
[06:16:28 CEST] <jamrial> what's the thread?
[06:17:03 CEST] <jamrial> if it's old i'd say send the four patches on their own. they will probably get more attention that way
[06:17:33 CEST] <ramiro> jamrial: ok, thanks
[16:19:05 CEST] <ramiro> atomnuker: regarding "lavu/frame: add mb_types side data", could you point me to the previous discussions you mentioned?
[16:21:28 CEST] <atomnuker> not right now, it was last august or so when michaelni wanted to readd the debug interface (which you kinda reimplemented)
[16:22:05 CEST] <atomnuker> anyway, I'd really rather have a json or xml output in the style of what av1's web analyzer uses
[16:22:48 CEST] <ramiro> atomnuker: thanks. I'll have a closer look at av1
[16:24:30 CEST] <atomnuker> ramiro: take a look at examples/inspect.c in the aom repo
[16:24:54 CEST] <jamrial_> what's wrong with a side data for this? we're using that for qp tables
[16:25:11 CEST] <atomnuker> its debug info only
[16:25:23 CEST] <atomnuker> I'd rather not have tons of side data types per codec
[16:25:32 CEST] <ramiro> atomnuker: btw I have a fork (which I don't plan on submitting upstreams because it is way too intrusive) that's geared towards extracting/manipulating values deep inside the codecs <https://github.com/ramiropolla/ffglitch-core/tree/ffedit-0.6>
[16:26:08 CEST] <ramiro> jamrial_: I agree with atomnuker. it will be much more work, but much better to work with
[16:29:10 CEST] <kierank> t
[16:29:42 CEST] <atomnuker> ramiro: thanks, its nice to have someone who cares enough to do it properly
[16:33:48 CEST] <atomnuker> you can express mb_types using the transformType json block, as well as by making all the blockSize data identical (as that's the case for all old codecs mostly)
[16:34:48 CEST] <atomnuker> someone wrote a non-web analyzer last year in feburary, I'll try to see if I can find it again
[16:35:32 CEST] <atomnuker> this is the repo for the web analyzer - https://github.com/mbebenita/aomanalyzer
[16:44:26 CEST] <nevcairiel> the way I see it, formatting the data into any format for an external analyzer should not fall to the codec
[16:44:46 CEST] <nevcairiel> we should have a clear internal representation, binary probably since thats far easier to handle from C code
[17:01:14 CEST] <jkqxz> Where can I find the SA10180.vc1 test file?
[17:10:39 CEST] <bencoh> 66
[17:40:51 CEST] <cone-912> ffmpeg 03Jerome Borsboom 07master:3d028b7b72a0: avcodec/vaapi_vc1: slice_vertical_position starts from zero for the second field
[17:40:51 CEST] <cone-912> ffmpeg 03Jerome Borsboom 07master:89651c82b86e: avcodec/vc1: support multi-slice field interlaced pictures with hwaccel
[18:07:34 CEST] <jamrial_> jkqxz: michaelni, carl or jerome should have it
[18:07:43 CEST] <jamrial_> google is not being very helpful
[19:45:52 CEST] <cone-912> ffmpeg 03Danil Iashchenko 07master:b41b6b323417: libavfilter/opencl: Add macro for setting opencl kernel arguments
[19:57:37 CEST] <atomnuker> nevcairiel: yep, that's the idea
[00:00:00 CEST] --- Mon Jun 18 2018