[Ffmpeg-devel-irc] ffmpeg-devel.log.20150424

burek burek021 at gmail.com
Sat Apr 25 02:05:02 CEST 2015


[01:25:27 CEST] <cone-969> ffmpeg 03Andreas Cadhalpun 07master:4b657a1b1eed: alac: reject rice_limit 0 if compression is used
[01:56:45 CEST] <cone-969> ffmpeg 03Michael Niedermayer 07master:c0d847e457c1: tests/fate-run: do not attempt to parse tiny_psnrs output if it failed
[03:01:22 CEST] <cone-969> ffmpeg 03Martin Storsjö 07master:eb7ddb506684: rtsp: Don't warn about unparsed time ranges
[03:01:23 CEST] <cone-969> ffmpeg 03Michael Niedermayer 07master:f8951d36dfca: Merge commit 'eb7ddb5066843ab16caf33b05eafb2576648b797'
[03:22:16 CEST] <cone-969> ffmpeg 03Andrey Utkin 07master:1bd2646a6da8: rtpenc_jpeg: Handle case of picture dimensions not dividing by 8
[03:22:17 CEST] <cone-969> ffmpeg 03Michael Niedermayer 07master:cd69c0e0181c: Merge commit '1bd2646a6da808f6f9881525771db098c54bc3d2'
[04:01:04 CEST] <cone-969> ffmpeg 03wm4 07master:0f776df09e0b: avformat/wvdec: fix seeking
[04:13:46 CEST] <Timothy_Gu> michaelni: are your release branch fate clients using "branch" variable in fate_config.sh? It's messing the server up
[04:14:44 CEST] <Timothy_Gu> BTW the new server finally got deployed at http://fatebeta.ffmpeg.org/ \o/
[04:24:34 CEST] <michaelni> they set branch somewhere yes
[04:25:43 CEST] <michaelni> i should add some other branches maybe not just releases if the server cannot yet handle that
[04:26:18 CEST] <michaelni> or is there a problem on the client side ?
[05:10:37 CEST] <cone-969> ffmpeg 03Timothy Gu 07release/2.4:0014b243733d: fate: Include branch information in the payload header
[08:58:11 CEST] <cone-359> ffmpeg 03James Zern 07master:acb07f33c40c: git-howto.texi: fix a typo
[09:07:37 CEST] <cone-359> ffmpeg 03James Zern 07master:238ec505e263: libvpxenc: only set noise reduction w/vp8
[09:07:38 CEST] <cone-359> ffmpeg 03James Zern 07master:e07ef1ada3d8: libvpxenc: cosmetics: reindent
[09:07:39 CEST] <cone-359> ffmpeg 03James Zern 07master:db3f86871eb1: libvpxdec: cosmetics: reindent
[10:57:12 CEST] <cone-359> ffmpeg 03Clément BSsch 07master:8e985b727181: avutil/frame: fix everythnig/everything typo
[16:19:53 CEST] <BBB> michaelni: do you still need a github tree to merge outstanding patches?
[16:20:08 CEST] <BBB> (I have one, its a little old and theres some cruft in it, but it may make your life easier)
[16:35:49 CEST] <cone-359> ffmpeg 03wm4 07master:537ab680534e: avformat/mp3: large id3 tags break concatenated file detection
[16:36:26 CEST] <michaelni> BBB, i can just apply the patches from the ML, its np
[16:38:55 CEST] <BBB> ok ty
[16:39:06 CEST] <BBB> I have a current tree here: https://github.com/rbultje/ffmpeg/commits/vp9-profile1-wip
[16:39:10 CEST] <BBB> dont take the top 2 patches
[16:39:23 CEST] <BBB> they make 444 work but they break 420 (oops)
[16:53:55 CEST] <cone-359> ffmpeg 03Ronald S. Bultje 07master:345e3be7baa7: vp9: add lf_deltas fate test.
[16:53:56 CEST] <cone-359> ffmpeg 03Ronald S. Bultje 07master:4b02032fb4cf: vp9: add fate tests for show-existing-frame feature.
[17:11:20 CEST] <cone-359> ffmpeg 03Ronald S. Bultje 07master:3de13d5212b6: vp9: remove another optimization branch in iadst16 which causes overflows.
[17:11:21 CEST] <cone-359> ffmpeg 03Ronald S. Bultje 07master:ed45edb2f6fe: vp9: use aligned size to write segmentation map into cache.
[17:11:22 CEST] <cone-359> ffmpeg 03Ronald S. Bultje 07master:0eccf7d43af7: vp9: add fate test for segmentation image-edge issue.
[17:11:23 CEST] <cone-359> ffmpeg 03Ronald S. Bultje 07master:640d878bc161: vp9: fix intraonly frame decoding.
[17:11:24 CEST] <cone-359> ffmpeg 03Ronald S. Bultje 07master:5d914a4a4334: vp9: add fate test for intraonly frames.
[17:29:46 CEST] <BBB> michaelni: ty
[17:30:03 CEST] <michaelni> np
[17:30:08 CEST] <BBB> michaelni: whats your opinion on merging a (now working) not-yet-bitexact version of profile 1 decoding?
[17:30:21 CEST] <BBB> basically everything works and it looks mostly ok, but the loopfilter code hasnt been fixed yet
[17:30:52 CEST] <BBB> I wont have much time next week, so we can merge as-is as experimental feature, or just hold off until I get to fixing loopfilter 2 weeks from now
[17:30:56 CEST] <BBB> care either way?
[17:31:12 CEST] <BBB> (profile 1 = 444/422 support in vp9)
[17:31:29 CEST] <michaelni> sounds usefull so i would favor merging
[17:32:38 CEST] <BBB> I also havent security-tested it much yet
[17:32:44 CEST] <BBB> (Im seeing it segfault in bits)
[17:33:03 CEST] <BBB> how about we keep it in my tree for a bit and Ill poke you next week to merge if Ive had time to at least debug the segfaults I see?
[17:33:05 CEST] <michaelni> security issues would be not so good
[17:33:17 CEST] <michaelni> ok
[17:34:19 CEST] <BBB> I think 444/420 work, 422/440 don't
[17:34:23 CEST] <BBB> so thats sort of an issue
[17:34:24 CEST] <BBB> ohwell
[17:41:33 CEST] <BBB> hm ok thats fixed now
[17:41:47 CEST] <BBB> its not tested for performance yet, and again, loopfilter for non-420 is broken
[17:41:57 CEST] <BBB> but if you dare: https://github.com/rbultje/ffmpeg/commits/vp9-profile1-wip
[17:42:29 CEST] <BBB> test samples available at http://downloads.webmproject.org/test_data/libvpx/vp91-2-04-yuv4{22,40,44}.webm
[17:42:50 CEST] <BBB> now I gotta grab my family and run for the airport, bye
[18:50:24 CEST] <cone-359> ffmpeg 03Shivraj Patil 07master:97f074f13404: avcodec/mips: MSA (MIPS-SIMD-Arch) optimizations for HEVC copy and hv mc functions
[20:48:08 CEST] <cone-359> ffmpeg 03Martin Storsjö 07master:b90adb0aba07: rtsp: Make sure we don't write too many transport entries into a fixed-size array
[20:48:08 CEST] <cone-359> ffmpeg 03Michael Niedermayer 07master:10a28e8a08d6: Merge commit 'b90adb0aba073f9c1b4abca852119947393ced4c'
[20:49:03 CEST] <BtbN> Looks like i successfully eliminated allmost all floating point math. Only thing i couldn't find a replacement for is sqrt
[20:51:21 CEST] <michaelni> ff_sqrt()
[20:51:34 CEST] <BtbN> That's in libavcodec, i'm in avfilter.
[20:52:30 CEST] <michaelni> it could be moved to libavutil, where it was originally
[20:52:30 CEST] <BtbN> Was aparently moved there from libavutil, because only libavcodec used it and there were some issues with the table beeing used accross libs.
[20:55:21 CEST] <BtbN> It's decently fast with just sqrt. The only major issue left is the pixel format mapping. Do you happen to know if there is some way to map formats?
[20:55:43 CEST] <BtbN> BGR0/BGRA->BGRA, RGB0/RGBA->RGBA, and so on.
[20:55:52 CEST] <BtbN> In a filter.
[20:56:12 CEST] <BtbN> Currently the filter chain just picks whatever output format it likes, even though i don't intend to do format conversions
[20:56:46 CEST] <cone-359> ffmpeg 03Vittorio Giovara 07master:5bba3ab0cf7a: internal: Make dlog/tlog a no-op when disabled
[20:56:47 CEST] <cone-359> ffmpeg 03Michael Niedermayer 07master:036162a37870: Merge commit '5bba3ab0cf7a0238ee1ea31ca2da08ce860fd8f9'
[21:12:46 CEST] <jamrial> if the table is not too big, it could be duplicated in libavfilter
[21:13:52 CEST] <BtbN> https://github.com/FFmpeg/FFmpeg/blob/HEAD/libavcodec/mathtables.c#L63 it's not too horrible.
[21:15:53 CEST] <cone-359> ffmpeg 03Vittorio Giovara 07master:3c04ec04df8c: riff: Add GLV4 fourcc for mpeg4
[21:15:54 CEST] <cone-359> ffmpeg 03Michael Niedermayer 07master:9c56cf09ca6d: Merge commit '3c04ec04df8cba12f3a7408709ff5c1fddfe8d82'
[21:25:45 CEST] <wm4> that subtitle guy doesn't seem too bad, although he/she/it keeps posting patches with commented code
[21:27:19 CEST] <cone-359> ffmpeg 03Vittorio Giovara 07master:b17cbb0b1c4f: riff: Add GMP4 fourcc for mpeg4
[21:27:20 CEST] <cone-359> ffmpeg 03Michael Niedermayer 07master:861dec5e9025: Merge commit 'b17cbb0b1c4fdf8324eff24d40a02201bfaebfd3'
[21:34:51 CEST] <cone-359> ffmpeg 03Vittorio Giovara 07master:a4edaeb50fc7: hq_hqa: Fix table data for profile 17
[21:34:52 CEST] <cone-359> ffmpeg 03Michael Niedermayer 07master:0d3821a6df96: Merge commit 'a4edaeb50fc7510f28a5d79349a7926a182c9930'
[21:48:41 CEST] <cone-359> ffmpeg 03Vittorio Giovara 07master:d90133b77bf6: asvenc: Mark codec as init-thread-safe and init-cleanup
[21:48:42 CEST] <cone-359> ffmpeg 03Michael Niedermayer 07master:a649650a0ea0: Merge commit 'd90133b77bf69667d10e54de9aae7da223c6876a'
[22:00:49 CEST] <cone-359> ffmpeg 03Vittorio Giovara 07master:43171886e08b: huffyuvenc: Mark codec as init-thread-safe and init-cleanup
[22:00:50 CEST] <cone-359> ffmpeg 03Michael Niedermayer 07master:f2978aa7ed0f: Merge commit '43171886e08b6a2f20a1e2f3ecc95a7984b591cd'
[22:21:04 CEST] <cone-359> ffmpeg 03Vittorio Giovara 07master:bb428e00ac15: hqx: Mark codec as init-thread-safe and init-cleanup
[22:21:05 CEST] <cone-359> ffmpeg 03Michael Niedermayer 07master:8d5088a168d1: Merge commit 'bb428e00ac158244d6691bf135be404e85b66a8b'
[22:24:42 CEST] <wm4> uh what the heck, why does the png decoder have a delay?
[22:24:54 CEST] <wm4> if I feed it one packet, I don't get a picture
[22:25:03 CEST] <wm4> it takes a second flush call to get one
[22:26:31 CEST] <Compn> hmm, never noticed this in the strings gxavc.dll ... G:\GeoSourceCode\MyCodec\8400_GAVC_EncCheckSize_HalfD1_ChgHdr_POS_176x144_UVPitch_SSE3_Di\x264_v628_vc60\encoder\ratecontrol.c
[22:27:16 CEST] <Compn> lots of x264 stufff, strings GXAVC.dll | grep -i x264 
[22:47:23 CEST] <cone-359> ffmpeg 03Vittorio Giovara 07master:74a1cad7e3ba: lclenc: Mark codec as init-thread-safe and init-cleanup
[22:47:24 CEST] <cone-359> ffmpeg 03Vittorio Giovara 07master:eae7338e1592: libx264: Make codec use the init-cleanup flag and mark it as init-thread-safe
[22:47:25 CEST] <cone-359> ffmpeg 03Michael Niedermayer 07master:9ba1c62775ca: Merge commit '74a1cad7e3ba79e5b1e5b2e2bcf6179520442679'
[22:47:26 CEST] <cone-359> ffmpeg 03Michael Niedermayer 07master:c180f0f6d398: Merge commit 'eae7338e1592f4a398b7c3cb9d1ac854b7a44ff8'
[23:20:38 CEST] <kierank> Compn: is it a gpl violation?
[23:23:16 CEST] <cone-359> ffmpeg 03Vittorio Giovara 07master:18db1286b045: libxvid: Make codec use the init-cleanup flag and mark it as init-thread-safe
[23:23:17 CEST] <cone-359> ffmpeg 03Michael Niedermayer 07master:437bdf482e92: Merge commit '18db1286b04557aa2d2df7efbcb65ae825d5a469'
[23:23:18 CEST] <cone-359> ffmpeg 03Michael Niedermayer 07master:c949a4500e14: avcodec/libxvid: remove now redundant init cleanup code
[23:54:57 CEST] <cone-359> ffmpeg 03Andreas Cadhalpun 07master:cfdaa4de6c49: dss_sp: use lowercase codec name without whitespace
[23:54:58 CEST] <cone-359> ffmpeg 03Michael Niedermayer 07master:cea62a90c0b7: Merge commit 'cfdaa4de6c496b0b761c763cd18067cb1af268a7'
[00:00:00 CEST] --- Sat Apr 25 2015


More information about the Ffmpeg-devel-irc mailing list