[Ffmpeg-devel-irc] ffmpeg-devel.log.20160326

burek burek021 at gmail.com
Sun Mar 27 03:00:04 CEST 2016


[00:00:01 CET] <nevcairiel> that looks fine
[00:11:23 CET] <atomnuker> oh wow, huge vf_colormatrix and swscale patches
[00:12:31 CET] <michaelni> Theres a student searching for a mentor for "Integration of Subtitles into libavfilter" (https://summerofcode.withgoogle.com/dashboard/organization/6504812191416320/proposal/6413595109752832/)
[07:09:08 CET] <Timothy_Gu> ethe: https://gist.github.com/TimothyGu/23ec15872465cd295444
[07:12:09 CET] <Timothy_Gu> ethe: partially done translation: http://sprunge.us/SRJB. it doesn't work fully yet, and I can't find time to fix it
[08:23:42 CET] <relaxed> is Carl Hoyos in this channel?
[10:02:50 CET] <wm4> relaxed: he rarely is
[10:20:30 CET] <cone-180> ffmpeg 03Martin Vignali 07master:495f08edb444: avcodec/exr: add support for B44 and B44A compression
[11:04:04 CET] <cone-180> ffmpeg 03Michael Niedermayer 07master:c06bdc60c90d: avformat/rtpdec: Remove stray debug av_log()
[11:20:01 CET] <Gramner> apparently someone dumped all the internal facebook iOS headers. there's some quality enterprise design principles there, like https://github.com/quellish/Facebook-iOS-Internal-Headers/blob/master/Headers/FBProfileSetEventsCalendarSubscriptionStatusMutationOptimisticPayloadFactoryProtocol-Protocol.h
[11:20:38 CET] <Gramner> also 18.000+ classes in one app, because it's web-scale
[11:32:36 CET] <sfan5> does it use node.js?
[11:32:40 CET] <sfan5> otherwise it's not true web-scale
[11:32:42 CET] <sfan5> also mongodb
[11:53:38 CET] <wm4> Gramner: holy shit what
[11:53:52 CET] <wm4> that's truly enterprise
[12:18:17 CET] <thardin> looks like I was right about that bad mxf file
[12:24:49 CET] <ethe> thanks Timothy_Gu 
[12:55:11 CET] <Compn> relaxed : cehoyos is his irc nick, but he frequently reads the irc logs
[12:55:14 CET] Action: Compn waves at carl
[14:50:49 CET] <cehoyos> Hi, did we already inform vlc that the mail-hunk is not working?
[14:51:02 CET] <cehoyos> (Or our mail-admins?)
[14:56:48 CET] <michaelni> cehoyos, what ?
[14:57:38 CET] <cone-180> ffmpeg 03Carl Eugen Hoyos 07master:de1a0d43595b: lavc/flicvideo: Implement padding in COPY chunks.
[14:58:05 CET] <cehoyos> michaelni: http://ffmpeg.org/pipermail/ffmpeg-cvslog/2016-March/date.html does not show todays commits.
[15:08:44 CET] <michaelni> who messed with the ffmpeg-cvslog filter rules last time ?
[15:10:42 CET] <michaelni> either way ive disabled the rule that i belived caused this
[15:12:30 CET] <RiCON> michaelni: 19:10:37 <    llogan@> due to spam the cvslog entries for web commits may be delayed because messages "from" ffmpeg-cvslog are not set to go to queue until i make an appropriate filter
[15:18:32 CET] <michaelni> what spam ?
[15:19:31 CET] <michaelni> also it blocked everything not just web 
[15:23:21 CET] <ethe> cehoyos: the attachment for "[PATCH] remove the hls_wrap option" looks fine
[15:23:39 CET] <cehoyos> michaelni:
[15:24:06 CET] <cehoyos> michaelni:http://ffmpeg.org/pipermail/ffmpeg-cvslog/2016-March/098804.html http://ffmpeg.org/pipermail/ffmpeg-cvslog/2016-March/098930.html
[15:24:43 CET] <cehoyos> ethe: It looks broken here: http://ffmpeg.org/pipermail/ffmpeg-devel/2016-March/192157.html
[15:24:51 CET] <michaelni> ahh, i guess my spam filter filters these 
[15:25:00 CET] <cehoyos> And here: http://thread.gmane.org/gmane.comp.video.ffmpeg.devel/212297
[15:25:34 CET] <ethe> cehoyos: the attachment is fine though
[15:26:04 CET] <ethe> cehoyos: https://gist.github.com/joshdekock/4fc616b2de2f500a4db5
[15:28:08 CET] <cehoyos> Three applications here believe that the mail has no attachments...
[15:29:19 CET] <cehoyos> Ok, found the patch in one of the mail parts, I still believe it can look better in pipermail with little effort.
[15:35:32 CET] <ethe> cehoyos: odd. I'm using thunderbird and it said there was an attachment
[15:51:11 CET] <cehoyos> ethe: Yes, I just missed it, sorry!
[16:36:34 CET] <cone-180> ffmpeg 03Dan Dennedy 07master:28688d7c2b53: avcodec/audiotoolboxenc: Fix compile error on OS X 10.8.
[17:34:38 CET] <cone-180> ffmpeg 03Michael Niedermayer 07master:d4424d7de2c9: fate: Modify a random h264 test to also test the -framerate option
[17:34:39 CET] <cone-180> ffmpeg 03Michael Niedermayer 07master:5a19506088c3: fate: Remove vsync drop from some h264 tests
[18:08:30 CET] <durandal_1707> whats tep2 merge status?
[18:08:50 CET] <durandal_1707> have all demuxers been updated?
[18:35:50 CET] <wm4> durandal_1707: see etherpad
[18:39:34 CET] <Compn> did anyone ever want to mail a hd with samples ?
[18:39:41 CET] <Compn> i forgot if i ever got a response.
[18:39:58 CET] <Compn> for money of course
[18:40:08 CET] <Compn> for some reason we get requests for samples on hd
[18:40:15 CET] <Compn> once in a while
[18:41:42 CET] <ethe> As in, people ask FFmpeg for sample outputs on a hard drive?
[18:42:00 CET] <JEEB> probably something like the FATE samples archive
[18:42:02 CET] <Compn> for a copy of http://samples.ffmpeg.org on a hd.
[18:42:08 CET] <ethe> oh right
[18:42:25 CET] <Compn> i asked this guy what price he wants to pay and what country he is in
[18:58:52 CET] <J_Darnley> Last time you asked I lacked a spare harddrive, the samples, and you never said how quickly the guy wanted it.
[18:58:59 CET] <J_Darnley> I still lack those first two.
[19:30:42 CET] <rcombs> what, throttled internet service?
[20:21:47 CET] <Compn> J_Darnley : the person never responded.
[20:21:59 CET] <Compn> rcombs : mostly i'd guess people are just confused by the README file
[20:22:05 CET] <Compn> into thinking thats the only way to get a copy
[20:22:09 CET] <Compn> some people are just weird though
[20:22:13 CET] Action: rcombs points at wget
[20:22:35 CET] <rcombs> (this I say without looking at the readme)
[20:22:44 CET] <Compn> i'd be more afraid of not having the actor age information on any porn clips ... which is why i dont send hds full of clips :P
[20:22:50 CET] <Compn> from the uuusssayyy
[20:49:32 CET] <RiCON> rcombs: the readme says "if you're thinking of downloading the whole thing, please limit yourself to 50kB/s"
[20:58:19 CET] <ethe> rcombs: rsync is suggested as well (over wget I presume)
[21:33:00 CET] <wm4> lol drama
[21:34:38 CET] <J_Darnley> now what?
[21:38:12 CET] <llogan> wm4: does it have to do with a CoC?
[21:38:24 CET] <atomnuker> nope
[21:39:25 CET] Action: llogan is too hungover to care about drama
[21:42:50 CET] <llogan> michaelni: Can you take a look at "Spam Filter Rule 3" in cvslog? I think if I set it to Discard then that should stop spam. i had it set to hold just for testing.
[21:45:55 CET] <michaelni> llogan, i suggest to leave it on "hold" for a bit longer to ensure nothing valuable gets discarded
[21:46:45 CET] <llogan> i'll give it a week on Hold then.
[21:46:54 CET] <michaelni> thx
[21:49:49 CET] <llogan> michaelni: also, i removed ffmpeg-cvslog at ffmpeg from auto hold due to filter #3
[21:55:25 CET] <kierank> BBB: https://software.intel.com/en-us/articles/google-vp9-optimization
[21:55:27 CET] <kierank> (lol)
[21:57:12 CET] <JEEB> intel poking libvpx, I see
[21:59:42 CET] <wm4> is it still slower than lavc?
[22:00:46 CET] <kierank> ah they are optimising for atom
[22:00:50 CET] <kierank> where pshufb is sloooooow
[22:03:29 CET] <BBB> hm...
[22:03:31 CET] <BBB> so...
[22:03:39 CET] <BBB> where is the numbers?
[22:03:52 CET] <BBB> I mean, all I see is car sales terms like very, significant, impressive and lasting"
[22:03:58 CET] <kierank>  The team improved user-level performance by up to 16 percent (6.2 frames per second) in 64-bit mode and by about 12 percent (1.65 frames per second) in 32-bit mode. 
[22:04:00 CET] <BBB> horizon, expanded"
[22:04:10 CET] <BBB> ah, ok, so its hidden somewhere
[22:04:15 CET] <BBB> 32bit? omg
[22:04:17 CET] <BBB> poor people
[22:04:24 CET] <BBB> so wheres the results versus ffvp9? :-p
[22:04:39 CET] <BBB> last time I checked we were 30% faster on x86-64
[22:04:41 CET] <JEEB> they were probably just told to optimize chromium's decoding
[22:04:47 CET] <JEEB> so welp
[22:05:06 CET] <BBB> chromeos?
[22:05:11 CET] <kierank> no it's for android
[22:05:11 CET] <BBB> I mean, why else bother with 32bit
[22:05:13 CET] <kierank> in software
[22:05:15 CET] <BBB> ah
[22:05:17 CET] <kierank> on atoms on phones
[22:05:33 CET] <JEEB> yeah, android is IIRC 32bit user space + 64bit kernel on intel atm
[22:05:36 CET] <BBB> then use ffvp9 *ding ding*
[22:05:54 CET] <JEEB> that would make sense and require actual work on switching components in Chromium :D
[22:07:03 CET] <BBB> hiring intel to optimize libvpx is less effort than switching chrome to ffvp9 as they did for ffvp8?
[22:07:11 CET] <BBB> maybe
[22:07:26 CET] <nevcairiel> maybe intel volunteered to push  atom performance
[22:09:23 CET] <jamrial> maybe they could do the same for ffvp9. all of the atom optimized simd functions were added some time ago for a couple decoders, then nobody with it bothered again
[22:10:17 CET] <jamrial> "pmaddubsw xmm1, xmm16" uhh
[22:10:21 CET] <jamrial> wonder if whoever wrote that article has been playing with avx512 too much :p
[22:12:12 CET] <BBB> atom :-p
[22:17:12 CET] <jkqxz> That's the first time I've ever seen the name "Tremont".
[22:18:23 CET] <jkqxz> Ha, and Google agrees <https://www.google.co.uk/#q=intel+goldmont+%22tremont%22>...
[22:20:50 CET] <Gramner> possibly the 10nm shrink then
[22:24:08 CET] <jkqxz> Or just another "we haven't managed to make any significant changes yet but marketing demands a new version" release.
[22:24:56 CET] <BBB> I remember talking to some intel engineers at the vp9 summit last year
[22:25:09 CET] <Gramner> true. could be a "refresh" with some minor bug fixes/tweaks and a new codename slapped on
[22:25:16 CET] <BBB> I tried to explain to them some of the techniques that make ffvp9 much faster than libvpx (they were generally working on libvpx optimizations back then)
[22:25:26 CET] <BBB> they didnt understand all too much of it, sadly :(
[22:36:15 CET] <cone-609> ffmpeg 03KO Myung-Hun 07release/3.0:9e4d297ba74c: MAINTAINERS: add myself as an OS/2 maintainer
[22:36:15 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:c3e7a7ef3f85: swscale/x86/output: Move code into yuv2planeX_mainloop
[22:36:15 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:be5acd6cb1f3: swscale/x86/output: Fix yuv2planeX_16* with unaligned destination
[22:36:15 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:c4765a41b956: avcodec/h264: Execute error concealment before marking the frame as done.
[22:36:15 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:bd79dbfa2b8f: avutil/pixdesc: Make get_color_type() aware of CIE XYZ formats
[22:36:16 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:ca5c6394461e: swscale/input: Fix GBRAP16 input
[22:36:16 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:fb8676eb1cdd: swscale/utils: Fix chrSrcHSubSample for GBRAP16
[22:36:17 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:0d7343f8ddf6: avcodec/mjpegdec: Fix decoding slightly odd progressive jpeg
[22:36:17 CET] <cone-609> ffmpeg 03Rodger Combs 07release/3.0:bf8f2fae2ad8: lavf/mov: fix sidx with edit lists (cherry picked from commit 3617e69d50dd9dd07b5011dfb9477a9d1a630354)
[22:36:18 CET] <cone-609> ffmpeg 03Rodger Combs 07release/3.0:3b179b630253: lavf/mov: downgrade sidx errors to non-fatal warnings; fixes trac #5216 (cherry picked from commit 22dbc1caaf13e4bb17c9e0164a5b1ccaf490e428)
[22:36:19 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:b4be953c5596: avformat/cache: Fix memleak of tree entries
[22:36:20 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:27a61302b792: fate/aac: Increase fuzz from of fate-aac-pns-encode from 72 to 74 for Loongson
[22:36:21 CET] <cone-609> ffmpeg 03foo86 07release/3.0:da6ccfe18ebb: avcodec/dca: clear X96 channels if nothing was decoded
[22:36:22 CET] <cone-609> ffmpeg 03Boris Nagels 07release/3.0:56c4dca5ae78: avformat/rtpenc: Fix integer overflow in NTP_TO_RTP_FORMAT
[22:36:23 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:756f4b586ae7: avformat/utils: fix dts from pts code in compute_pkt_fields() during ascending delay
[22:36:24 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:64f2b48be062: avformat/concatdec: set safe mode to enabled instead of auto
[22:36:25 CET] <cone-609> ffmpeg 03Martin Cracauer 07release/3.0:934bc32accae: avutil/channel_layout: AV_CH_LAYOUT_6POINT1_BACK not reachable in parsing
[22:36:26 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:c8977ec6e9b0: avutil/random_seed: Add the runtime in cycles of the main loop to the entropy pool
[22:36:27 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:e1589367128a: avformat/hlsenc: Fix passing options, regression since bc9a5965c815cf7fd998d8ce14a18b8e861dd9ce
[22:36:28 CET] <cone-609> ffmpeg 03Luca Barbato 07release/3.0:b0e836466f2a: indeo2data: K&R formatting cosmetics
[22:36:29 CET] <cone-609> ffmpeg 03Luca Barbato 07release/3.0:53fffc9fc4f2: indeo2: Fix banding artefacts
[22:36:30 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:bf80b1d88d39: avcodec/resample: Remove disabled and faulty code
[22:36:31 CET] <cone-609> ffmpeg 03Mark Thompson 07release/3.0:d6adcab84280: lavc/hevc: Allow arbitrary garbage in bytestream as long as at least one NAL unit is found.
[22:36:32 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:50a639a62aa3: avcodec/mjpegenc_common: Store approximate aspect if exact cannot be stored
[22:36:33 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:0fae52d7e37a: avformat/file: Add crypto to default whitelist
[22:36:34 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:e15a48d35afa: avformat/mpegtsenc: Keep track of the program for each service
[22:36:35 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:6e5648ad42b7: avformat/mpegtsenc: Fix used service
[22:36:36 CET] <cone-609> ffmpeg 03Ico Doornekamp 07release/3.0:daa6dc0a3b39: avformat/rtpdec_jpeg: fix low contrast image on low quality setting
[22:36:37 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:2d0e415482ca: avcodec/libutvideodec: copy frame so it has reference counters when refcounted_frames is set
[22:49:37 CET] <cone-609> ffmpeg 03Michael Niedermayer 07release/3.0:526c7b21865f: Update for 3.0.1
[22:55:05 CET] <jamrial> atomnuker: you didn't backport any of the aac or vc2 patches yet
[00:00:00 CET] --- Sun Mar 27 2016


More information about the Ffmpeg-devel-irc mailing list