[Ffmpeg-devel-irc] ffmpeg-devel.log.20151022

burek burek021 at gmail.com
Fri Oct 23 02:05:03 CEST 2015

[00:43:22 CEST] <cone-848> ffmpeg 03Andreas Cadhalpun 07master:957f92729a3e: avcodec: drop 2 suffix from avpriv_ac3_parse_header2
[00:43:24 CEST] <cone-848> ffmpeg 03Andreas Cadhalpun 07master:a4680e71707e: avcodec: rename avpriv_color_frame to ff_color_frame
[01:49:45 CEST] <Timothy_Gu> kierank: thx for the heads up. dont think ill be using it
[01:56:31 CEST] <kagami_> Hi. Is it possible to fix pkt_duration inside filter? I'm setting different value but it still doesn't change in the output. (I hope this is the appropriate channel for the question.)
[02:13:26 CEST] <cone-848> ffmpeg 03Alexis Ballier 07master:94ec82f14ab1: libavformat/mxfdec.c: cosmetics: Add missing space after '?' in log message.
[02:13:27 CEST] <cone-848> ffmpeg 03Alexis Ballier 07master:12628e3369de: libavformat/mxfdec.c: Report dark metadata keys only when they match no parser at all, not everytime they fail to match one.
[04:22:13 CEST] <atomnuker> was ff_clz removed?
[04:23:57 CEST] <atomnuker> nvm, it never existed, was thinking of ff_ctz
[04:56:31 CEST] <pross> i have forgotten Paul Mahols nick
[04:57:22 CEST] <pross> Paul: the samples you sent, were not relevant to your patch!
[05:01:40 CEST] <atomnuker> pross: he's durandal_1707
[05:01:58 CEST] <cone-848> ffmpeg 03Timothy Gu 07master:ab5f43e6342c: vc1dsp: Port ff_vc1_put_ver_16b_shift2_mmx to yasm
[05:10:22 CEST] <Compn> pross : durandal_1707 is paul i think
[05:15:30 CEST] <pross> thx
[05:23:47 CEST] <durandal_1707> pross: how so?
[05:24:10 CEST] <cone-848> ffmpeg 03Marton Balint 07master:a1240c0522d1: ffmpeg: log failed av_write_trailer
[05:24:11 CEST] <cone-848> ffmpeg 03Marton Balint 07master:f4730a584542: ffmpeg: exit on av_write_trailer failure if exit_on_error is set
[05:24:12 CEST] <cone-848> ffmpeg 03Marton Balint 07master:e2b416b68e85: ffmpeg: factorize checking decoder result
[05:24:13 CEST] <cone-848> ffmpeg 03Marton Balint 07master:b69b43e2c471: ffmpeg: exit on corrupt packets or decoded frames if exit_on_error flag is present
[08:59:23 CEST] <cone-206> ffmpeg 03James Almer 07master:73353af6e559: x86/Makefile: move decoder/encoder objects out of the subsystems section
[10:11:46 CEST] <ubitux> http://ffmpeg.org/pipermail/ffmpeg-devel/2015-October/181803.html
[10:11:50 CEST] <ubitux> someone please explain :(
[10:13:25 CEST] <flux> yes, please add those presets. k thx bai.
[10:44:47 CEST] <wm4> _julian: hi
[10:45:01 CEST] <_julian> wm4: hi :)
[10:46:36 CEST] <cone-206> ffmpeg 03Julian Scheel 07master:7ebe12fc5559: mmaldec: Add mpeg2 decoding support
[10:57:40 CEST] <kagami_> Could someone at least give a hint regarding pkt_duration? It just shouldn't work that way or I'm doing something wrong?
[11:01:58 CEST] <cone-206> ffmpeg 03Paul B Mahol 07master:0d7c02748335: avformat/electronicarts: support ADPCM PSX
[11:01:59 CEST] <cone-206> ffmpeg 03Paul B Mahol 07master:6debfce6a31b: avformat/electronicarts: fix demuxing of certain eam files
[11:11:00 CEST] <_julian> wm4: are you running on Pi or Pi2, btw?
[11:12:16 CEST] <wm4> pi2
[11:13:18 CEST] <nevcairiel> (everyone should be, its like 200x faster)
[11:23:45 CEST] <cone-206> ffmpeg 03Carl Eugen Hoyos 07master:c52c78cc568f: lavf/mpjpegdec: Return 0 if an allocation inside the probe function fails.
[11:38:32 CEST] <cone-206> ffmpeg 03Carl Eugen Hoyos 07master:775b84e30ef2: lavc/x86/vc1dsp_init: Fix compilation with --disable-yasm.
[11:38:39 CEST] <nevcairiel> man what happened with git 2.6, rebasing has gotten like 20 times faster
[11:57:02 CEST] <JEEB> nevcairiel: wow
[11:57:52 CEST] <durandal_1707> ok to push shuffleframes?
[12:27:35 CEST] <maurobaraldi> Hello world. I have a question. Does the ffmpeg load entire file when read? Where can I find it on source code?
[12:36:04 CEST] <wm4> maurobaraldi: for most types of files it usually doesn't
[12:40:19 CEST] <maurobaraldi> But it open only once? I mean, it doesn't not open and reopen along the process?
[12:43:29 CEST] <wm4> usually not
[12:56:51 CEST] <maurobaraldi> Thanks!
[13:13:16 CEST] <cone-206> ffmpeg 03Peter Ross 07master:cdfc61de4eea: libavformat/electronicarts: also demux mpeg audio layer 2
[14:21:08 CEST] <cone-206> ffmpeg 03Michael Niedermayer 07master:00efaa798377: avutil/intmath: fix undefined behavior in ff_ctzll_c()
[15:47:32 CEST] <cone-206> ffmpeg 03Anton Khirnov 07master:772c87c5a658: qsvenc: support passing arbitrary external buffers to the encoder
[15:47:33 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:760dbdd3c561: Merge commit '772c87c5a658f36d7c0612f5da583fc2bfa54f79'
[15:49:55 CEST] <cone-206> ffmpeg 03Anton Khirnov 07master:f6f32fc93d39: qsvenc: set the timestamp for PIX_FMT_QSV frames as well
[15:49:56 CEST] <cone-206> ffmpeg 03Anton Khirnov 07master:2ec96b6bd5bf: qsvenc: cosmetics, reindent
[15:49:57 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:2fece1e4d5ee: Merge commit 'f6f32fc93d39caf329869c1bff8ad024ccab1d42'
[15:49:58 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:a440886a35e3: Merge commit '2ec96b6bd5bf7b22978711bcf2cee702bee89c6f'
[15:57:26 CEST] <cone-206> ffmpeg 03Anton Khirnov 07master:dc923bc23b3e: qsvenc: add an API for allocating opaque surfaces
[15:57:27 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:866a4174dba0: Merge commit 'dc923bc23b3efd949d0bf67ff1abdb95059e5843'
[16:23:13 CEST] <cone-206> ffmpeg 03Anton Khirnov 07master:fb472e1a11a4: avconv: add support for Intel QSV-accelerated transcoding
[16:23:14 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:dd8a4b0f8c39: Merge commit 'fb472e1a11a4e0caed2c3c91da01ea8e35d9e3f8'
[17:02:58 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:b54d64511607: qsvdec: fix get_format and hwaccel_context handling
[17:59:55 CEST] <mateo`> hello there o/ regarding the hwaccel api, is it possible for the end_frame to return no frame ? is so, which code is supposed to be returned ? EAGAIN leads hwaccel to fail
[18:00:19 CEST] <nevcairiel> hwaccel is always supposed to return a frame, or return a failure code
[18:00:52 CEST] <nevcairiel> it operates at a slice level, it gets handed all the slices in one frame, and its supposed to decode those and only those
[18:00:57 CEST] <nevcairiel> if it cant, it should return an error
[18:01:01 CEST] <nevcairiel> but nothing else is valid
[18:01:31 CEST] <mateo`> meh
[18:02:37 CEST] <mateo`> but if the hw api needs more input to actually output a frame ?
[18:02:49 CEST] <nevcairiel> thats not something it can decide
[18:03:04 CEST] <nevcairiel> sounds like you dont want to use the hwaccel API if thats questions you have to ask
[18:04:23 CEST] <wm4> mateo`: if the hw API doesn't match the ffmpeg hwaccel framework, you'll have to create a separate decoder for it
[18:04:28 CEST] <wm4> like e.g. qsv does
[18:05:46 CEST] <mateo`> i guess i will have to do that
[18:06:24 CEST] <nevcairiel> the hwaccel API is designed for the "traditional" hwaccels which all work on a slice level
[18:06:37 CEST] <nevcairiel> like VDPAU, VAAPI and DXVA
[18:07:06 CEST] <nevcairiel> many "newer" and vendor specific APIs work on a bitstream level instead and want the full thing, which just doesnt fit into there
[18:09:31 CEST] <mateo`> good to know ..., I would have been more inspired to closely look at how it is working before rushing the mediacodec support over there ^^
[18:12:42 CEST] <mateo`> wm4, nevcairiel: thanks for the information
[19:02:44 CEST] <cone-206> ffmpeg 03Marton Balint 07master:5821244b225b: ffmpeg: fix ffmpeg.h trailing whitespace
[19:12:03 CEST] <cone-206> ffmpeg 03Marton Balint 07master:ddc6bd8c9586: ffmpeg: add abort_on option to allow aborting on empty output
[19:42:22 CEST] <cone-206> ffmpeg 03Vittorio Giovara 07master:ac981d16415e: APIchanges: Fill in missing dates and hashes
[19:42:23 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:e9a2ab3ac783: Merge commit 'ac981d16415e7fd99683e10297781c7d9ec1a8cd'
[19:44:05 CEST] <cone-206> ffmpeg 03Vittorio Giovara 07master:3496cec43304: msrle: Use AVFrame instead of AVPicture
[19:44:06 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:d3cc258a6167: Merge commit '3496cec43304ac040d6d05f9d500a6f93cc049e7'
[19:52:57 CEST] Action: durandal_1707 started RE xma codec
[20:02:01 CEST] <wm4> nevcairiel: seems like nobody is able to review your schannel patch
[20:02:13 CEST] <nevcairiel> of course not
[20:02:15 CEST] <nevcairiel> its windows code
[20:02:50 CEST] <cone-206> ffmpeg 03Luca Barbato 07master:fcc1280acb6e: rawenc: Replace avpicture functions with imgutils
[20:02:51 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:6d34ab1a08c0: Merge commit 'fcc1280acb6e6f682b34c2101b075b82f83d71ba'
[20:09:09 CEST] <nevcairiel> and no actual feedback on hwaccel+mt either yet, other than "vlc uses it", well duh, i knew that :)
[20:09:19 CEST] <cone-206> ffmpeg 03Luca Barbato 07master:ef3a3519c106: rawdec: Replace avpicture functions with imgutils
[20:09:20 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:9a6f1eea3e1d: Merge commit 'ef3a3519c10620c4206738595bf03fc0bed71802'
[20:09:39 CEST] <lotharkript> any one interested to review this "[PATCH] Adds support parsing the QuickTime Metadata Keys." 
[20:11:33 CEST] <jamrial> nevcairiel: if you get no actual feedback or concrete arguments agains the hwaccel patch then just push it
[20:11:47 CEST] <J_Darnley> All I have to say about that is who thought it would be a good idea to store floats as metadata values.
[20:12:21 CEST] <nevcairiel> jamrial: thats what I'll do in a couple days, dont want to rush it and get blamed later
[20:12:44 CEST] <nevcairiel> my vp9 hwaccel is kidna waiting for it since its somewhat broken with mt
[20:13:34 CEST] <nevcairiel> i'll probably write a proper commit message tomorrow and put it on the ML again, if i dont get anything back then either, thats that
[20:15:45 CEST] <llogan> lotharkript: if nobody replies after a few days you can reply with "ping". we sometimes just forget.
[20:18:25 CEST] <Daemon404> ill look at it.
[20:19:12 CEST] <lotharkript> thanks.. my coworker is on my back :-(
[20:24:14 CEST] <cone-206> ffmpeg 03Vittorio Giovara 07master:f0a106578d75: roqvideodec: Replace avpicture functions with imgutils
[20:24:15 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:42b87ed01eb3: Merge commit 'f0a106578d759de6183eea3c75f8373b6d3153c1'
[20:24:16 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:ef4fbee79df6: roqvideodec: use av_frame_copy
[20:31:30 CEST] <cone-206> ffmpeg 03Vittorio Giovara 07master:48c063868316: dpx: Replace avpicture functions with imgutils
[20:31:32 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:8fe1433ad4f8: Merge commit '48c06386831604921bdaf4fb77ea02766cd615f4'
[20:34:51 CEST] <cone-206> ffmpeg 03Vittorio Giovara 07master:13bddab7de10: nuv: Replace avpicture functions with imgutils
[20:34:52 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:422fab7fbe51: Merge commit '13bddab7de10aebf6efb98aa6d7ff0c51bb0e364'
[20:41:41 CEST] <kierank> Daemon404: surely it's a $9.99 subscription/mon to look at patches now =p
[20:44:08 CEST] <Daemon404> lotharkript, review sent
[20:44:48 CEST] <Daemon404> kierank, harr harr
[20:45:17 CEST] <cone-206> ffmpeg 03Lou Logan 07master:4c46f1d493f4: avfilter/vf_zscale: fix typo
[20:45:27 CEST] <Daemon404> oho zscale went in
[20:45:31 CEST] <Daemon404> neat.
[20:46:06 CEST] <nevcairiel> great now i have to rebase my merge
[20:46:10 CEST] Action: nevcairiel curses
[20:46:47 CEST] <nevcairiel> thank the git gods for rerere
[20:46:55 CEST] <Daemon404> for what now
[20:47:23 CEST] <nevcairiel> rerere!
[20:47:31 CEST] <nevcairiel> it remembers conflict resolutions
[20:47:45 CEST] <Daemon404> sounds like an anime
[20:47:46 CEST] <Daemon404> <_<
[20:47:53 CEST] <nevcairiel> and since you cant really rebase merges, you have to re-do them
[20:48:07 CEST] <Daemon404> or merge again.
[20:48:11 CEST] <nevcairiel> but as long as no new conflicts appear, rerere does that for you
[20:48:15 CEST] <Daemon404> just in the opposite directin
[20:48:16 CEST] <J_Darnley> (Why isn't your git history a straight line?)
[20:48:45 CEST] <nevcairiel> cant merge libav changes into a straight line =p
[20:50:01 CEST] <cone-206> ffmpeg 03Vittorio Giovara 07master:f890677d05bc: Replace any remaining avpicture function with imgutils
[20:50:02 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:470204218f37: Merge commit 'f890677d05bc4e8b494a73373ab4cc19791bf884'
[20:50:49 CEST] <llogan> nevcairiel: heh, sorry. yet another one of my super useful commits.
[20:51:05 CEST] <nevcairiel> i'm just happy that i found out rerere exists
[20:57:51 CEST] <lotharkript> Deamon404: thanks. Will let my co-worker know. 
[21:02:42 CEST] <Plorkyeran> git rebase --preserve-merges works for simple merges
[21:02:50 CEST] <Plorkyeran> but maybe the merges from libav aren't simple...
[21:03:57 CEST] <nevcairiel> all that does is try to re-apply the merge after the rebase
[21:04:08 CEST] <nevcairiel> it doesnt remember any of your changes or conflict resolutions
[21:04:58 CEST] <NitLord> think about that makes me feel like a real gitard.
[21:05:25 CEST] <NitLord> so does reading what I type.
[21:32:48 CEST] <cone-206> ffmpeg 03Michael Niedermayer 07master:ea5a1d148530: avcodec/x86/vc1dsp: Remove unused macro
[21:32:49 CEST] <cone-206> ffmpeg 03Michael Niedermayer 07master:37498a4b202e: avcodec/nuv: Fix 'libavcodec/nuv.c:83:19: warning: passing argument 3 of av_image_copy from incompatible pointer type'
[21:43:15 CEST] <cone-206> ffmpeg 03Vittorio Giovara 07master:a17a7661906b: lavc: Add data and linesize to AVSubtitleRect
[21:43:16 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:ee573b4d31d4: Merge commit 'a17a7661906ba295d67afd80ac0770422e1b02b3'
[21:44:03 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:ce9736362896: avcodec/libzvbi: Update for AVSubtitleRect changes
[21:46:27 CEST] <cone-206> ffmpeg 03Vittorio Giovara 07master:ff7956fcbf8e: avplay: Replace avpicture functions with imgutils
[21:46:28 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:3c3997188ef9: Merge commit 'ff7956fcbf8e59b21654b95038de3ed88a850a9e'
[21:48:05 CEST] <nevcairiel> anyone feel like updating ffplay to no longer use AVPicture? its used all over the place
[21:49:13 CEST] <ubitux> only for bitmap subs, isn't it?
[21:49:13 CEST] <cone-206> ffmpeg 03Vittorio Giovara 07master:3ee2c60cc296: utils: Use data buffers directly instead of an AVPicture
[21:49:14 CEST] <cone-206> ffmpeg 03Hendrik Leppkes 07master:dde8e5ad02ad: Merge commit '3ee2c60cc296eee3f63d7b5fee9b4332eeeac9fa'
[21:49:29 CEST] <nevcairiel> ubitux: also for generic video things, from the looks of it
[21:50:31 CEST] <nevcairiel> (also, i hope none of the bitmap sub decoders broke now)
[21:51:27 CEST] <jamrial> nevcairiel: poke marton balint. he's the main ffplay developer afaik
[21:59:15 CEST] <ubitux> erm
[21:59:33 CEST] <ubitux> isn't the AVSubtitleRect breaking ABI?
[21:59:40 CEST] <ubitux> so we're still in that period?
[22:02:27 CEST] <nevcairiel> noone has specifically closed it yet i su ppose
[22:02:42 CEST] <nevcairiel> and barely anyone bothered to answer my mail on the subject
[22:02:44 CEST] <nevcairiel> so shrug
[22:04:17 CEST] <nevcairiel> i do wonder why koda didnt put the AVSubtitleRect.pict compat code into the main decode subtitle function though
[22:15:55 CEST] <cone-206> ffmpeg 03Ganesh Ajjanagadde 07master:8507b98c10d9: avfilter,swresample,swscale: use fabs, fabsf instead of FFABS
[22:19:34 CEST] <wm4> nevcairiel: me too
[22:36:34 CEST] <RiCON> nevcairiel: schannel with gmp for rtmpe seems to be working fine with mpv
[22:36:40 CEST] <RiCON> using your patch and https://raw.githubusercontent.com/jb-alvarado/media-autobuild_suite/master/build/patches/ffmpeg-0004-configure-Add-options-to-use-gmp.patch
[22:56:32 CEST] <BBB> durandal_1707: which docs?
[22:56:42 CEST] <BBB> durandal_1707: I dont tihnk the last line of the av_log output is documented
[22:57:15 CEST] <BBB> durandal_1707: or maybe it is and Im blind?
[23:00:29 CEST] <durandal_1707> I meant if its put in stats file
[23:00:48 CEST] <durandal_1707> But it looks like not
[23:11:01 CEST] <cone-206> ffmpeg 03Michael Niedermayer 07master:3f85552e402c: avcodec/libzvbi-teletextdec: Add variable to fix build
[23:16:50 CEST] <BBB> durandal_1707: yeah I didnt want to update that just yet, this is just the log
[23:16:57 CEST] <BBB> durandal_1707: so patch ok w/o doc adjustments then?
[23:17:40 CEST] <durandal_170> BBB: yes
[23:19:04 CEST] <BBB> ok, ty
[23:35:29 CEST] <kierank> llogan: that larbi guy
[23:35:36 CEST] <kierank> sent me emails, added me on linkedin
[23:35:39 CEST] <kierank> ...
[23:37:11 CEST] <llogan> kierank: heh. you are now friends.
[23:37:59 CEST] <atomnuker> how'd you get yourself a stalker online?
[23:38:53 CEST] <llogan> talking about stalking, when do you think experimental flag will be removed?
[23:39:16 CEST] <nevcairiel> atomnuker needs to stalk claudio more, apparently
[23:41:08 CEST] <nevcairiel> Last I heard we were waiting for a few small fixes of his
[23:41:33 CEST] <kierank> atomnuker: that's what people do
[23:41:48 CEST] <kierank> not sure why
[23:42:32 CEST] <cone-206> ffmpeg 03Ronald S. Bultje 07master:81d7f118e20c: vf_psnr: remove %0.2f format specifiers for stream summary line.
[23:42:34 CEST] <cone-206> ffmpeg 03Ronald S. Bultje 07master:e7d9061d4f93: vf_ssim: print per-channel dB values.
[23:43:09 CEST] <llogan> it takes a certain type of person to send their questions to ffmpeg-*-owner at ffmpeg.org then then email the list asking to unsubscribe
[23:44:27 CEST] <atomnuker> nevcairiel: he needs a few more days, and I think I'll wait, basically we found a huge retarded oversight which was what ruined the new coder on some techno tracks and need time to do regression testing
[23:44:28 CEST] <kierank> llogan: you haven't seen the facebook ones
[23:44:44 CEST] <kierank> either they pm you or they complain on the ffmpeg page
[23:44:45 CEST] <wm4> isn't facebook all about stalking
[23:44:50 CEST] <llogan> i'm surprised there aren't more help requests on twitter
[23:45:15 CEST] <kierank> I wonder if you could get twitter verified
[23:45:41 CEST] <llogan> AFAIK/IIRC they just have to anoint you with that on their own regard.
[23:45:53 CEST] <TD-Linux> ffmpeg has a facebook page?
[23:46:05 CEST] <nevcairiel> atomnuker: we've been waiting years, we can hold out a couple more days/weeks
[23:46:07 CEST] <nevcairiel> but not months
[23:46:10 CEST] <nevcairiel> just to be clear
[23:46:10 CEST] <nevcairiel> :D
[23:46:13 CEST] <llogan> TD-Linux: fcaebook/ffmpeg.org I think
[23:48:56 CEST] <atomnuker> nevcairiel: well if you need to know the details IS marked bands and replaced their scalefactor indices but since those are transmitted as the difference between the previous and the next value IS was being an inconsiderate prick and resulted in sometimes an assertion being triggered and completely messing up the decoder even though the IS bands were fine on their own
[23:49:35 CEST] <atomnuker> yeah, that's a very definite derp
[23:51:15 CEST] <atomnuker> this basically resulted in warbling heard on some synth-heavy tracks and at low bitrates manifested itself as a freaking 440ish Hz tone (dependant on samplerate)
[23:51:22 CEST] <nevcairiel> unrelated: how do i download stuff from the incoming ftp again? I cant remember, or remember where to find the info
[23:54:15 CEST] <wm4> you need to login with a secret login
[23:54:21 CEST] <wm4> (which everyone knows)
[23:56:34 CEST] <llogan> the silence is a result of nobody knowing.
[23:57:01 CEST] <llogan> i asked carl once, then also lost/forgot it and didn't have the foresight to write it down
[23:59:21 CEST] <wm4> so was the login changed?
[23:59:38 CEST] <BBB> durandal_1707: that psnr stdout stats patch is broken btw
[00:00:00 CEST] --- Fri Oct 23 2015

More information about the Ffmpeg-devel-irc mailing list