[Ffmpeg-devel-irc] ffmpeg-devel.log.20160104

burek burek021 at gmail.com
Tue Jan 5 02:05:02 CET 2016


[00:01:28 CET] <cehoyos> I prefer to call them thieves and liars but some seem to dislike this;-)
[00:02:45 CET] <durandal_170> stop talking nonsense
[00:03:05 CET] <rcombs> the dark side of the force
[00:03:14 CET] <rcombs> (we have cookies!)
[00:03:24 CET] <cehoyos> What nonsense?
[00:03:38 CET] Action: rcombs gives cehoyos a cookie
[00:03:45 CET] <cehoyos> Thx!
[00:04:38 CET] <durandal_170> you are brown side
[00:05:09 CET] <rcombs> it's sort of a grey area
[00:05:27 CET] <durandal_170> :)
[00:05:39 CET] <cehoyos> The copyright violations? Or the claims on their homepage? Or the MPlayer infrastructure?
[00:06:07 CET] <rcombs> the cookies
[00:06:51 CET] <durandal_170> trocadero
[00:08:20 CET] <durandal_170> I see nothing wrong with foo decoder
[00:38:13 CET] <Compn> is it the uk requiring cookies
[00:38:15 CET] <Compn> or the whole eu ?
[00:38:17 CET] <Compn> anyone know?
[00:38:25 CET] <Compn> because i'm pissed at these cookie notices on all these websites
[00:38:39 CET] <Compn> its how the internet has worked for hte last 18+ years you damn cookie monsters
[00:38:43 CET] <Compn> REEEEEEEEEEEEEEEEEEEEEEEEE
[00:39:32 CET] <J_Darnley> It was the EU we can thank for that.
[00:39:49 CET] <Compn> great.
[00:40:17 CET] <atomnuker> so what, no one can touch you if you don't show the notice
[00:40:30 CET] <Compn> too many double negatives atomnuker 
[00:41:04 CET] <atomnuker> double negatives enhance the negative effect in many languages
[00:42:05 CET] <atomnuker> also, MEGA just dumps the file into browser cache and you browser then copies the file from there AFAIK
[00:42:10 CET] <Compn> i know
[00:42:21 CET] <atomnuker> so that's far worse than cookies yet no notices or anything
[00:42:22 CET] <Compn> that mega cache is tasking the memory leaks in many web browsers hahah
[00:42:32 CET] <Compn> web browser can barely hang on :D
[01:48:28 CET] <Compn> wm4 : does audio over hdmi work in mpv ?
[01:48:31 CET] <Compn> linux
[01:48:57 CET] <Compn> asking for a user, i should probably get some hdmi equipment...
[01:49:04 CET] <Compn> sitting here back in 2003. :p
[01:50:45 CET] <atomnuker> if your HDMI output appears as a separate sound card output (which it should) it'll work with anything
[01:53:05 CET] <Compn> should work, yes
[01:53:12 CET] <Compn> betting on linux audio? mmm
[01:54:40 CET] <atomnuker> pulse and alsa never failed me
[01:56:22 CET] <Compn> never?
[01:56:35 CET] <Compn> do you watch films with a media player (if so which one?)
[01:57:16 CET] <atomnuker> mpv
[01:57:50 CET] <atomnuker> I've tried to remember an instance where pulseaudio broke (and it wasn't mine or my distro's fault) and I can't
[01:59:22 CET] <philipl> Compn: yes, it works. It's how I use it.
[01:59:25 CET] <philipl> the app doesn't care.
[01:59:49 CET] <Compn> cool
[02:45:25 CET] <cone-301> ffmpeg 03Michael Niedermayer 07master:4da2ac5c7a49: avcodec/h264: Fix regression caused by removial of default_ref_list
[03:43:14 CET] <cone-301> ffmpeg 03Mats Peterson 07master:bf42a7ef6d07: lavc/qtrle: Use AV_PIX_FMT_PAL8 for 1-bit video
[03:43:15 CET] <cone-301> ffmpeg 03Mats Peterson 07master:b6c61b7d43fa: lavf/qtpalette: Ignore greyscale bit in certain cases
[03:45:46 CET] <jamrial> huh, it made it in
[04:00:02 CET] <Compn> persistant, i would have given up long ago :D
[04:28:43 CET] <RiCON> i'd have given up right at the trac ticket
[09:39:41 CET] <j-b> 'morning
[09:58:41 CET] <durandal_170> 'evening
[10:48:19 CET] <mateo`> hello o/
[10:49:43 CET] <wm4> michaelni: is there any deeper reason why yadif doesn't support nv12?
[11:17:30 CET] <ubitux>  
[11:17:35 CET] <ubitux> oups
[11:18:35 CET] <nevcairiel> wm4: other than the code not being written for it ? :D
[11:20:18 CET] <michaelni> wm4, no deep reason
[11:24:51 CET] <wm4> it somehow supports some 10/16 bit per channel formats, so it can't be just restrained to 1 byte per pixel
[11:25:08 CET] <wm4> but I have trouble seeing how it actually works
[11:25:28 CET] <nevcairiel> the interleaving probably needs some extra magic to be supported
[11:32:08 CET] <cone-546> ffmpeg 03Anton Khirnov 07master:9d74012761bc: h264: improve behaviour with invalid reference lists
[11:32:08 CET] <cone-546> ffmpeg 03Maxym Dmytrychenko 07master:a43905f4ae26: qsvenc_hevc: improve the default settings
[11:32:08 CET] <cone-546> ffmpeg 03Maxym Dmytrychenko 07master:3b6473b43eb6: qsvdec: properly handle the warning from MFXVideoCORE_SyncOperation
[11:32:08 CET] <cone-546> ffmpeg 03Hendrik Leppkes 07master:1fec347ef56b: Merge commit '9d74012761bc3ee676fe43321d5699e4877fde5b'
[11:32:08 CET] <cone-546> ffmpeg 03Hendrik Leppkes 07master:ad53c94d8fe0: Merge commit 'a43905f4ae261bdde87c300901d867b31961f57b'
[11:32:09 CET] <cone-546> ffmpeg 03Hendrik Leppkes 07master:9507f68deb66: Merge commit '3b6473b43eb69fc3faaf69f7fd0b83b51db7607f'
[11:32:58 CET] <cone-546> ffmpeg 03Diego Biurrun 07master:58170b027144: lavc: Add missing #includes for ff_dlog()
[11:33:00 CET] <cone-546> ffmpeg 03Hendrik Leppkes 07master:3e337f0e4014: Merge commit '58170b027144f4840018c7a60fb8d07924cc6242'
[11:34:20 CET] <ubitux> nevcairiel: not sure if it's because of a recent merge, but did you see http://fate.ffmpeg.org/log.cgi?time=20160104090113&log=compile&slot=x86_64-archlinux-gcc-ddebug ?
[11:41:10 CET] <cone-546> ffmpeg 03Diego Biurrun 07master:2080bea4a74b: h264_refs: Remove broken trace debug code
[11:41:11 CET] <cone-546> ffmpeg 03Hendrik Leppkes 07master:e1c5170c638e: Merge commit '2080bea4a74b6b1598caa806a11ea807ba546f94'
[11:41:19 CET] <nevcairiel> ubitux: ^^
[11:41:38 CET] <ubitux> lol ok thanks
[12:21:21 CET] <kurosu> hi
[13:08:30 CET] <cone-546> ffmpeg 03Andreas Cadhalpun 07master:e7a7b3135a4e: vorbisdec: reject rangebits 0 with non-0 partitions
[13:21:28 CET] <cone-546> ffmpeg 03Andreas Cadhalpun 07master:d7d37c479fa7: brstm: make sure an ADPC chunk was read for adpcm_thp
[13:21:29 CET] <cone-546> ffmpeg 03Andreas Cadhalpun 07master:bcf4ee26a0a1: brstm: also allocate b->table in read_packet
[13:49:28 CET] <cone-546> ffmpeg 03Andreas Cadhalpun 07master:1cb2331eca0d: brstm: fix missing closing brace
[14:07:36 CET] <cone-546> ffmpeg 03Paul B Mahol 07master:6defda1cdb41: avfilter/avf_showspectrum: add option to set gain for calculating color intensity
[18:30:11 CET] Action: Daemon404 pokes ubitux 
[18:30:23 CET] <ubitux> yup?
[18:30:39 CET] <Daemon404> do you have a comment on carl's decimate change
[18:31:13 CET] <ubitux> i need to look more closely, it's been a long time
[18:31:27 CET] <ubitux> i need to "sync" it with upstream since a long time
[18:31:55 CET] <ubitux> i'll check tonight (in ~3h probably)
[18:32:07 CET] <ubitux> but feel free to push if it's urgent
[18:32:11 CET] <Daemon404> it's not
[18:59:33 CET] <cone-546> ffmpeg 03Matthieu Bouron 07master:44913d19457d: swscale/arm/yuv2rgb: fix typo
[18:59:34 CET] <cone-546> ffmpeg 03Matthieu Bouron 07master:e4e9b9454e97: swscale/arm/yuv2rgb: add ff_yuv422p_to_{argb,rgba,abgr,bgra}_neon_{16,32}
[20:52:32 CET] <nevcairiel> I found a 32-bit FLAC file, which apparently libFLAC reference cant even decoder, but FFmpeg can .. wonder who or what produced this file
[20:53:24 CET] <ubitux> grep -i lavc bla.flac? :)
[20:53:42 CET] <nevcairiel> the metadata claims libflac encoded it
[20:53:49 CET] <nevcairiel> but why cant it decode it then
[20:54:27 CET] <TD-Linux> 32 bit flac files are not a thing
[20:54:46 CET] <nevcairiel> this one would disagree with you
[20:54:55 CET] <TD-Linux> how are you even determining that it is 32 bit?
[20:55:09 CET] <nevcairiel> the file encoded the bitdepth in the header
[20:56:30 CET] <TD-Linux> ah OK I see it is actually possible.
[20:59:28 CET] <nevcairiel> yeah streaminfo says 32-bit
[20:59:57 CET] <nevcairiel> now to figure out why the libflac decoder fails
[21:00:00 CET] <nevcairiel> maybe my wrapper is wonky
[21:00:17 CET] <J_Darnley> Did ffmpeg once mark 24-bit as 32-bit because it couldn't hold 24-bits from the input?
[21:00:38 CET] <J_Darnley> The reference only supports 24 bit
[21:00:45 CET] <TD-Linux> nevcairiel, I don't think libflac supports 32 bit. which doesn't explain the metadata.
[21:00:47 CET] <J_Darnley> (up to 24 that is)
[21:01:10 CET] <TD-Linux> from google, flake is the only encoder I could find that can output 32 bit
[21:01:12 CET] <nevcairiel> TD-Linux: they wouldnt be as crazy as supporting 32-bit encoding but not decoding, would they
[21:01:35 CET] <TD-Linux> nevcairiel, maybe it's someone's hacked up libflac?
[21:02:21 CET] <nevcairiel> TD-Linux: i suppose thats a possibility, but if the reference cant even decoder it, handing out such files would feel odd
[21:03:13 CET] <nevcairiel> apparently its from some crazy website that sells these in s ome form
[21:03:47 CET] <TD-Linux> nevcairiel, indeed. I originally said 32-bit flac files weren't a thing because I knew libflac didn't support them
[21:04:07 CET] <TD-Linux> I suspect the standardization of flac in CELLAR will explicitly ban them
[21:04:11 CET] <nevcairiel> Yeah I remembered that about libflac, but ffmpeg flac decoder supports it
[21:04:18 CET] <nevcairiel> sounds fine at least
[21:04:21 CET] <nevcairiel> no clue if lossless
[21:06:53 CET] <nevcairiel> 32-bit integer audio for distribution is just silly anyway
[21:07:04 CET] <nevcairiel> but you know, users of a media player expect their media to play
[21:07:34 CET] <wm4> nevcairiel: that's such a cliched and outdated point of view
[21:07:43 CET] <wm4> media players should be able to browse the web
[21:07:57 CET] Action: wm4 goes trolling somewhere else
[21:08:10 CET] <Compn> winamp actually has winamp browser
[21:08:11 CET] <Compn> :D
[21:08:44 CET] <fritsch> wm4: haha - sadly with all the cripled DRM content out there ... open source media centers sadly need to do that - to play that stuff with chrome html5 drm and so on
[21:08:59 CET] <wm4> reality trolls better
[21:13:06 CET] <jamrial> nevcairiel: header aside, is it truly 32bit?
[21:13:23 CET] <nevcairiel> who knows
[21:13:53 CET] <nevcairiel> its encoded as such in any case, if it outputs full 32-bit audio information i couldnt tell you
[21:14:02 CET] <nevcairiel> but my real problem is that libflac fails on it anyway :)
[21:14:15 CET] <wm4> why do you care about libflac?
[21:16:53 CET] <nevcairiel> because thats what work uses for flac handling
[21:17:15 CET] <wm4> you can't use libavcodec?
[21:18:00 CET] <nevcairiel> need libflac anyway for tagging, then using avcodec for decoding would be awkward
[21:19:17 CET] <nevcairiel> http://files.1f0.de/samples/big/Flac32.flac if someone is curious
[21:24:29 CET] <J_Darnley> I'm curious but not 200M curious.
[21:26:32 CET] <jamrial> BBB_:  for some reason my last email was sent to you alone instead of the ml
[21:26:39 CET] <jamrial> guess it's becase you cc me
[21:27:56 CET] <BBB> weird
[21:27:59 CET] <BBB> sorry about that
[21:56:14 CET] <cone-546> ffmpeg 03Michael Niedermayer 07master:9434ec5f767c: avcodec/h264_refs: Fix and add back trace code removed in e1c5170c638e39a8017827339ac687c5f6ff9b35
[22:10:01 CET] <atomnuker> Daemon404: it's been like a day and no one has responed against removing libstagefright
[22:10:15 CET] <Daemon404> atomnuker, im catching a tarsatlatnic flight tonight
[22:10:18 CET] <Daemon404> ill give it 1-2 more days
[22:10:26 CET] <Daemon404> so nobody yells at me
[22:10:35 CET] <atomnuker> yeah, I'd wait a little bit longer too
[22:11:29 CET] <J_Darnley> Today was only the first business day of new year.  Waiting is probably a good idea.
[22:15:23 CET] <atomnuker> good luck on your tarsatlatnic flight, hope not many kids to go tarsatlatnica this time of the year
[22:19:32 CET] <Daemon404> tarsatlatnica is great this time of year though
[22:19:47 CET] <Compn> we are always at war with tarsatlatnica.
[23:16:11 CET] <llogan> FFmpeg (2.9|3.0) "Tarsatlatnic"
[23:16:43 CET] <Daemon404> lulz
[23:35:18 CET] <Compn> seems like a good start to the year, trolling hardcore.
[00:00:00 CET] --- Tue Jan  5 2016


More information about the Ffmpeg-devel-irc mailing list