[Ffmpeg-devel-irc] ffmpeg-devel.log.20130607

[burek]

[00:18] <cone-101> ffmpeg.git 03Carl Eugen Hoyos 07master:d5978c8678ba: Add unscaled converter from GBRP > 8bit to RGBx48/64.
[01:08] <michaelni> Daemon404, will the integer overflow checker fate client and the other clients you did run come back ?
[01:48] <cone-101> ffmpeg.git 03Clément BSsch 07master:f97e28ebe523: lavfi/lut3d: add sanity checks.
[03:44] <Daemon404> michaelni, ive been putting off updating their known_hosts
[03:44] <Daemon404> so lazy
[03:44] <Daemon404> i SHOULD
[10:17] <cone-135> ffmpeg.git 03Martin Storsjö 07master:2681924b0f3f: libavformat: Fix standalone compilation of the webm muxer
[10:17] <cone-135> ffmpeg.git 03Michael Niedermayer 07master:a18bf95a8ec4: Merge commit '2681924b0f3f35840dd9d9825df9f9239c448b22'
[10:23] <cone-135> ffmpeg.git 03Martin Storsjö 07master:aa2c918f7d1a: rtpdec: Fix the alphabetical ordering in registering depacketizers
[10:23] <cone-135> ffmpeg.git 03Michael Niedermayer 07master:9c9d2e9c7793: Merge remote-tracking branch 'qatar/master'
[11:15] <cone-135> ffmpeg.git 03Atli Thorbjornsson 07master:f3c51215cef3: flvdec: Fix mistakenly discarding metadata at dts==0
[11:18] <durandal_1707> michaelni: if id3v2 tags are at begining of file, file is misdetected as mp3 when it is tta ^
[11:28] <cone-135> ffmpeg.git 03Paul B Mahol 07master:379fcc4955d3: id3v2: read all textual chapter subframes
[11:36] <cone-135> ffmpeg.git 03Paul B Mahol 07master:739d6a78b7ea: alacenc: don't allocate unused avctx->coded_frame
[11:36] <cone-135> ffmpeg.git 03Paul B Mahol 07master:5285a7549f6e: pcmenc: don't allocate unused avctx->coded_frame
[11:36] <cone-135> ffmpeg.git 03Paul B Mahol 07master:dc5c029f0285: sonicenc: don't allocate unused avctx->coded_frame
[11:54] <cone-135> ffmpeg.git 03Paul B Mahol 07master:5bafe0ce4463: tta: replace datalen with nb_samples
[12:26] <durandal_1707> huh ;)
[12:34] <durandal_1707> thing is files is not demuxable anyway if extension is not .tta
[12:38] <durandal_1707> which means id3 stuff is not handled correctly internally in lavf
[12:39] <cone-135> ffmpeg.git 03Paul B Mahol 07master:369684f10924: tta: unbreak demuxing of files with id3 tags at start of file
[12:57] <durandal_1707> nice, i now get segv if i try to parse id3v2 in tta
[13:00] <durandal_1707> why is id3v2 enabled only for aac and mp3?
[13:00] <durandal_1707> see libavformat/utils.c line 676
[13:02] <durandal_1707> looks like its normal to add 2047x2047 jpeg id3v2 apic frames to tta files
[13:03] <av500> but jpg is not lossless
[13:03] <durandal_1707> so what is your point?
[13:04] <durandal_1707> id3v2 can be added to virtually any file, even video...
[13:04] <av500> it should be a png
[13:04] <av500> :)
[13:05] <durandal_1707> tell them...
[13:15] <mateo`_> :q!
[13:15] <mateo`_> oups
[13:16] <durandal_1707> [av500 at home ~/torrents/] $
[13:19] <michaelni> durandal_1707, the probing of the tta fails due to PROBE_BUF_MAX
[13:22] <michaelni> parsing apics is only enabled for formats that support it because not every demuxer can handle an extra stream being inserted behind its back
[13:28] <durandal_1707> michaelni: i dont get it, i see nothing special in mp3 demuxer
[13:31] <michaelni> it doesnt need anything special IIRC but for example a demuxer that does for(i=0; i<nstreams; i++) myarray[i] will crash if it gets an extra stream from outside
[13:33] <durandal_1707> another funny thing is i file have both idv2 and ape picture tag, you will get segv
[13:34] <michaelni> do you have a sample ?
[13:35] <durandal_1707> its in #2638
[13:35] <durandal_1707> but only if you apply patch i sent on ml
[13:35] <durandal_1707> that enables extra stuff reading from id3v2 added to tta
[13:36] <durandal_1707> actually nope, you need to add both id3v2 & ape picture tag
[13:36] <durandal_1707> with Mp3tag for example
[13:39] <michaelni> ETOOLAZY
[13:44] <Compn> i didnt know my qcelp sample would cause so much problem
[13:46] <durandal_1707> michaelni: speaking of that ^, what is better solution: adding funny parser or adding CODEC_CAP_SUBFRAMES ?
[14:38] <cone-135> ffmpeg.git 03Michael Niedermayer 07master:bce2ed55596a: smvjpegdec: only extract picture when a picture has been decoded.
[14:39] <ubitux> 13:27:03 < JEEB> proper PDF for HEVC/H.265 out now http://www.itu.int/rec/T-REC-H.265-201304-I
[14:53] <durandal_1707> michaelni: so you are no more working on jpeg2000dec?
[15:00] <michaelni> durandal_1707, ATM no
[15:16] <ubitux> michaelni: any objection to the mov dts shift patch?
[15:55] <ubitux> ./ffmpeg -f lavfi -i color=#1c2c3c -vf format=rgb565 -frames:v 1 -y test.png  any idea why there is some kind of weird dithering happening here?
[15:55] <ubitux> (the output is not uniform)
[15:58] <michaelni> ubitux, no objection to the dts patch
[15:59] <ubitux> ok, will apply soon then, thx
[15:59] <durandal_1707> ubitux: what weird means?
[15:59] <ubitux> durandal_1707: well output is not uniform
[15:59] <ubitux> try to zoom in with gimp or whatever
[15:59] <durandal_1707> but its 565
[16:01] <durandal_1707> how should #1c2c3c looks like in 565 ?
[16:04] <ubitux> 0x1967 everywhere?
[16:05] <durandal_1707> nope, as you are losing last bit
[16:05] <ubitux> (or #182c38 everywhere)
[16:05] <ubitux> alright so there is some general averaging?
[16:06] <ubitux> is there a sws flag to disable that?
[16:06] <durandal_1707> if there is not enought bitdept, some kind of dithering should be done
[16:07] <ubitux> i don't want it with my input here; i want to keep the uniformity over the overall accuracy
[16:07] <ubitux> (uniformity is more important in my case)
[16:08] <durandal_1707> see libswscale
[16:09] <ubitux> i've tried a few flags, but none seems to do the trick
[16:15] <TimNich> I've found a typo in the il filter usage description.
[16:16] <durandal_1707> great!
[16:18] <TimNich> filters.texi  "@item chroma_mode, s"  should be   "@item chroma_mode, c"
[16:19] <durandal_1707> send patch
[16:27] <TimNich> done
[16:28] <cone-135> ffmpeg.git 03Paul B Mahol 07master:c59c0488ec14: cmdutils: print threading capabilities for audio codecs
[16:29] <TimNich> useful filter though :)
[16:29] <durandal_1707> for what?
[16:30] <cone-135> ffmpeg.git 03Alexandre Sicard 07master:8912029031d3: avformat/mov: compute dts_shift with trun cts
[16:31] <TimNich> makes showing up when frames are wrongly packed fieldwise
[16:32] <michaelni> ubitux, if you want want to disable dithering in sws, it should work to set the various dither fields to a constant like 0 
[16:34] <cone-135> ffmpeg.git 03Tim.Nicholson 07master:0ea135613788: doc/filters: fix alias option name for chroma_mode in 'il' filter.
[16:38] <TimNich> Ta!
[16:39] <TimNich> fastest commit I've ever had ;P
[20:51] <cone-454> ffmpeg.git 03Michael Niedermayer 07master:467e7a8f26e5: avcodec/smvjpegdec: check that frames_per_jpeg cleanly divides height
[21:39] <cone-454> ffmpeg.git 03Carl Eugen Hoyos 07master:7ce011e054d1: Do not offer smvjpeg lowres capabilities, the decoder does not support it.
[21:39] <cone-454> ffmpeg.git 03Michael Niedermayer 07master:5b9917ec2b13: Merge remote-tracking branch 'cehoyos/master'
[23:26] <ubitux> Daemon404: there is no way on a embedded vimeo video to jump to get/jump to the vimeo url?
[23:27] <Daemon404> ... right click
[23:27] <ubitux> i don't see anything relevant in that menu
[23:27] <Daemon404> link
[23:27] <Daemon404> we do let people customize their player / viewability
[23:27] <Daemon404> they can hide it from vimeo.com
[23:27] <ubitux> http://www.platige.com/en/page/61-The_Kinematograph
[23:28] <Daemon404> looks like they disabled it
[23:28] <Daemon404> user's choice
[23:29] <ubitux> why do you allow this?
[23:30] <Daemon404> ... because they pay for a service
[23:31] <Daemon404> why SHOULDNT we allow it?
[23:31] <Daemon404> so satisify FREEDOM FIGHTERS?
[23:31] <llogan> pleaase allow Cinepak encoding.
[23:31] <ubitux> Daemon404: ah, i see
[23:31] <Daemon404> ubitux, you need plus or pro to disable/customize it
[23:32] <Daemon404> it's one of the selling points
[23:32] <ubitux> ok, interesting
[23:32] <Daemon404> think like
[23:32] <Daemon404> if you are sing it o na professional site
[23:32] <Daemon404> would oyu want big ol vimeo links?
[23:32] <Daemon404> i wouldnt.
[23:33] <ubitux> i see; it seems that video isn't available directly on vimeo at all, unless i'm missing something
[23:33] <Daemon404> that could be the case
[23:33] <Daemon404> i didnt check
[23:34] <ubitux> <@Daemon404> why SHOULDNT we allow it? // i had in mind that would like to see vimeo comments for any embedded video for instance, but sure i get the commercial point now
[23:34] <ubitux> anyway, sorry channel for OT ;)
[23:37] <Daemon404> ah
[23:37] <Daemon404> ok
[23:38] <durandal_1707> ubitux: do you want to be punished?
[23:42] <ubitux> durandal_1707: no :(
[00:00] --- Sat Jun  8 2013