[Ffmpeg-devel-irc] ffmpeg-devel.log.20120321

burek burek021 at gmail.com
Thu Mar 22 02:05:02 CET 2012


[00:04] <michaelni> Compn, did you say something before "otherwise" ? iam having a minor issue with me english parser here ;)
[00:07] <michaelni> maybe someone wants to go to linuxtag 2013 ? 
[00:08] <michaelni> i know its a little early to throw this question in ;)
[00:08] <michaelni> but better too early than too late ... :)
[00:10] <durandal_1707> lol
[00:58] <Compn> michaelni : i meant 'i'm far from linuxtag, i'm not interested in going, thanks for asking'
[01:01] <michaelni> Compn, ok
[01:02] <Compn> i used to like travelling. but the TSA is crazy around here :(
[02:35] <CIA-17> ffmpeg: 03Anton Khirnov 07master * r713f3062a7 10ffmpeg/libavformat/id3v2.c: id3v2: set the keyframe flag on attached pictures.
[02:35] <CIA-17> ffmpeg: 03Aaron Colwell 07master * r30011bf201 10ffmpeg/libavcodec/vp8.c: (log message trimmed)
[02:35] <CIA-17> ffmpeg: vp8: avoid race condition on segment map.
[02:35] <CIA-17> ffmpeg: This change avoids accessing the segment map of the previous frame if
[02:35] <CIA-17> ffmpeg: segmentation is not enabled for the current frame. The caller of
[02:35] <CIA-17> ffmpeg: decode_mb_mode() only calls ff_thread_await_progress() on the reference
[02:35] <CIA-17> ffmpeg: segmentation index array if segmentation is enabled, so Chromium's TSAN
[02:35] <CIA-17> ffmpeg: will report a race when accessing this data while segmentation is not
[02:35] <CIA-17> ffmpeg: 03Anton Khirnov 07master * r01fcc42b90 10ffmpeg/libavformat/utils.c: 
[02:35] <CIA-17> ffmpeg: lavf: requeue attached pictures after seeking.
[02:35] <CIA-17> ffmpeg: This allows the caller to get them without special code even after
[02:35] <CIA-17> ffmpeg: seeking before receiving any data.
[02:35] <CIA-17> ffmpeg: 03Anton Khirnov 07master * r0d0b81f941 10ffmpeg/ (4 files in 2 dirs): Generate manpages for AV{Format,Codec}Context AVOptions.
[02:35] <CIA-17> ffmpeg: 03Anton Khirnov 07master * r4fea8959d8 10ffmpeg/ (doc/avconv.texi libavcodec/options_table.h): 
[02:35] <CIA-17> ffmpeg: doc/avconv: remove entries for AVOptions.
[02:35] <CIA-17> ffmpeg: Documentation for those will be generated automatically.
[02:35] <CIA-17> ffmpeg: 03Martin Storsjö 07master * r4fa57d524f 10ffmpeg/libavformat/ (mpegenc.c mpegtsenc.c options_table.h rtpenc.c rtsp.c): 
[02:35] <CIA-17> ffmpeg: libavformat: Set the default for the max_delay option to -1
[02:35] <CIA-17> ffmpeg: Make the muxers/demuxers that use the field handle the default
[02:35] <CIA-17> ffmpeg: -1 in the same way as 0.
[02:35] <CIA-17> ffmpeg: This allows distinguishing an intentionally set 0 from the default
[02:35] <CIA-17> ffmpeg: value where the user hasn't set it.
[02:35] <CIA-17> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[02:35] <CIA-17> ffmpeg: 03Anton Khirnov 07master * r5626697104 10ffmpeg/ (4 files in 2 dirs): 
[02:35] <CIA-17> ffmpeg: Move AVFormatContext/AVCodecContext option tables to separate files.
[02:35] <CIA-17> ffmpeg: frag_keyframe) are combined with OR, cutting fragments at the
[02:35] <CIA-17> ffmpeg: first of the conditions being fulfilled.
[02:35] <CIA-17> (49 lines omitted)
[02:36] <CIA-17> ffmpeg: 03Paul B Mahol 07master * rb36872bdb6 10ffmpeg/tests/ (fate/adpcm.mak ref/fate/adpcm-ea-maxis-xa): 
[02:36] <CIA-17> ffmpeg: FATE: add test for adpcm-ea-maxis-xa
[02:36] <CIA-17> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[02:36] <CIA-17> ffmpeg: Signed-off-by: Justin Ruggles <justin.ruggles at gmail.com>
[02:36] <CIA-17> ffmpeg: 03Justin Ruggles 07master * rf63412fc74 10ffmpeg/ (libavcodec/Makefile libavformat/oggparsevorbis.c): (log message trimmed)
[02:36] <CIA-17> ffmpeg: oggdec: output correct timestamps for Vorbis
[02:36] <CIA-17> ffmpeg: Takes encoder delay into account by comparing first the coded page
[02:36] <CIA-17> ffmpeg: duration with the calculated page duration. Handles last packet duration
[02:36] <CIA-17> ffmpeg: if needed, also by comparing coded duration with calculated duration.
[02:36] <CIA-17> ffmpeg: Also does better handling of timestamp generation for packets in the
[02:37] <CIA-17> ffmpeg: first page for streamed ogg files where the start time is not
[02:37] <CIA-17> ffmpeg: 03Anton Khirnov 07master * ra6733202cc 10ffmpeg/libavformat/ (avformat.h gxfenc.c internal.h utils.c version.h): 
[02:37] <CIA-17> ffmpeg: lavf: make av_interleave_packet_per_dts() private.
[02:37] <CIA-17> ffmpeg: There is no reason for it to be public, it's only meant to be used
[02:37] <CIA-17> ffmpeg: internally.
[02:37] <CIA-17> ffmpeg: 03Paul B Mahol 07master * r791d6df4ae 10ffmpeg/tests/ (fate/demux.mak ref/fate/maxis-xa): 
[02:37] <CIA-17> ffmpeg: FATE: change fate-maxis-xa to a normal demuxing test
[02:37] <CIA-17> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[02:37] <CIA-17> ffmpeg: Signed-off-by: Justin Ruggles <justin.ruggles at gmail.com>
[03:26] <CIA-17> ffmpeg: 03Paul B Mahol 07master * r841e669a39 10ffmpeg/libavcodec/cdxl.c: 
[03:26] <CIA-17> ffmpeg: cdxl: swap CHUNKY and BYTE_PLANAR
[03:26] <CIA-17> ffmpeg: This is how it is defined in Amiga Developer CD from year 1992 and
[03:26] <CIA-17> ffmpeg: is consistent with files created with ADPro.
[03:26] <CIA-17> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[03:26] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[03:26] <CIA-17> ffmpeg: 03Paul B Mahol 07master * r8a90148dfe 10ffmpeg/libavcodec/smc.c: 
[03:26] <CIA-17> ffmpeg: smc: use bytestream2_size()
[03:26] <CIA-17> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[03:26] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[03:26] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r3eaf712053 10ffmpeg/libavcodec/sgienc.c: 
[03:26] <CIA-17> ffmpeg: sgienc: fix packet size.
[03:26] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[03:26] <CIA-17> ffmpeg: 03Paul B Mahol 07master * rff05fd6249 10ffmpeg/libavcodec/xxan.c: 
[03:26] <CIA-17> ffmpeg: xxan: remove write-only variable
[03:26] <CIA-17> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[03:27] <CIA-17> ffmpeg: xxan: use bytestream2_size()
[03:27] <CIA-17> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[03:27] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[03:27] <CIA-17> ffmpeg: 03Paul B Mahol 07master * r1a7a707f74 10ffmpeg/libavcodec/eatgq.c: 
[03:27] <CIA-17> ffmpeg: tgq: use bytestream2_get_bytes_left()
[03:27] <CIA-17> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[03:27] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[03:27] <CIA-17> ffmpeg: 03Paul B Mahol 07master * r0ee5be4ee4 10ffmpeg/libavcodec/bytestream.h: 
[03:27] <CIA-17> ffmpeg: bytestream: add functions for accessing size of buffer
[03:27] <CIA-17> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[03:27] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[04:33] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r9e69d3c6d4 10ffmpeg/libavcodec/zerocodec.c: 
[04:33] <CIA-17> ffmpeg: zerocodec: factorize loop
[04:33] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[04:33] <CIA-17> ffmpeg: 03Paul B Mahol 07master * r594a3d6315 10ffmpeg/libavcodec/bink.c: 
[04:33] <CIA-17> ffmpeg: bink: no need to increase width twice
[04:33] <CIA-17> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[04:33] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[04:33] <CIA-17> ffmpeg: 03Paul B Mahol 07master * r15e07348fe 10ffmpeg/libavcodec/tta.c: 
[04:33] <CIA-17> ffmpeg: ttadec: refactor ttafilter_process()
[04:33] <CIA-17> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[04:33] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[10:51] <ubitux> didn't nicolas said he'd push bprint api this last w-e?
[10:51] <ubitux> :(
[10:52] <CIA-17> ffmpeg: 03Michael Niedermayer 07release/0.10 * rabfafb6c81 10ffmpeg/libavcodec/pngenc.c: 
[10:52] <CIA-17> ffmpeg: pngenc: Fix incorrect mask used for interlaced mode.
[10:52] <CIA-17> ffmpeg: Fixes Ticket1109
[10:52] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[10:52] <CIA-17> ffmpeg: (cherry picked from commit 15db6a959057b92245a384909ec7d413d5c16461)
[13:26] <CIA-17> ffmpeg: 03Tomas Härdin 07release/0.10 * re687d77d15 10ffmpeg/libavformat/mxfdec.c: 
[13:26] <CIA-17> ffmpeg: mxfdec: Only parse next partition pack if parsing forward
[13:26] <CIA-17> ffmpeg: This fixes ticket #1099.
[13:26] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[13:26] <CIA-17> ffmpeg: (cherry picked from commit 393b81f0934866bd7fff0a2b113623dd9ee6808f)
[13:26] <CIA-17> ffmpeg: 03Tomas Härdin 07release/0.9 * r515b7f08b1 10ffmpeg/libavformat/mxfdec.c: 
[13:26] <CIA-17> ffmpeg: mxfdec: Only parse next partition pack if parsing forward
[13:26] <CIA-17> ffmpeg: This fixes ticket #1099.
[13:26] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[13:26] <CIA-17> ffmpeg: (cherry picked from commit 393b81f0934866bd7fff0a2b113623dd9ee6808f)
[13:40] <ubitux> michaelni: was the recently added random tests in fate for this: https://banu.com/blog/42/openbsd-bug-in-the-random-function/ ? :)
[13:59] <CIA-17> ffmpeg: 03Nicolas George 07master * rb75c67dc01 10ffmpeg/ (5 files in 2 dirs): lavu: add av_bprintf and related.
[13:59] <CIA-17> ffmpeg: 03Nicolas George 07master * r6cf53927c4 10ffmpeg/libavfilter/graphdump.c: graphdump: use av_bprintf API.
[14:00] <ubitux> \o/
[14:08] <michaelni> ubitux, the openbsd (s)random() is not strictly buggy IMHO
[14:08] <michaelni> it seems a correct implementation of a poor algorithm
[14:08] <michaelni> and (s)random() has always been a poor algorithm on every platform ive seen
[14:10] <michaelni> the openbsd one is a bit poorer than the other ones though
[14:11] <ubitux> ok :)
[14:11] <ubitux> did they fix their shell btw?
[14:11] <ubitux> i saw you sending them some patches
[14:11] Action: ubitux is watching you
[14:12] <michaelni> IIRC they fixes it
[14:12] <michaelni> fixed
[14:12] <ubitux> ok, nice :)
[14:17] <michaelni> ubitux, is there some thread @openbsd somewhere about this bug ?
[14:17] <michaelni> the srandom() one i mean
[14:18] <ubitux> i didn't saw it
[14:18] <ubitux> i was just subscribing to bugs at openbsd since a while because of an old request
[14:18] Action: michaelni had some better random number generators laying around ...
[14:19] <michaelni> which is why i aasked ...
[14:19] <ubitux> i'm not really following openbsd dev, it was just an historical subscribtion
[14:19] <ubitux> it's mainly a spam mailing list btw :p
[14:19] Action: michaelni is just lookibng at gmane of bugs at openbsd
[14:19] <michaelni> so i see it too :)
[14:20] <ubitux> :)
[14:34] Action: michaelni wants to reply to that blog but theres no way to comment :/
[14:34] <michaelni> ohh well
[14:35] <av500> not even with a facebook account?
[14:35] <michaelni> the way the issue is presented is that openbsd srandom(0) leads to a sequence of 0,0,0, ... that sucks
[14:35] <michaelni> facebook ?
[14:35] Action: michaelni doesnt touch such things
[14:35] Action: michaelni is too paranoid
[14:36] <ubitux> the article is on HN
[14:36] <ubitux> you can comment there
[14:37] <ubitux> no need for an email to create an account
[14:37] <av500> they ask you to sign in too
[14:37] <ubitux> it's instant
[14:37] <ubitux> just use pwgen to create a login/pass, and you're automatically log'in
[14:38] <ubitux> strange i though the article was on HN but it seems not
[14:39] <ubitux> http://news.ycombinator.com/item?id=3730348
[14:39] <ubitux> here it is.
[14:41] <spaam> http://www.openbsd.org/cgi-bin/cvsweb/src/lib/libc/stdlib/random.c.diff?r1=1.15;r2=1.16  ze fix for that bug
[14:41] <spaam> fyi
[14:42] <spaam> commited ~2h ago.
[14:42] <ubitux> 2005 ’ 2012
[14:42] <ubitux> nice
[14:54] Action: michaelni finally wrote the comment :)
[14:57] <av500> michaelni: there, upvoted you
[15:22] <CIA-17> ffmpeg: 03Stefano Sabatini 07master * re71e65ff1d 10ffmpeg/libavfilter/vf_aspect.c: 
[15:22] <CIA-17> ffmpeg: lavfi/aspect: check for a negative code from av_parse_ratio()
[15:22] <CIA-17> ffmpeg: Check on < 0 rather than on != 0, this is more correct as a positive
[15:22] <CIA-17> ffmpeg: error code from av_parse_ratio() value doesn't mean an error.
[15:22] <CIA-17> ffmpeg: 03Stefano Sabatini 07master * r8e0d3c0369 10ffmpeg/ (doc/filters.texi libavfilter/version.h libavfilter/vf_ass.c): 
[15:22] <CIA-17> ffmpeg: lavfi/ass: add dar option
[15:22] <CIA-17> ffmpeg: Allow to specify the display aspect ratio adopted for rendering
[15:22] <CIA-17> ffmpeg: subtitles.
[15:22] <CIA-17> ffmpeg: 03Stefano Sabatini 07master * rc9399538b7 10ffmpeg/libavfilter/vf_ass.c: 
[15:22] <CIA-17> ffmpeg: lavfi/ass: use a default DAR value of 1.0
[15:22] <CIA-17> ffmpeg: Previously it was using the same value of the input video DAR, which is
[15:22] <CIA-17> ffmpeg: inconsistent with most implementations.
[15:22] <CIA-17> ffmpeg: Fix trac ticket #1098.
[15:31] <ubitux> hey i was wondering, would it make sense for subtitles to use dts and pts?
[15:31] <ubitux> for instance the ass subtitles may not be ordered in the presentation timestamp
[15:32] <ubitux> internally iirc there is a qsort() run on them or sth like that
[15:33] <av500> they are not ordered?
[15:33] <kierank> ubitux: why are they not ordered
[15:33] <kierank> that's crazy
[15:33] <Plorkyeran> they're sorted when they're muxed into mkv
[15:33] <Plorkyeran> but raw ass files don't have to be
[15:33] <kierank> oh
[15:34] <av500> so you want to not sort them in mkv?
[15:34] <av500> and instead assing dts and pts?
[15:34] <av500> and have somebody re-order them at decode time?
[15:34] <av500> hmm, maybe p-subs that depend on earlier I-subs
[15:35] <ubitux> it's sometimes interesting to have them sorted not in the presentation time
[15:36] <ubitux> for instance with ass you can hack the "signs" in a video
[15:36] <ubitux> so you might want to regroup all the sign events
[15:36] <Plorkyeran> when editing you often want them out of order, but because of vsfilter bugs the end result should be sorted
[15:36] <ubitux> and keep the dialogs clean
[15:37] <Plorkyeran> (there's some weird border rendering issues with unsorted subtitles that makes zero sense)
[15:37] <ubitux> this is for ass, but it can be applied to other formats as well
[15:37] <ubitux> where some decoders might be able to handle that situation properly
[15:37] <ubitux> and vsfilter could be fixed too in the future :)
[15:38] <Plorkyeran> finishing the non-fontconfig libass stuff would probably be easier
[15:39] <Plorkyeran> and a better end result
[15:39] <ubitux> where is code doing the reordering in ffmpeg btw?
[15:40] <saste> the more i know about ass the more insane it looks
[15:41] <ubitux> i think it's a good idea to allow such operation
[15:41] <ubitux> of course, not in a muxed situation
[15:41] <ubitux> but for a standalone subtitles it clearly helps editing
[15:42] <av500> isnt subtitle editing outside of the scope of ffmpeg?
[15:42] <av500> libavsubedit?
[15:42] <ubitux> i want to ffmpeg -i a.ass -sf 'sed=s/I/l/' ;)
[15:42] <saste> libavbringmeacoffe
[15:43] <ubitux> av500: even if it's not the purpose of ffmpeg to edit subtitles, supporting them for convert is important
[15:43] <ubitux> it is actually doing it
[15:43] <ubitux> afaik, when you do ffmpeg -i a.ass a.srt, it's doing an internal reordering of the events
[15:44] <ubitux> i was just asking if that would make sense to use the pts/dts stuff for that :)
[15:44] <ubitux> so we would put the subtitles id in the dts, and the subtitles starting time in the pts
[15:44] <ubitux> or sth like that
[15:45] <av500> and then mux this?
[15:45] <av500> or reorder in the muxer?
[15:45] <ubitux> keeping the original order might be interesting
[15:45] <av500> how long would you delay the muxing for that?
[15:46] <av500> and other demuxers/player would know how to handle out of order subs?
[15:46] <ubitux> i don't know how it would work with muxed stream
[15:46] <ubitux> but i guess it must not be supported with muxed ones
[15:47] <ubitux> i just don't know :)
[15:52] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r87a72b9122 10ffmpeg/libswscale/utils.c: 
[15:52] <CIA-17> ffmpeg: swscale: Merge a hunk from qatar that seems to have been forgotten or lost.
[15:52] <CIA-17> ffmpeg: Author of the code from qatar is Ronald S. Bultje
[15:52] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:01] <CIA-17> ffmpeg: 03Nicolas George 07master * r8ec0832743 10ffmpeg/libavfilter/ (Makefile drawutils.c drawutils.h): 
[16:01] <CIA-17> ffmpeg: drawutils: new API.
[16:01] <CIA-17> ffmpeg: This new API stores useful data in a dedicated structure
[16:01] <CIA-17> ffmpeg: and has clearly delimited init functions.
[16:01] <CIA-17> ffmpeg: Hopefully, uses of the old API can be replaced quickly.
[16:01] <CIA-17> ffmpeg: 03Nicolas George 07master * r53b7a3fe08 10ffmpeg/ (libavfilter/vf_pad.c tests/ref/lavfi/pixfmts_pad): vf_pad: port to new drawutils API.
[16:01] <CIA-17> ffmpeg: 03Nicolas George 07master * r7084985173 10ffmpeg/libavfilter/vsrc_color.c: vsrc_color: port to new drawutils API.
[16:01] <CIA-17> ffmpeg: 03Nicolas George 07master * r3493390d47 10ffmpeg/ (5 files in 3 dirs): lavfi: add tile video filter.
[16:02] <ubitux> saste: btw, about the -af
[16:02] <ubitux> do you think it would be better to use a common avfilter context, or independent ones?
[16:03] <ubitux> having a common filtergraph looks like it will ease a/v sync between the two
[16:03] <ubitux> but it might require more complex usage of it
[16:07] <saste> ubitux: do you mean -avf ... against -af, -vf?
[16:08] <saste> that's not a bad idea, but looks more complicate, -af looks the more straightforward approach, then we can merge the two paths
[16:11] <juanmabc> hint: i would like to have a epts (estimated presentation time stamp), for the decoded frame. Any chances you'll implement that to reduce player (user) (re)coding?
[16:12] <juanmabc> i.e. move some ffplay.c code to epts
[16:23] <durandal_1707> how do you convert float to AVRational?
[16:27] <merbanan> could be something in avutil
[16:27] <nevcairiel> av_d2q
[16:29] <durandal_1707> doesn't double take more thant 32bits?
[16:29] <juanmabc> it's a bad thing (tm) that after all that coding work to solve all that formats, the pts is so tricky in player making
[16:30] <nevcairiel> double is 64-bits, but no harm in casting the float to double, the AVRational is 64-bit anyway
[16:33] <Tjoppen> the real question is why you have float to begin with
[16:33] <durandal_1707> ask guys who created such format
[17:44] <ubitux> <@saste> ubitux: do you mean -avf ... against -af, -vf? // no i meant internally
[17:44] <ubitux> the avfilter graph context could be shared
[17:44] <ubitux> but yeah more complicated
[18:02] <ubitux> btw, not related
[18:02] <ubitux> any reason we have OPT_FUNC2 but no OPT_FUNC?
[18:30] <CIA-17> ffmpeg: 03Tomas Härdin 07master * r8d393ce330 10ffmpeg/libavformat/aiffdec.c: 
[18:30] <CIA-17> ffmpeg: aiffdec: Fix SIGFPE on pcm_f32be
[18:30] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[18:30] <CIA-17> ffmpeg: 03Paul B Mahol 07master * r98ca4d217d 10ffmpeg/ (libavcodec/adpcm.c libavformat/westwood_vqa.c): 
[18:30] <CIA-17> ffmpeg: westwood_vqa: fix SND0 chunk handling
[18:30] <CIA-17> ffmpeg: Version from vqa header does not dictate which sound chunks may
[18:30] <CIA-17> ffmpeg: appear in file.
[18:30] <CIA-17> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[18:30] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[18:30] <CIA-17> ffmpeg: 03Paul B Mahol 07master * r5ca595facb 10ffmpeg/libavformat/westwood_vqa.c: 
[18:30] <CIA-17> ffmpeg: westwood_vqa: set video stream duration
[18:30] <CIA-17> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[18:30] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[18:52] <saste> ubitux: OPT_FUNC2/OPT_FUNC... there was some mess going on when merging from libav
[18:53] <saste> i removed OPT_FUNC2 at some point but then it was readded in a merge
[18:53] <saste> btw now that i think at it ffprobe -i FILE is still broken...
[19:08] <ubitux> i'm not sure but i think there is sth wrong with a bool+func option
[19:13] <ubitux> mmh
[19:13] <ubitux> OPT_FUNC might be for this
[19:15] <ubitux> i just need a OPT_BOOL + func ptr so
[19:15] <ubitux> and no OPT_FUNC2
[19:36] <CIA-17> ffmpeg: 03Nicolas George 07master * r38c3b68e3a 10ffmpeg/libavutil/ (avutil.h pixdesc.c pixdesc.h): pixdesc: put components always in the same order.
[20:15] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * r89411ae699 10ffmpeg/libavcodec/x86/sbrdsp.asm: 
[20:15] <CIA-17> ffmpeg: Replace SSE2 instruction by SSE equivalent.
[20:15] <CIA-17> ffmpeg: This is even potentially faster in this use-case.
[20:15] <CIA-17> ffmpeg: Should fix AAC SBR decoding on machines with SSE but not
[20:15] <CIA-17> ffmpeg: SSE2, fixing track issue #1041.
[20:15] <CIA-17> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[20:23] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * r2a4f1d38d9 10ffmpeg/libavformat/utils.c: 
[20:23] <CIA-17> ffmpeg: Print a warning when using wild guesses as time stamps.
[20:23] <CIA-17> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[21:15] <pasteeater> anyone ever encounter a xpa video file from security camera? a PI dropped it off expecting a conversion.
[21:18] <Daemon404> pasteeater, random google makes it seem encrypted? O.o
[21:19] <pasteeater> i saw something like that too, but also some conflicting opinions. IIRC.
[21:19] <Daemon404> so dude just dropped it off without anything, expectign a conversion
[21:19] <Daemon404> ?
[21:20] <pasteeater> yes, but he has no further information other than it's from a national chain of stores and the video is evidence for a court case.
[21:21] <pasteeater> and the proprietary players from the assumed manufacturer would not play the alleged video.
[21:22] <Daemon404> ... he could have at least pointed you to WHAT the proprietary player was
[21:56] <Daemon404> hey durandal_1707 
[21:57] <Daemon404> i found an vqa file on a game of mine that has another unknown chunk type (undocumented)
[21:57] <durandal_1707> give chunk type
[21:57] <Daemon404> LINF
[21:58] <durandal_1707> and what about it? have you looked in hex editor where and how big it is?
[21:58] <Daemon404> you were the one poking vqa
[21:58] <Daemon404> so i figured it might be intersting to you.
[21:58] <durandal_1707> there is nothing interesting left for vqa
[21:59] <Daemon404> unknwon block types arent interesting?
[21:59] <durandal_1707> just existing code is soo ugly that it first need to be completly rewritten to add support for hicolor files
[22:00] <Daemon404> yes
[22:00] <Daemon404> seems i have a hicolor file
[22:00] <durandal_1707> that specific chunk is irrelevant in decoding video
[22:00] <Daemon404> doesnt mean it is entirely irrelevant to decoding.
[22:00] <durandal_1707> i have already made patch with bunch of chunks which need to be ignored....
[22:00] <Daemon404> its nto even documented
[22:01] <Daemon404> do you know what LINF is?
[22:01] <Daemon404> or are you talking out your ass?
[22:01] <durandal_1707> Daemon404: there is vqadll.dll which decodes such files fine and do not use LINF
[22:02] <durandal_1707> as I can see in hex editor there is only one LINF chunk per file
[22:02] <Daemon404> doesnt mean it shouldnt be parsed
[22:02] <durandal_1707> and it's size is 22 bytes
[22:02] <Daemon404> if it has metadata
[22:03] <durandal_1707> if you can dechiper what that metadata means ...
[22:03] <Daemon404> i have relevant dlls
[22:03] <Daemon404> i might poke it.
[22:04] <Daemon404> also other fun files to poke with this game
[22:04] <Daemon404> like thing.bin
[22:04] <CIA-17> ffmpeg: 03gs_gail 07master * r361c4cdca9 10ffmpeg/libavcodec/h264_mp4toannexb_bsf.c: h264_mp4toannexb_bsf: support 24bit lengths
[22:05] Action: durandal_1707 currently working on this crap http://wiki.multimedia.cx/index.php?title=AMF
[22:06] <Daemon404> have fun
[22:08] <durandal_1707> no fun at all, can't get decoding to work correctly (funny thing is i can see few frames with carts)
[22:19] <durandal_1707> how to use test/probetest ?
[22:20] <durandal_1707> Failure of jv probing code with score=100 type=3 p=1C2 size=4
[22:20] <durandal_1707> Failure of swf probing code with score=100 type=3 p=162 size=4
[22:21] <durandal_1707> what those numbers means?
[22:26] <Daemon404> no clue
[22:26] Action: Daemon404 is busy finding old lulzy video codecs on his old unlabelled discs
[22:26] <Daemon404> like angel potion
[22:26] <JEEB> :<
[22:26] <Daemon404> ?
[22:27] <CIA-17> ffmpeg: 03Carl Eugen Hoyos 07master * rd3640096aa 10ffmpeg/libavcodec/libspeexdec.c: 
[22:27] <CIA-17> ffmpeg: Fix libspeex decoding.
[22:27] <CIA-17> ffmpeg: This parly reverts 85469f.
[22:27] <CIA-17> ffmpeg: Fixes ticket #1091.
[22:27] <CIA-17> ffmpeg: 03Vidar Madsen 07master * r17431f4a4e 10ffmpeg/ (libavformat/mmf.c tests/ref/lavf/mmf): Fix type and codetype fields in the MMF header for some phones.
[22:29] <Daemon404> gwsdos.zip
[22:29] <Daemon404> wow.
[22:30] Action: Daemon404 moves it to 'interesting files'
[22:47] <durandal_1707> michaelni: what probetest is supposed to show?
[22:56] <michaelni> durandal_1707, probetest tests the demuxers with various forms of random data
[22:56] <Daemon404> http://i.imgur.com/vOoV1.png <-- any of these particularily interesting?
[22:57] <michaelni> if one demuxer claims positive detections on such random data thats a "bug"
[22:57] <michaelni> the higher the score the worse
[22:58] <durandal_1707> Daemon404: we already have most of it
[22:59] <durandal_1707> but EXE/ELF one could be nice :)
[22:59] <Daemon404> lol.
[22:59] <Daemon404> decoding dos self-displaying images. seems silly.
[23:00] <funman> --enable-qemu ?
[23:00] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rf1fdd208cc 10ffmpeg/libavcodec/h264_mp4toannexb_bsf.c: 
[23:00] <CIA-17> ffmpeg: h264_mp4toannexb_bsf: use i instead of misusing unit_type.
[23:00] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:00] <durandal_1707> Daemon404: i mean encoding
[23:01] <Daemon404> lol.
[23:01] <Daemon404> this old backup has fun stuff and im more than entertained for now
[23:01] <Daemon404> just found Attack of the Mutant Space Bats of Doom for DOS 
[23:05] <durandal_1707> jv probing is failing but probe functions actually checks for long ascii string...
[23:05] <ohsix> did you solve the halting problem in the demuxer :D
[23:05] <durandal_1707> ?
[23:08] <durandal_1707> i wrote perfect probe function for new demuxer
[23:49] <Compn> Daemon404 : you could make samples for each one
[23:49] <Compn> for some reason, ffmpeg has a lot of image format demuxers :P
[23:50] <Daemon404> lol
[00:00] --- Thu Mar 22 2012


More information about the Ffmpeg-devel-irc mailing list