[Ffmpeg-devel-irc] ffmpeg-devel.log.20150501

burek burek021 at gmail.com
Sat May 2 02:05:02 CEST 2015


[02:29:37 CEST] <jamrial> michaelni: http://fate.ffmpeg.org/report.cgi?time=20150430232606&slot=x86_64-archlinux-gcc-valgrindundef both are the same bug, svq3 not freeing the (H264Context)h->last_pic_for_ec.f frame allocated after calling ff_h264_decode_init()
[02:31:46 CEST] <jamrial> not sure what would be the cleanest way to deal with that, if either moving the av_frame_free() call in h264.c somewhere else, or duplicating it in svq3.c, or something else
[02:34:40 CEST] <debianuser> Hello. `ffmpeg -i in.vob -c:a copy -c:v libx264 out.mkv` prints "[matroska @ 0xb5f7aa0] Codec for stream 1 does not use global headers but container format requires global headers" since n2.6-dev-3056-g9598946 merge from libav. in.vob sample: datafilehost.com/d/dadbd9ed Does it count as ffmpeg bug?
[02:36:03 CEST] <michaelni> debianuser, yes, please open a ticket on trac
[02:38:26 CEST] Action: debianuser now needs to register on one more bugtracker... figure out some login and password...
[02:38:28 CEST] <debianuser> Heh... Maybe bugreport over IRC counts? ;)
[02:43:02 CEST] <llogan> this was already discussed somewhere recently
[02:43:07 CEST] <llogan> IIRC
[02:44:17 CEST] <llogan> [FFmpeg-devel] [PATCH]lavf/mkv: Fix AAC remuxing
[02:44:25 CEST] <llogan> assuming this is the same issue
[02:45:19 CEST] <cone-747> ffmpeg 03Michael Niedermayer 07master:4e8d01f20ce8: avformat/mpegts: Factorize version checking code out
[02:45:19 CEST] <cone-747> ffmpeg 03Michael Niedermayer 07master:4b6be54bed27: avformat/mpegts: reset last_ver on corrupted packets
[02:45:19 CEST] <cone-747> ffmpeg 03Michael Niedermayer 07master:e0153145f6f8: avformat/mpegts: Detect changes in packet through CRC instead of just the 5bit version
[03:30:19 CEST] <debianuser> michaelni: done
[03:32:26 CEST] Action: debianuser sighs... bugreporting over IRC could be faster and more interactive...
[03:33:39 CEST] <jamrial> harder to keep track of reports that way
[03:36:11 CEST] <debianuser> There must be some IRC bot automatically archiving bugreports from IRC and discussions about it!
[03:53:24 CEST] <cone-747> ffmpeg 03Michael Niedermayer 07master:3cd666cfbc7d: avcodec/svq3: Fix memleak of last_pic_for_ec
[03:53:52 CEST] <michaelni> jamrial, fixed
[04:19:29 CEST] <cone-747> ffmpeg 03Shivraj Patil 07master:02001ada5cdd: avcodec/mips: MSA (MIPS-SIMD-Arch) optimizations for H264 lpf and weight/biweight functions
[08:46:57 CEST] <BBB> michaelni: new patchset (w/o the one you applied to fix the mt bug) on github, if you want to merge from there later today
[12:01:19 CEST] <wm4> ubitux: if lavfi used a proper scripting language, filter sub-graphs could be functions, and you could deal with the issue that individual filters != use cases
[12:26:57 CEST] <ubitux> wm4: that's not a lavfi scripting that you are looking for, that's a scripting for ffmpeg api / AVFrame, on top of lavfi (so you could create multiple "graph"/filter instances) and feed them how you want
[12:27:28 CEST] <wm4> these things are different
[12:27:33 CEST] <wm4> but it shows the inflexibility of lavfi
[12:27:56 CEST] <wm4> lavfi needs everything in a "configured" graph, and it's apparently not easy to compose multiple such graphs
[12:28:24 CEST] <wm4> what I meant was more in the context of filters which do some basic work, and need to be combined with others to make up a real use-case
[12:28:46 CEST] <wm4> like combining findandcover and delogo (as it was suggested)
[12:30:21 CEST] <ubitux> the flexibility you are asking for seems to be allowing the user to drive the filter himself
[12:30:48 CEST] <ubitux> but i have the feeling it will be a hindrance for simple usages
[12:30:54 CEST] <ubitux> (or maintainability)
[12:31:08 CEST] <wm4> or easily connect filters with other filters
[12:31:50 CEST] <ubitux> why would the user have to do that glue code between filters
[12:32:07 CEST] <ubitux> if we need better communication between filters, we have such things
[12:32:17 CEST] <ubitux> we can use C directly, or more abstract mechanism like metadata
[12:32:29 CEST] <ubitux> or simply streams of data
[12:32:36 CEST] <ubitux> (or commands..)
[12:34:04 CEST] <ubitux> wm4: anyway, did you see my suggestion to rcombs the other day about the subtitles tests?
[12:34:21 CEST] <ubitux> you seemed to be willing to add some tests or something
[12:34:28 CEST] <ubitux> and i hadn't the time to followup on this
[12:35:23 CEST] <wm4> I think this is more about the gritty details of the API... lavfi is not made for flexible filter configuration, e.g. it needs to "reconfigure" the while graph to do anything, and then wants control over the whole graph (as I understand)
[12:35:33 CEST] <wm4> I don't know what rcombs suggested for subtitles
[12:37:26 CEST] <ubitux> wm4: https://dpaste.de/Q18B/raw
[12:49:25 CEST] <wm4> ubitux: ah, I was thinking more about adding tests for features which I've added
[12:50:15 CEST] <ubitux> ah, sure
[12:50:20 CEST] <ubitux> what features?
[12:50:27 CEST] <ubitux> utf16?
[13:18:09 CEST] <wm4> yeah
[13:41:33 CEST] <cone-596> ffmpeg 03Michael Niedermayer 07master:c9ac9adb46f2: avcodec/h264: Remove failure on data partitioning
[14:07:20 CEST] <jlbhshluekg> .
[14:07:21 CEST] <jlbhshluekg> did usa intelligence supply isis with weapons like they did with al-qaeda to justify creating wars?
[14:07:21 CEST] <jlbhshluekg> did usa excute the creative mess in the middle east like they said they will, does the creative mess include explosions with uncertain responsibles to create wars?
[14:07:22 CEST] <jlbhshluekg> plz, send my qs to help limiting usa & israel aggression against others& may then lessen number of people killed in the middle east.
[14:07:22 CEST] <jlbhshluekg> .did usa intelligence supply isis with weapons like they did with al-qaeda to justify creating wars?
[14:25:46 CEST] <anshul_mahe> ubitux: did u find time, to review my cc patch. 
[14:27:23 CEST] <anshul_mahe> there is also webvtt patch romaing around from lots of days, if someone have time to give input, what changes are bad and need to improved. it would merge in mainstream
[17:46:27 CEST] <Daemon404> so uh
[17:46:36 CEST] <Daemon404> whats with the italian email on ffmpeg-devel
[17:46:41 CEST] <Daemon404> saste ^
[17:48:37 CEST] <BtbN> It's importanti
[17:48:44 CEST] <Daemon404> indeed
[17:49:25 CEST] <JEEBsv> something regarding budgeting, revenue sources and donations
[17:50:20 CEST] <Daemon404> that sounds very italian indeed
[18:33:51 CEST] <saste> Daemon404, oh my god
[18:33:57 CEST] <saste> mutt did a mess
[18:34:06 CEST] <saste> that was not supposed to be broadcasted
[18:34:13 CEST] Action: saste palmface...
[18:34:38 CEST] <Daemon404> oh dear
[18:34:54 CEST] <JEEBsv> my condolences
[20:30:20 CEST] <Compn> its been a while since i've seen any video clip without "lavf" as its creator... ;D
[20:32:25 CEST] <j-b> creating broken files since years \o/
[20:36:15 CEST] <wm4> j-b: like the slight timeshift in mkv file, and nobody cares
[20:36:19 CEST] <wm4> lipsync is for losers
[20:37:20 CEST] <Compn> yeah, should be using mencoder ;)
[20:39:00 CEST] <wm4> mencoder creates mkv files nothing but mplayer can play
[20:39:05 CEST] <wm4> (messed up timestamps)
[20:39:09 CEST] <Compn> it works! :)
[20:40:48 CEST] <jamrial> most rips i've seen were made with mkvtoolnix anyway
[21:51:57 CEST] <cone-596> ffmpeg 03Vittorio Giovara 07master:6503cbf842a0: ffv1enc: Add const attribute to input frame
[21:51:58 CEST] <cone-596> ffmpeg 03Michael Niedermayer 07master:9aff7d41f65a: Merge commit '6503cbf842a026faec517eb980551089168c7d8a'
[22:22:43 CEST] <cone-596> ffmpeg 03Carl Eugen Hoyos 07master:3ecb82dd4110: dnxhddec: Fix pixel format change
[22:22:44 CEST] <cone-596> ffmpeg 03Michael Niedermayer 07master:c8de8f7e6438: Merge commit '3ecb82dd4110108244373f007710df4c2ae572b6'
[22:30:52 CEST] <cone-596> ffmpeg 03Ronald S. Bultje 07master:a03b057ede53: vp9: profile 1 header decoding.
[22:30:53 CEST] <cone-596> ffmpeg 03Ronald S. Bultje 07master:2d0bea471958: vp9: increase buffer sizes for non-420 chroma subsamplings.
[22:30:54 CEST] <cone-596> ffmpeg 03Ronald S. Bultje 07master:6019002f0fb4: vp9: support non-420 chroma subsampling for profile 1 token decoding.
[22:30:55 CEST] <cone-596> ffmpeg 03Ronald S. Bultje 07master:ed3e0cc715a9: vp9: take chroma subsampling into account when walking the block tree.
[22:30:56 CEST] <cone-596> ffmpeg 03Ronald S. Bultje 07master:d2aa6f65db1a: vp9: use correct chroma subsampling for profile 1 intra block recon.
[22:30:57 CEST] <cone-596> ffmpeg 03Ronald S. Bultje 07master:d382cc0fad1d: vp9: use correct chroma subsampling for profile 1 inter block recon.
[22:30:58 CEST] <cone-596> ffmpeg 03Ronald S. Bultje 07master:844799b033d6: vp9: invert order of two conditions.
[22:30:59 CEST] <cone-596> ffmpeg 03Ronald S. Bultje 07master:ba59a5ec4242: vp9: split out loopfilter luma rows/cols functions from loopfilter_sb().
[22:31:00 CEST] <cone-596> ffmpeg 03Ronald S. Bultje 07master:afa9d00b433f: vp9: merge uv loopfilter code into generic filter_plane_rows/cols().
[22:31:01 CEST] <BBB> oh yay
[22:31:01 CEST] <cone-596> ffmpeg 03Ronald S. Bultje 07master:e4f77e623977: vp9: don't create special u/v filter masks for 444.
[22:31:02 CEST] <cone-596> ffmpeg 03Ronald S. Bultje 07master:f052ea6ba54e: vp9: add fate test for profile 1 444.
[22:31:03 CEST] <cone-596> ffmpeg 03Ronald S. Bultje 07master:cf62f2d2d6e7: vp9: more specifically specify mask destination to mask_edges().
[22:31:04 CEST] <cone-596> ffmpeg 03Ronald S. Bultje 07master:2a7d7bef6280: vp9: fix mask_edges and filter_plane_rows/cols() for 440.
[22:31:05 CEST] <cone-596> ffmpeg 03Ronald S. Bultje 07master:7126c80f0d2c: vp9: add yuv440 fate test.
[22:31:06 CEST] <cone-596> ffmpeg 03Ronald S. Bultje 07master:6f63bdbae65d: vp9: copy bug in libvpx for 4:2:2 chroma bs=8x4/4x4 prediction.
[22:31:07 CEST] <cone-596> ffmpeg 03Ronald S. Bultje 07master:b005d0976461: vp9: add fate test for 422.
[22:31:08 CEST] <cone-596> ffmpeg 03Michael Niedermayer 07master:d56245f77035: Merge remote-tracking branch 'rbultje/vp9-profile1-wip'
[22:34:51 CEST] <nevcairiel> yay spam
[22:35:05 CEST] <nevcairiel> does that support everything in profiel 1 now?
[22:35:38 CEST] <nevcairiel> reads like its only chroma, not yet bitdepth
[22:36:03 CEST] <nevcairiel> (or is that another profile then?)
[22:36:49 CEST] <nevcairiel> Ah found the info, thats profile 2 then
[22:37:06 CEST] <nevcairiel> or 2/3 rather
[00:00:00 CEST] --- Sat May  2 2015


More information about the Ffmpeg-devel-irc mailing list