[FFmpeg-devel-irc] IRC log for 2010-02-06

[irc at mansr.com]

[00:06:28] <CIA-17> ffmpeg: stefano * r21649 /trunk/ffplay.c:
[00:06:29] <CIA-17> ffmpeg: Remove unused variable ref_clock, fix the warning:
[00:06:29] <CIA-17> ffmpeg: ffplay.c: In function ?compute_frame_delay?:
[00:06:29] <CIA-17> ffmpeg: ffplay.c:1064: warning: unused variable ?ref_clock?
[00:11:29] <CIA-17> ffmpeg: stefano * r21650 /trunk/ffmpeg.c:
[00:11:30] <CIA-17> ffmpeg: Add parentheses in boolean expression: (A && B || C) => ((A && B) || C),
[00:11:30] <CIA-17> ffmpeg: fix the warning:
[00:11:30] <CIA-17> ffmpeg: ffmpeg.c: In function ?output_packet?:
[00:11:30] <CIA-17> ffmpeg: ffmpeg.c:1317: warning: suggest parentheses around && within ||
[00:34:11] <CIA-17> libswscale: stefano * r30518 /trunk/libswscale/swscale.h:
[00:34:11] <CIA-17> libswscale: Satisfy consistency need:
[00:34:11] <CIA-17> libswscale: int * sws_getCoefficients -> int *sws_getCoefficients.
[00:34:11] <CIA-17> libswscale: stefano * r30519 /trunk/libswscale/swscale.h: Apply misc fixes to sws_getCoefficients() doxy.
[00:34:11] <CIA-17> libswscale: stefano * r30520 /trunk/libswscale/utils.c: Simplify how usesVFilter and usesHFilter are set.
[02:38:39] <elenril> MKV files don't work on bloody anything reliably except VLC, even though they're theoretically an h264 variant.
[02:38:48] * elenril lols at slashdot comments
[02:56:48] <CIA-17> ffmpeg: michael * r21651 /trunk/libavcodec/h264_direct.c:
[02:56:48] <CIA-17> ffmpeg: Long term references behave different from short term in spatial direct MV calculation.
[02:56:48] <CIA-17> ffmpeg: Fixes FRExt_MMCO4_Sony_B / issue1251
[03:21:18] <microchip_> elenril: waaaa? LMAO!
[03:28:13] <saintdev> lmao: http://lists.mplayerhq.hu/pipermail/ffmpeg-devel-irc/2010-February/000008.html
[03:29:06] <saintdev> looks like trollbot was sufficiently confuzzled
[03:30:13] <Kovensky> lol
[03:31:37] <microchip_> saintdev: why i see music notes in michael's commit's? lol
[03:32:07] <elenril> because he's secretly japanese?
[03:32:20] <microchip_> lol
[03:32:27] <Kovensky> microchip_: lol
[03:32:34] <Kovensky> because you're using a wrong encoding :P
[03:32:57] <ohsix> the mail is
[03:33:01] <Kovensky> those are mIRC color codes that trollbot doesn't strip
[03:33:28] <saintdev> Kovensky: no they're not
[03:35:11] <Kovensky> ugh
[03:35:13] <saintdev> Kovensky: Ｔｈｅｙ＇ｒｅ ｕｔｆ８ ｆｕｌｌｗｉｄｔｈ ｃｈａｒａｃｔｅｒｓ．
[03:35:23] <Kovensky> yes, I just saw down the log
[03:35:27] <Kovensky> I was thhinking about the CIA commits
[03:35:35] <saintdev> oic
[03:35:37] <Kovensky> the \x0f need to be stripped too
[03:36:45] <microchip_> it'll be nice to CIA to report michael's mood... when he's listening to music, you get notes, when he's angry or mischievous, you get lil devils, etc :p
[03:38:10] <saintdev> Kovensky: Ｍｙ ｆａｖｏｒｉｔｅ ｉｓ ｔｈｅ ｓｅｎｔｅｎｃｅ ｗｉｔｈ ａ ｐｌａｉｎｔｅｘｔ michael ｔｈｒｏｗｎ ｉｎ ｔｈｅ ｍｉｄｄｌｅ ｏｆ ｉｔ ＝Ｄ
[03:39:56] <saintdev> Ｈｅ＇ｓ ｇｏｉｎｇ ｔｏ ｈａｔｅ ｍｅ ａｎｄ ｔｈｉｎｋ Ｉ＇ｍ ａｌｗａｙｓ ｓａｙｉｎｇ ｂａｄ ｔｈｉｎｇｓ ａｂｏｕｔ ｔｈｅ H.264 ｄｅｃｏｄｅｒ．
[03:40:58] <Dark_Shikari> lol
[03:41:07] <Dark_Shikari> mru: ping
[03:41:20] * Dark_Shikari spent the past two days at Green Hills Software, thought you might be interested
[03:41:28] <Dark_Shikari> Also, they have an ARM compiler that's 10% better than armcc
[03:42:53] <Kovensky> saintdev: ｏ ｕ
[03:43:36] * elenril wonders why are so many people awake at 4 in the morning
[03:43:46] <Dark_Shikari> it's 7:40 PM
[03:44:15] <Kovensky> it's 0:45
[03:44:17] <elenril> evil lies
[03:44:30] <Compn> 10:54 here :P
[03:44:35] <Compn> but my clock is fast
[03:44:39] <Kovensky> Dark_Shikari: west coast?
[03:45:02] <Compn> Dark_Shikari : you know of any shakey-cam removal / video smoothing software?
[03:45:20] <Compn> i want to watch a shakey cam movie without vomitting
[03:45:26] <Dark_Shikari> Kovensky: yes
[03:45:29] <Dark_Shikari> Compn: avisynth, depan
[03:45:33] <Kovensky> deen the motion away
[03:45:37] <Kovensky> :D
[03:45:43] <Compn> deen?
[03:45:44] <Compn> thanks Dark_Shikari
[03:45:59] <Kovensky> srsly, there's an anime that I'm watching this week
[03:46:05] <Kovensky> the encode is practically an eDeen() demo
[03:46:18] <Compn> NOOOOOOOOOOOOOOOOOOOO
[03:46:30] <Compn> twitter results live in google search
[03:46:44] <Kovensky> Compn: what
[03:46:56] <Compn> google has live twitter in its search results...
[03:49:57] <Kovensky> o_o'
[04:39:31] <Compn> http://www1.mplayerhq.hu/~compn/shaken.avi
[04:39:54] <Compn> this is the future of movies guys ^^
[04:40:43] <Dark_Shikari> LOL
[04:41:00] <Dark_Shikari> can someone say pan and scan
[04:41:19] <Kovensky> pan and scan
[04:41:21] <Kovensky> there
[04:41:28] <Kovensky> anyway, night ._.
[07:46:35] * elenril wonders what would be a good place to set muxer ident string globally
[07:47:31] <kshishkov> to set what?
[07:48:36] <elenril> encoder="Lavf xx.xx"
[07:49:55] <elenril> doing it as a special case in each muxer is annoying
[07:54:52] <jai> avformat.h?
[07:55:29] <kshishkov> wince all muxers do it in their own way, no problem with special cases
[07:57:58] <elenril> no there is a problem
[07:58:17] <elenril> with the new metadata system i should be able to do 'for tag in metadata: write_tag()'
[07:59:04] <elenril> having to handle writing muxer ident separately makes it more complicated than necessary
[08:01:07] <elenril> hmm, av_write_header() seems like a nice place
[08:02:45] <jai> question : is it okay to commit trivial cleanups?
[08:02:50] <jai> or do i send patches
[08:14:08] <kshishkov> trivial is fine
[08:24:49] <jai> kshishkov: k, thx
[10:35:41] * _av500_ is at ff booth ff is not :(
[10:39:33] * janneg guesses ff had a beer too much yesterday
[10:45:11] <jai> heh
[10:45:24] <jai> _av500_: try and hook up with mru maybe
[11:12:01] <CIA-17> ffmpeg: michael * r21652 /trunk/ffmpeg.c: (log message trimmed)
[11:12:01] <CIA-17> ffmpeg: Revert r21615 | cehoyos | 2010-02-02 23:23:09 +0100 (Tue, 02 Feb 2010) | 5 lines
[11:12:01] <CIA-17> ffmpeg:  Guess the duration before converting video and write guessed duration
[11:12:01] <CIA-17> ffmpeg:  into flv header.
[11:12:01] <CIA-17> ffmpeg:  Patch by Justin Johnson, justin D johnson3 A gmail
[11:12:02] <CIA-17> ffmpeg: The change to ffmpeg.c is wrong as input and output files arent connected 1:1
[11:12:02] <CIA-17> ffmpeg: (i seem to have missed that)
[12:03:20] <jai> elenril: is there a reason why "ISFT" is commented out?
[12:03:44] <jai> i havent read the thread in detail so might be missing something though
[12:04:07] <elenril> because it means 'encoder' so it's treated differently
[12:05:21] <elenril> i'm working on a patch that would remove this special case atm
[12:07:07] <jai> right, asked because i saw it in the conv table
[12:16:36] <elenril> can you apply it?
[12:19:52] <jai> elenril: sure
[12:21:39] <jai> elenril: running regtests
[12:21:48] <elenril> they should pass
[12:21:56] * elenril tested this morning
[12:33:36] <CIA-17> ffmpeg: jai_menon * r21653 /trunk/libavformat/ (avidec.c Makefile avi.c avi.h avienc.c):
[12:33:36] <CIA-17> ffmpeg: Add AVI metadata conversion table.
[12:33:36] <CIA-17> ffmpeg: Patch by Anton Khirnov (gmail{wyskas}).
[12:34:26] <elenril> thanks
[12:34:41] <jai> np
[12:39:33] <CIA-17> ffmpeg: jai_menon * r21654 /trunk/libavcodec/alac.c: alac.c : Use av_freep instead of av_free.
[12:40:35] <kierank> damn this stupid software seems to find a way to figure out that it's being run in a vm
[12:40:51] <elenril> lol
[12:41:03] <elenril> what's so secret about it anyway?
[12:41:55] <elenril> hmm, i get ridiculous pts when encoding with ffaac
[12:42:01] <elenril> is this expected/known?
[12:42:01] <kierank> kept secret so that it can be sold for thousands of pounds
[12:42:09] <jai> elenril: Red pill blue pill
[12:42:27] <kshishkov> elenril: it's ridiculous you encode with ffaac
[12:42:50] <elenril> just wanted something to mux to avi
[12:43:01] <elenril> probably should have used flac
[12:43:01] <kshishkov> not that encoder did something with pts though
[12:49:16] <elenril> who is the maintainer of ogg/flac
[12:49:26] <CIA-17> ffmpeg: jai_menon * r21655 /trunk/libavcodec/alac.c: Remove redundant use of numchannels since it is 1 for mono.
[12:53:11] <CIA-17> ffmpeg: jai_menon * r21656 /trunk/libavcodec/alac.c: Avoid using an intermediate variable.
[12:53:23] <elenril> bcoudrier is on fosdem?
[12:54:17] <CIA-17> ffmpeg: reimar * r21657 /trunk/libavcodec/ (h261enc.c h261dec.c): Add include necessary for ff_set_qscale
[12:54:30] <jai> elenril: possibly justin or Yuvi
[12:55:18] * elenril wants vorbiscomment writing patch applied
[12:57:43] <jai> elenril: was it approved?
[12:59:12] <elenril> i don't think so
[13:00:03] <elenril> it seems jruggles approved it, but wanted some comments from others
[13:01:47] <jai> right, its been 3 days since the updated patch was sent
[13:02:02] <jai> so maybe wait for a week to go by and ping again
[13:05:10] <jai> also, i applied a few trivial patches i had for ffalac, if michael is reading this, i hope you are okay with it :)
[13:16:07] <CIA-17> ffmpeg: jai_menon * r21658 /trunk/ffplay.c: cosmetics : re-indent after r21639.
[13:55:49] * kierank reads kshishkov's blog post
[13:59:53] * kshishkov did not
[14:01:29] <janneg> the blog is write only memory?
[14:02:00] <kshishkov> looking at most blogs it is
[15:05:23] <kierank> could qemu be used for disassembly
[15:09:43] <jai> kierank: wat
[15:10:13] <kierank> i think that might work actually
[15:10:56] <kierank> because you can dump the current instructions without the kernel module getting involved
[15:31:11] <kshishkov> it can
[16:35:49] <elenril> who runs git.ffmpeg.org?
[16:36:21] <kierank> KotH
[16:37:33] * elenril summons KotH from depths of hell
[16:38:05] <kshishkov> well, "Belgium" is shorter
[16:39:42] * elenril should go to fosdem too some time
[16:39:59] * kshishkov should go wherever
[16:40:58] <kierank> i think the html5 talk starts now
[16:41:07] <kshishkov> heh
[16:41:15] <elenril> that could be funnny
[16:41:16] <kshishkov> tell us when HTML5 flames start
[16:43:54] <kierank> might be a video stream
[16:53:30] <CIA-17> ffmpeg: siretart * r21659 /trunk/configure:
[16:53:30] <CIA-17> ffmpeg: libgsm installs headers in a subdirectory, use gsm.h from that subdirectory.
[16:53:30] <CIA-17> ffmpeg: Approved by Diego at FOSDEM.
[16:56:52] <_av500_> gm
[16:57:05] <kshishkov> nv
[16:57:41] <_av500_> html talk was too packed to get in....
[16:59:08] <Compn> ooooo man
[16:59:15] * Compn wants to see more html5 flames
[16:59:17] <Compn> FLAME ON
[16:59:40] <ohsix> :M
[17:10:31] <elenril> http://arstechnica.com/tech-policy/news/2010/02/studios-crushed-isp-cant-be-forced-to-play-copyright-cop.ars/ << awesome
[17:13:00] <jai> elenril: lol @ treasure analogy
[17:14:16] <jai> Compn: was that a Human Torch reference
[17:17:36] <Compn> i guess so
[17:42:33] <Kovensky> lolaustralia
[17:42:48] <Kovensky> 14:21.29 <&Emess> well
[17:42:49] <Kovensky> 14:21.33 <&Emess> after seeing that
[17:42:54] <Kovensky> 14:21.34 <&Emess> I lol'd
[17:43:04] <Kovensky> 14:21.37 <&Emess> because
[17:43:04] <Kovensky> 14:21.39 <&Emess> most ISP's here dont give a shit
[17:43:04] <Kovensky> 14:21.49 <&Emess> seeing as all their employees fileshare like teh fist of the north star
[17:43:07] <Kovensky> 14:21.55 <&Emess> AND iinet is a west coast isp
[17:43:20] <Kovensky> 14:22.13 <&Emess> the big west coast isp's have something called WAIX
[17:43:20] <Kovensky> 14:22.18 <&Emess> essentially a high speed no cost network over the whole state
[17:43:24] <Kovensky> 14:22.24 <&Emess> that lets people share stuff without a quota
[17:43:26] <Kovensky> 14:22.32 <&Emess> although iinet withdrew from it last year the fuckers
[17:43:28] <Kovensky> 14:22.37 <&Emess> anyway
[17:43:31] <Kovensky> 14:22.44 <&Emess> most isps here are like talk to the cops, we dont care
[17:43:33] <Kovensky> 14:22.56 <&Emess> my own isp hosts 4 trackers that I am aware of even
[17:43:36] <Kovensky> 14:23.08 <&Emess> along with an FTP that downloads latest tv caps and shit in HD every week
[17:43:43] <twnqx> come on
[17:43:53] <twnqx> i know more than one european or US isp where it's the same
[17:43:56] <Kovensky> hi charlieb
[17:44:03] <twnqx> it's just normal.
[17:44:11] <kierank> not any more now this channel is logged ;)
[17:44:35] <twnqx> this channel was always logged
[17:44:49] <kshishkov> now it's explicitly logged
[17:44:50] <kierank> not publically
[17:44:57] <twnqx> i know
[17:45:08] <twnqx> but it doesn't change a thing
[17:45:54] * Kovensky has logs dating back to early 2009
[17:46:27] <kierank> but the logs appear in google
[17:46:47] <kierank> i.e. I search for stuff about smpte specs and find discussion of me
[17:47:10] <kierank> be nice to have robots.txt on that mailing list
[17:48:15] <Kovensky> I wonder how seriously will the logs here be taken
[17:49:24] <kshishkov> kierank: so you are expecting a visit from nice guys of D*lby yet?
[17:49:40] <kierank> not them i'm worried about
[17:50:05] <kierank> people quoting me from here out of context is more of a problem
[17:51:54] <kierank> also it seems this stupid drm can detect qemu
[17:54:55] <twnqx> wow
[17:55:00] <twnqx> that's pretty advanced
[17:55:43] <kshishkov> no, virtual machine or debugger detection is rather old
[17:56:00] <kierank> works nicely in virtualbox for some reason
[17:56:24] <twnqx> debugger is old indeed
[17:56:28] <ohsix> qemu is easy to detect too; virtualbox/vmware use secret knocks and stuff :>
[17:56:49] <twnqx> i once ran into a linux virus that would detect if it was straced >_>
[17:57:03] <kshishkov> wow
[17:57:10] <kshishkov> linux virus FTW
[17:57:18] <ohsix> ptracing yourself is oldschool too
[17:57:20] <twnqx> that was like... hm... 10 years back or so
[17:57:53] <twnqx> i noticed it more because the infected binaries wwouldn't run at all any more
[17:58:03] <twnqx> and it infected /bin
[17:58:52] <kierank> this thing  goes as far as crashing qemu
[17:59:52] <kshishkov> try virtualbox then
[18:00:35] <kierank> virtualbox doesn't have much in the way of built in debugging
[18:08:22] <Compn> hey
[18:08:39] <Compn> heres a spam site installing vlc and ffdshow and other ffmpeg related stuff bundled with spyware >> http://www.video-buzz.com/
[18:09:02] <kshishkov> wanna ask them for malware sources?
[18:09:46] <Compn> no, but you can
[18:13:48] <kierank> lol it has TrustE
[18:14:22] <kierank> whoops wrong camelcase, it's TRUSTe
[18:14:36] <Compn> must be trustworthy
[18:14:46] <Compn> kierank : feel free to put on roundup for me :D
[18:14:56] <Compn> kshishkov : so are you going to merge x264 into ffmpeg?
[18:15:04] * Compn trolls around
[18:15:34] <kshishkov> Compn: no, I was merely saying that in N years it won't be so rapidly developed and will be ready for merge
[18:16:47] <Compn> ah
[19:26:03] <pscherer> Does anyone know if FFmpeg will be part of the google summer of code 2010?
[19:26:17] <kshishkov> only Google does
[19:26:34] <kshishkov> we want to participate though
[19:26:54] <rsk> didnt ffmpeg get the maximum ammount of slots 2009?
[19:27:18] <kshishkov> nope, other projects got slots as well
[19:28:17] <pscherer> If you are part of it again I am interested in doing something for FFmpeg
[19:29:00] * kshishkov is always interested to do something for FFmpeg
[19:29:11] <tetsuo55> hello
[19:29:19] <tetsuo55>  i think i found a sort of bug, an uninitilised variable in ituh263dec.c
[19:29:25] <tetsuo55> if s->mb_intra is true, and CONFIG_RV10_DECODER is not defined, and s->codec_id == CODEC_ID_RV10 is true, level is uninitialized
[19:29:34] <tetsuo55> level seems to have too broad scope, it would make more sense to declare it inside the if(s->mb_intra)
[19:29:42] <tetsuo55> what do you guys think?
[19:29:47] <kshishkov> CONFIG_* vars are always defined
[19:30:09] <tetsuo55> for every build option and platform?
[19:30:14] <kshishkov> yes
[19:30:18] <kshishkov> 0 or 1
[19:30:23] <tetsuo55> awesome
[19:30:23] <tetsuo55> :D
[19:31:39] <tetsuo55> oops
[19:31:57] <tetsuo55> its suppsed to say "when defined as 1" instead of "when not defined"
[19:32:47] <kshishkov> I suspect that condition should not happen in that file anyway
[19:33:48] <tetsuo55> still, better safe than sorry or "wontfix" ?
[19:34:30] <kshishkov> "shouldnothappen"
[19:35:08] <kshishkov> CONFIG_RV10_DECODER=1 enables rv10 decoder, so when it's zero, it will refuse to decode rv10 and that check won't happen
[19:37:36] <tetsuo55> i think thats what he mean
[19:37:38] * tetsuo55 asks
[19:39:09] <kshishkov> no need to be too safe in this case
[19:59:28] <jai> hmm, do we have any aobs with mlp on mphq?
[19:59:45] <kshishkov> aobs?
[20:00:02] <jai> mpeg ps dvd audio
[20:01:27] <kshishkov> no
[20:04:01] <kierank> jai: do you want some?
[20:04:35] <jai> kierank: if you have some, then sure
[20:04:59] <kierank> i can get some
[20:05:01] <jai> kierank: thanks :)
[20:06:18] <jai> kierank: wait, do they have an mlp stream?
[20:06:35] <jai> kierank: if they do, do they demux properly for you with lavf?
[20:10:25] <kierank> yes they are mlp
[20:10:27] <kierank> mediainfo says that
[20:10:30] <kierank> ffmpeg says:
[20:10:30] <kierank>     Stream #0.0[0xa1]: Audio: pcm_s16be, 32000 Hz, 3 channels, s16, 1536 kb/s
[20:10:30] <kierank> At least one output file must be specified
[20:12:45] <jai> kierank: exactly, we need to fix that
[20:13:13] <kierank> gimme a few minutes and i'll upload you a sample
[20:13:21] <jai> kierank: thanks
[20:18:52] <kierank> http://dl.dropbox.com/u/2701213/mlp.aob
[20:21:30] * elenril sighs
[20:21:46] <elenril> i suppose asking if anybody here understands how nut index works would be pointless
[20:22:59] <iive> why?
[20:24:39] <elenril> because i don't think anybody except people who wrote the specs know that ;)
[20:25:56] <iive> please, ask your question and don't smear around.
[21:11:10] <Compn> pscherer : get started working on a qualification task now, if so
[21:11:29] <Compn> check small tasks page on http://wiki.multimedia.cx
[22:43:35] <pscherer> Compn: will do
[23:26:09] <CIA-17> ffmpeg: stefano * r21660 /trunk/doc/APIchanges: Add missing entries for the recent lsws API changes.