[Ffmpeg-devel-irc] ffmpeg-devel.log.20131005

[burek]

[02:03] <cone-451> ffmpeg.git 03Michael Niedermayer 07master:8cb7d2056705: avcodec/dpx: fix reading 12bit dpx images, which have non zero padding bits
[02:03] <cone-451> ffmpeg.git 03Michael Niedermayer 07master:fb726ef09a40: avcodec/dpx: fix reading of the packing field
[02:03] <cone-451> ffmpeg.git 03Michael Niedermayer 07master:23f105a83086: avcodec/dpx: check encoding
[02:04] <cone-451> ffmpeg.git 03Michael Niedermayer 07master:a927276fcc20: avcodec/dpx: check packing
[03:21] <BBB> ubitux: cool - ask if you need any help
[03:21] <BBB> ubitux: I'm trying to do the 32x32 2d-idct
[03:21] <BBB> ubitux: you were doing the 4x4 right?
[03:27] Action: Compn wonders why he has a 1gb sample of avrp codec
[03:36] <Compn> oh man
[03:37] <Compn> i love that pdf file with the broadcast rules
[03:37] <Compn> uk has a lot of crazy laws against 'the occult' :D
[03:37] <Compn> also , Rule 2.11 Subliminal Images
[03:37] <Compn> An image (however brief) that can be seen by viewers is not subliminal.
[03:41] <Compn> max headroom was right! 
[06:08] <ubitux> BBB: yes i'm on 4x4
[10:57] <cone-30> ffmpeg.git 03Martin Storsjö 07master:b698542ad832: mov: Don't allocate arrays with av_malloc that will be realloced
[10:57] <cone-30> ffmpeg.git 03Michael Niedermayer 07master:a7630e3755de: Merge commit 'b698542ad83284fbb8c22404e3cafeb2dd739d38'
[11:28] <cone-30> ffmpeg.git 03Martin Storsjö 07master:d347a7b248d4: ismindex: Use the individual stream duration instead of the global one
[11:28] <cone-30> ffmpeg.git 03Michael Niedermayer 07master:5de251a3e32e: Merge commit 'd347a7b248d4ffdc278373fecf033b0ade030343'
[11:35] <cone-30> ffmpeg.git 03Martin Storsjö 07master:583a287f6f48: ismindex: Calculate the file duration among the included tracks
[11:35] <cone-30> ffmpeg.git 03Michael Niedermayer 07master:47a57c6832d0: Merge commit '583a287f6f48f263859df142e6f90993e4217c72'
[11:41] <cone-30> ffmpeg.git 03Martin Storsjö 07master:a4435f9235ee: ismindex: Change the duration field to int64_t
[11:41] <cone-30> ffmpeg.git 03Michael Niedermayer 07master:922596e3dac8: Merge remote-tracking branch 'qatar/master'
[14:01] <cone-30> ffmpeg.git 03Paul B Mahol 07master:9665217a75d9: avfilter/af_silencedetect: support more sample formats
[14:30] <cone-30> ffmpeg.git 03Paul B Mahol 07master:c3e0fbc507fc: avcodec/dpx: return proper error code for unsupported files
[14:30] <cone-30> ffmpeg.git 03Paul B Mahol 07master:0841652b548b: avcodec/dpx: reformat avpriv_report_missing_feature messages
[14:30] <cone-30> ffmpeg.git 03Paul B Mahol 07master:fc83c99c0b4e: avcodec/dpx: cosmetics: fix identation
[14:35] <durandal_1707> michaelni_: lol, why you are optimizing configure, imho there are bunch of other things more worthy to speed up....
[14:37] Action: durandal_1707 doesn't run configure, modifies config.h/mak by hand
[14:56] <Compn> kostya says that viv h263 is just simple modified h263 , and the ivvideo.dll contains the codebooks in plaintext (strings ivvideo.dll ) , and an expert in h263 decoding would be able to get it working quickly :)
[14:56] <Compn> sample > http://samples.ffmpeg.org/vivo/Blur_-_Song_2.viv
[14:56] <Compn> probably not that easy tho for most people 
[14:58] <Compn> if someone wanted to work on that old old format :)
[14:59] <durandal_1707> there is also audio that need some work, there is even source code for it
[15:00] <Compn> the siren audio ?
[15:00] <Compn> wasnt there some siren patches? hmm
[15:03] <durandal_1707> no, there is extremly ugly patch that puts libsirensomething into lavc
[15:04] <durandal_1707> something like currently libmpcodecs is in lavfi
[15:09] <Compn> ehehe
[15:12] <durandal_1707> Compn: from where you got SCPR ?
[15:19] <Compn> you can download it http://www.infognition.com/ScreenPressor/
[15:19] <Compn> i think kostya might look at it. he loves lossless codecs ;)
[15:21] <Compn> Before ScreenPressor, world leadership in screen captured video compression was held by MSU Screen Capture Lossless Codec. Now it's obsolete and not supported anymore. The MSU codec's author Dmitry Popov, founder of Infognition Co. Ltd. (previously known as Dee Mon), is also the author of ScreenPressor.
[15:21] <durandal_1707> and speed?
[15:22] <Compn> same guy that did msu screen codec, so probably same speed as that
[15:22] <Compn> but there is a graph there that compares speed i think
[15:22] <Compn> same as tscc ?
[15:23] <durandal_1707> is doesn't compare with bunch of other lossles codecs.., just some old codecs
[15:24] <Compn> i havent tested
[15:24] <Compn> at least you can encode using mencoder :)
[15:24] <durandal_1707> and its from 2009...
[15:25] <Compn> yes another useless outdated lossless codec
[15:25] <Compn> did you think it was important ?
[15:25] <Compn> or were you curious about it vs ffv1 ?
[15:31] <durandal_1707> Compn: the lossy thing looks like just removing colors from video, you could do that with any encoder
[15:33] <durandal_1707> for lossless part one would need to inspect actual binary...
[15:35] <Compn> durandal_1707 : the binary is up on mphq if you want it
[15:35] <Compn> samples/drivers32/new/scpr.dll
[15:39] <durandal_1707> 'For feature parity with Windows.'
[17:39] <kierank> Compn: i think it's the same guy on doom9
[17:39] <kierank> who is paranoid of microsoft stealing his ideas
[17:59] <Compn> just another lossless codec to me, i dont care. :P
[18:00] <Compn> got one use report about it
[18:00] <durandal_1707> from where?
[18:08] <michaelni> fate vs. ocz vertex2, first round, fate wins ocz dead
[18:09] <durandal_1707> oh!, after single use?
[18:09] <michaelni> yes, single 800 day 24h per day use or so
[18:10] <durandal_1707> just for reading?
[18:10] <michaelni> yesterday: Error SMART Values Read failed: scsi error aborted command
[18:10] <michaelni> yesterday: SMART overall-health self-assessment test result: UNKNOWN!
[18:11] <michaelni> today: no ocz disk is recognized anymore
[18:52] <cone-605> ffmpeg.git 03Michael Niedermayer 07master:4b05906e8405: configure: add --tempprefix option
[19:06] <durandal_1707> saste: how i can use interleave filter?
[19:07] <durandal_1707> i get dropped frames
[19:22] <saste> durandal_1707, yes
[19:22] <saste> you can't prevent that
[19:22] <saste> unless we raise the buffer queue size
[19:23] <durandal_1707> it gives single frame output
[19:23] <durandal_1707> why it drops frames at all?
[19:23] <durandal_1707> it seems some pts issue...
[19:27] <durandal_1707> the interleave thins seems pretty trivial, why would it not work at all?
[19:31] <saste> durandal_1707, i tried to explain why it can't alwasy work in the documentation
[19:38] <durandal_1707> this is trivial filter, and not having it work points either at bug in filter itself or in dumb lavfi design or dumb lavfi implementation
[19:39] <durandal_1707> by having broken filter it hurts project more than not having it at all
[19:40] <durandal_1707> it would also need be some kind of obvious move to report it at least as bug on tracker
[20:01] <durandal_1707> saste: i just splitted stereo3d sbs2l sample to left and right file and than tried to use interleave2,stereo3d=al:sbsl. to get again same file
[20:01] <durandal_1707> and it did not work at all
[20:02] <durandal_1707> frame are supposed to have same pts in both cases
[20:02] <durandal_1707> it drops all frames except 1
[20:03] <durandal_1707> perhaps filter need to use framesync utility?
[20:03] <durandal_1707> because obviously something is really fishy in current implemenatation, or I'm really missing point
[20:04] <Daemon404> 42
[20:04] <durandal_1707> 69
[20:04] <Daemon404> heh.
[20:05] <durandal_1707> Daemon404: what channels is that?
[20:05] <Daemon404> a private one.
[20:06] <durandal_1707> so all day you just cycle from one irc window to another?
[20:08] <Daemon404> all day? no. but generally i pay attention to more than one irc channel.
[20:08] <Daemon404> thats kind of how irc works.
[20:14] <durandal_1707> saste: will you comment on avutil/parseutils: add more resolutions.
[20:22] <durandal_1707> Daemon404: what is the status of the lavfi deadlock?
[20:24] <Daemon404> the status is i am enjoying my weekend
[20:26] <durandal_1707> lol
[20:32] Action: durandal_1707 sigh
[20:54] <wm4> michaelni: you commented you want to allow setting filters via AVOption... if you want that, maybe you could implement that before my patch gets possibly accepted? because my patch would definitely need 2 new API functions to set the filters, as using the other functions to set it would interfere with my goals
[21:17] <cone-605> ffmpeg.git 03Paul B Mahol 07master:746e3142eee4: avfilter/f_interleave: use the name 's' for the pointer to the private context
[21:33] <michaelni> wm4, ok ill take a look
[21:39] <saste> durandal_1707, i'm mildly against name -> s patches
[21:39] <durandal_1707> saste: i'm fixing that filter
[21:39] <durandal_1707> and s is used in whole codebase
[21:40] <saste> i don't like "s", i'll change it to "x"
[21:40] <durandal_1707> do it after i fix it
[21:40] <saste> durandal_1707, did you manage to find the problem?
[21:41] <saste> i didn't test it yet, i was afk
[21:45] <saste> you can ignore the ending label, or reference it with -map
[21:46] <saste> otoh the example is confusing
[21:47] <saste> ^^ wrong chan...
[23:00] <saste> Docs are omitted as i expect long bikesheds on the syntax and which separator char to use
[23:00] <saste> :D
[00:00] --- Sun Oct  6 2013