[Ffmpeg-devel-irc] ffmpeg-devel.log.20120228

[burek]

[02:23] <darkfires> is there a way to force ffmpeg to this Stream #0.0: Video: mpeg4, yuv420p, 1280x720 [PAR 1:1 DAR 16:9], 25 fps, 25 tbr, 1200k tbn, 30 tbc?   I'm having a problem with an IP camera, it's got a bug in the firmware and the manufact wont respond to my emails, basically the first time you connect the proper vol header is sent, any subsequent connect without a reboot the header is corrupt
[02:23] <darkfires> i've been trying with various options but im getting no where
[03:26] <Compn> darkfires : can you relay the stream so we can access it
[03:26] <Compn> there is a program called socket2socket, part of desproxy project that can do it, if you cant figure it out in your router ...
[03:27] <Compn> and whats ffmpeg say about it after the header is corrupt ? 
[03:27] <Compn> its possible to force it, depends how much data is coming 
[03:27] <Compn> and the format the camera streams 
[03:27] Action: Compn goes afk
[03:53] <CIA-17> ffmpeg: 03Anton Khirnov 07master * rf5f5b154e7 10ffmpeg/doc/APIchanges: doc/APIchanges: fill in missing dates and hashes.
[03:53] <CIA-17> ffmpeg: 03Anton Khirnov 07master * r1d428004c6 10ffmpeg/libavcodec/avcodec.h: 
[03:53] <CIA-17> ffmpeg: lavc: reorder AVFrame fields.
[03:53] <CIA-17> ffmpeg: Put related fields together, move the important ones closer to the
[03:53] <CIA-17> ffmpeg: beginning, try to eliminate holes.
[03:53] <CIA-17> ffmpeg: 03Anton Khirnov 07master * rb1edfce08a 10ffmpeg/libavcodec/ (avcodec.h pthread.c): 
[03:53] <CIA-17> ffmpeg: lavc: reorder AVCodecContext fields.
[03:53] <CIA-17> ffmpeg: Put related fields together, move the important ones closer to the
[03:53] <CIA-17> ffmpeg: beginning, try to eliminate holes.
[03:53] <CIA-17> ffmpeg: 03Justin Ruggles 07master * rf234e8a32e 10ffmpeg/libavformat/movenc.c: 
[03:53] <CIA-17> ffmpeg: movenc: use timestamps instead of frame_size for samples-per-packet
[03:53] <CIA-17> ffmpeg: For encoding, AVCodecContext.frame_size is the number of input samples to
[03:53] <CIA-17> ffmpeg: send to the encoder and does not necessarily correspond directly to the
[03:53] <CIA-17> ffmpeg: timestamps of the output packets.
[03:53] <CIA-17> ffmpeg: 03Tommy Winther 07master * r1a6b9a98ce 10ffmpeg/libavformat/rtsp.c: (log message trimmed)
[03:53] <CIA-17> ffmpeg: rtsp: Handle requests from server to client
[03:53] <CIA-17> ffmpeg: This returns 200 OK for OPTIONS requests and 501 Not Implemented
[03:53] <CIA-17> ffmpeg: for all other requests.
[03:53] <CIA-17> ffmpeg: Even though this doesn't do much actual handling of the requests,
[03:53] <CIA-17> ffmpeg: it makes the code properly identify server requests as such, instead
[03:53] <CIA-17> ffmpeg: of interpreting it as a reply to the client's request as it did
[03:54] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * re60bdb7e5c 10ffmpeg/libavformat/flvdec.c: 
[03:54] <CIA-17> ffmpeg: flvdec: Remove the now redundant check for known broken metadata creator
[03:54] <CIA-17> ffmpeg: The index validation identifies these indexes as broken.
[03:54] <CIA-17> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[03:54] <CIA-17> ffmpeg: 03Martin Storsjö 07master * r7e297a46db 10ffmpeg/libavformat/flvdec.c: (log message trimmed)
[03:54] <CIA-17> ffmpeg: flvdec: Validate index entries added from metadata while reading
[03:54] <CIA-17> ffmpeg: By validating the index entries while reading, we don't need to
[03:54] <CIA-17> ffmpeg: seek at startup to validate the entries. If the error in the
[03:54] <CIA-17> ffmpeg: index entries is not pointing to (our definition of) the start
[03:54] <CIA-17> ffmpeg: of packets, and there is an index entry pointing at some of the
[03:55] <CIA-17> (18 lines omitted)
[04:03] <Compn> long commit messages :P
[07:37] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r8b1cd25ca7 10ffmpeg/libavformat/pmpdec.c: 
[07:37] <CIA-17> ffmpeg: pmpdec: Check for zero audio packets.
[07:37] <CIA-17> ffmpeg: This fixes a division by 0.
[07:37] <CIA-17> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[07:37] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[12:15] <trupheenix> Hi. I need to know if there are any good sources where I can understand about the fundamentals behind video codecs? 
[12:45] <Compn> wikipedia has that stuff trupheenix
[12:45] <av500> also the video from greg
[12:45] <trupheenix> Compn, any good books i can read? or there are no books on this subject?
[12:47] <av500> tried amazon?
[12:47] <trupheenix> av500, no
[12:47] <trupheenix> av500, what video from greg?
[12:50] <av500> http://wiki.xiph.org/A_Digital_Media_Primer_For_Geeks/Episode_01
[13:11] <Compn> trupheenix : oh, you mean video coding and dct and stuff
[13:11] <Compn> i thought you meant ... how codecs work, not the algorithms and such
[13:11] <Compn> yeah there are books
[13:11] <trupheenix> Compn, !!! yes!! :) finally! :)
[13:11] <j-b> where is episode 2?
[13:12] <Compn> trupheenix : http://codecs.multimedia.cx/?p=70
[13:12] <Compn> lists a couple good books
[13:13] <Compn> as for the rest, i dunno, best ask someone else, maybe kostya :)
[13:16] <trupheenix> Thanks Compn. I'm pretty bad at my calculus. Need to see these! :)
[14:09] <kierank> oh he's gone
[14:40] <funman> michaelni: you dont' want to use separate postproc repo because of fate? couldn't fate be changed to fetch postproc?
[15:35] <Compn> cant git do like svn externals ?
[15:36] <Compn> not that i care one way or the other
[15:36] <JEEBsv> yes
[15:36] <JEEBsv> git submodules
[15:51] <michaelni> JEEBsv, submodules are very annoying
[15:52] <av500> use repo
[15:52] <michaelni> funman, ill just keep external and internal in sync
[15:54] <funman> sounds like a plan
[15:55] <michaelni> ive looked at submodules and subtree and a few other things but it all looked like it would be a bigger mess
[15:55] <funman> in that case merging is free
[15:59] <michaelni> Daemon404, ping
[16:00] <michaelni> wth is this: Merge: 3a6e4e6 a792a83 8cf4ef5 59d8d9c 7876565 ???
[16:00] <michaelni> i mean yeah its a merge of 5 branches but
[16:00] <michaelni> why is it there in libpostproc repo
[16:03] <michaelni> also, whats "This file is part of FFmpeg/Libav." ??
[16:03] <michaelni> libpostproc was not developed in libav and isnt part of it either
[16:03] <michaelni> its either part of FFmpeg or part of libpostproc
[16:06] <michaelni> copyright (c) 2006 Michael Niedermayer <michaelni at gmx.at>
[16:06] <michaelni>  This file is part of Libav.
[16:07] <Compn> its part of XXX or YYY , the line is only to provide where the file came from
[16:07] <Compn> altho why it says libav when its from ffmpeg is another question
[16:07] <Compn> similar things were done to files in mplayer that mplayer collected over the years
[16:08] <Compn> i dont remember the reasoning for abandoning the original 'file is from' part, only so people dont copy over incorrect license headers i guess
[16:08] <Daemon404> [10:07] <@Compn> altho why it says libav when its from ffmpeg is another question <-- it has stuff from both sides.
[16:08] <Compn> definitely a question to ask dondiego 
[16:08] <Daemon404> :politics:
[16:08] <michaelni> libav and 2006 ...
[16:09] <Compn> Daemon404 : it should say @Copyright blah blah if that were the case
[16:09] <Daemon404> it has code from both sides.
[16:09] <Compn> the 'file is from' is not a copyright declaration
[16:09] <Daemon404> what
[16:09] <Daemon404> [10:00] < michaelni> wth is this: Merge: 3a6e4e6 a792a83 8cf4ef5 59d8d9c 7876565 <-- i wodnered this myself
[16:09] <Daemon404> it seems to be a biproduct of filter-branch
[16:10] <Daemon404> [10:03] < michaelni> libpostproc was not developed in libav and isnt part of it either <-- it has commits from them
[16:10] <Compn> Daemon404 : i just mean, if libav has changed the file, whoever changed the file should put his copyright on there, assuming it was enough code to add copyright ..
[16:10] <Daemon404> it is not a matter of copyright as far as im concerned
[16:10] <Daemon404> it;s a matter of where it lived 
[16:10] <Daemon404> it lived in both ffmpeg and libav.
[16:11] <Daemon404> and seriously, you gusy can take your politics and shove it.
[16:11] <Daemon404> i will not involve myself in them.
[16:11] <michaelni> <Daemon404> it has code from both sides.
[16:11] <michaelni> then git blame isnt working
[16:11] <Compn> i'm not trying to start a flame so i'll go away...
[16:11] <Compn> michaelni : are you saying there are no changes from libav ?
[16:11] <michaelni> Compn, not in x86_cpu.h
[16:12] <michaelni> or blame -C -w is lying to me
[16:12] <Daemon404> then change it
[16:12] <Daemon404> i seriously do not give a shit about this libav/ffmpeg political bullshit anymore.
[16:12] <Daemon404> do as you wish.
[16:13] <michaelni> Daemon404, what about the 5 head merge ?
[16:13] <Compn> Daemon404 : you sound upset :(
[16:13] <Daemon404> on the technical side there, i dont know what to do to fix that per se, michaelni 
[16:14] <Daemon404> do you want my git filter branch commandss?
[16:14] <michaelni> i could just clone ffmpeg and throw all bit libpostproc out and it wont have such monsters
[16:14] <Daemon404> perhaps you could spot it
[16:14] <Daemon404> Compn, i merely no longer wish to ever hear any of this politics crap again. too many things i do get plagued with it.
[16:14] <Daemon404> i prefer ignorant bliss now.
[16:15] <michaelni> if libpostproc is a true clone i also could do a true git merge to keep them in sync ...
[16:17] <Daemon404> it's not a clone
[16:17] <Daemon404> it's a filtered branch.
[16:17] <michaelni> i know
[16:17] <Daemon404> my intention was to filter out the noise of all the otehr unrelated commits
[16:17] <michaelni> i understand that
[16:17] <michaelni> i dont understand why it has 5 headed monsters in there
[16:19] <Daemon404> i think it may be teh 2nd command i ran... maybe
[16:19] <Daemon404> sec
[16:19] <Daemon404> (it has to do with the merges you did from libav)
[16:20] <Daemon404> http://pastebin.com/raw.php?i=gr9SUP3M
[16:20] <Daemon404> ^ what i used
[16:20] <Daemon404> just a ... bit ... ugly
[16:22] <michaelni> whats your oppinon on using a true clone, its kinda ... less work ?
[16:22] <Daemon404> far less work
[16:22] <Daemon404> less nice lookign history wise
[16:23] <Daemon404> it's up to you.
[16:23] <michaelni> if its up to me, ill pick a clone
[16:23] <Daemon404> ok.
[16:42] <bgmarete> Hello. I am looking for interlaced video sequences to test the yadif filter. Any down-loadable ones from the web? With varying field parity? Or how can I use FFmpeg to create such? Thanks!
[16:46] <cbsrobot> bgmarete: what is varying field parity ?
[16:47] <bgmarete> cbsrobot: I mean I would like some samples to be tff and others bff.
[16:47] <bgmarete> (If possible)
[16:48] <michaelni> i remember generating such (half framerate) sequences with mencoder long ago from progessive input 
[16:48] <michaelni> i dont remember how exactly i did it though ...
[16:49] <michaelni> Tjoppen, there are mxf patches on ffmpeg-dev
[16:56] <Compn> there are some interlaced mpeg2's on sample repo
[16:56] <Compn> but hes gone
[17:01] <mystica555_> heh
[17:02] <mystica555_> at one point i came up with a very good resizing interlace preserving commandline for 1080i to 720x480i dvds.. but, ive modified the script in the meantime and its not the one that actually worked
[17:03] <mystica555_> why i didn't keep using the commandline and up-enter, i'll smack myself for.. i store those logs <.<
[17:19] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r1ebc5e71a0 10ffmpeg/tests/ (lavfi-regression.sh ref/lavfi/pp): 
[17:19] <CIA-17> ffmpeg: tests: add libpostproc test
[17:19] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[17:19] <CIA-17> ffmpeg: 03Philip de Nier 07master * r539a8469ae 10ffmpeg/libavformat/utils.c: 
[17:19] <CIA-17> ffmpeg: add v210 to intra only list to allow "-vcodec copy"
[17:19] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[17:23] <mystica555_> apparently -vf yadif=1,scale=720:480,tinterlace=4  works well for mencoder; i was seeing audio sync issues so i just reencoded the audio on my test instead of copying, and it worked. heh. 
[17:24] <mystica555_> and michaelni that basically does your half framerate (tinterlace=4) from full framerate video (as given by yadif=1)
[17:33] <ubitux> michaelni: is it normal to have ./presets and ./ffpresets directories? (i think i already asked, but i can't remember any answer)
[18:03] <michaelni> ubitux, i dont remember the awnser either :(
[18:06] <ubitux> the option system is quite complicated btw :p
[18:06] <ubitux> for instance if you want to use libx264 profile
[18:07] <ubitux> you can use -profile, -vprofile, profile:v
[18:07] <ubitux> not sure how that work :p
[18:07] <ubitux> some presets reference vprofile, some profile
[19:05] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rddd86a2924 10ffmpeg/ffmpeg.c: 
[19:05] <CIA-17> ffmpeg: ffmpeg: Fix division by 0 due to invalid timebase
[19:05] <CIA-17> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[19:05] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:05] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r4f5d1468f5 10ffmpeg/libavformat/omadec.c: 
[19:05] <CIA-17> ffmpeg: omadec: Fix wrong number of array elements.
[19:05] <CIA-17> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[19:05] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:24] <michaelni> ubitux, -profile is a bit risky due to audio vs video
[19:43] <bcoudurier> kierank, kostya can join the ml as well ;)
[19:45] <Compn> did maxsim work on the encoder or just decoder ?
[19:46] <Compn> unrelated...
[19:46] <Compn> is there an easy way to diff two trunks ?
[19:46] <Compn> like diff two whole dirs ?
[19:49] <Compn> oh, diff does it already
[19:50] <Compn> not subdirs tho :\ crappy
[19:52] <Compn> ack CR vs LF diff
[19:52] <Compn> with git diff
[19:58] <Compn> ok , fix a few line endings and git diff works great
[19:58] <Compn> yay for subdirs 
[19:59] <nevcairiel> you can also make it ignore whitespace differences
[19:59] <Compn> reading git diff manual , i must have missed it
[19:59] <nevcairiel> i would hope that also counts for line ending differences
[20:00] <nevcairiel> its -w
[20:00] <nevcairiel> or --ignore-all-space if you like verbose
[20:00] <Compn> oh yeah its there
[20:00] <Compn> i'm blind i guess
[20:00] <Compn> diff -wb is old thing too, should have tried it :)
[20:37] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rbf72c0410d 10ffmpeg/libavcodec/cook.c: 
[20:37] <CIA-17> ffmpeg: cook: add const to quant_index_table of categorize()
[20:37] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:37] <CIA-17> ffmpeg: 03Philip de Nier 07master * r627cd3841b 10ffmpeg/libavformat/mxf.c: 
[20:37] <CIA-17> ffmpeg: mxf: fix v210 label matching len
[20:37] <CIA-17> ffmpeg: Must check all 16 bytes because there is a planar 10-bit format
[20:37] <CIA-17> ffmpeg: label that has equal first 15 bytes
[20:37] <CIA-17> ffmpeg: Review-by: Baptiste Coudurier <baptiste.coudurier at gmail.com>
[20:37] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:37] <CIA-17> ffmpeg: 03Philip de Nier 07master * r3592bd509c 10ffmpeg/libavformat/mxfdec.c: 
[20:37] <CIA-17> ffmpeg: mxfdec: set track edit rate num/den in expected order
[20:37] <CIA-17> ffmpeg: This matches the order used for the index table edit rate.
[20:37] <CIA-17> ffmpeg: Reviewed-by: Baptiste Coudurier <baptiste.coudurier at gmail.com>
[20:37] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:37] <CIA-17> ffmpeg: 03Philip de Nier 07master * rc2d11a8b86 10ffmpeg/libavformat/mxfdec.c: 
[20:37] <CIA-17> ffmpeg: mxfdec: detect uncomp pictures using essence container ul
[20:37] <CIA-17> ffmpeg: This supports detection of uncompressed picture in files that
[20:37] <CIA-17> ffmpeg: The properties of the CDCI Descriptor are insufficient to specify
[20:37] <CIA-17> ffmpeg: the pixel format for uncompressed picture data. SMPTE 377-1 and
[20:37] <CIA-17> ffmpeg: RP224v10 have defined a set of picture coding labels to indicate what
[20:37] <CIA-17> ffmpeg: formatting was used.
[20:37] <CIA-17> ffmpeg: This patch uses 2 labels to detect UYVY422 or YUYV422 pixel formats.
[20:37] <CIA-17> ffmpeg: 03Philip de Nier 07master * r51d632cc8f 10ffmpeg/libavformat/mxf.c: 
[20:37] <CIA-17> ffmpeg: mxf: add uncompressed 422 8-bit rawvideo label
[20:37] <CIA-17> ffmpeg: Reviewed-by: Baptiste Coudurier <baptiste.coudurier at gmail.com>
[20:37] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:54] <burek> if somebody is watching #ffmeg
[22:54] <burek> this is the perfect example
[22:54] <burek> why do we need static ffmpeg build
[22:54] <burek> :)
[23:12] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rf929abd0c3 10ffmpeg/libavcodec/adpcm.c: 
[23:12] <CIA-17> ffmpeg: adpcm_xa: Check filter validity.
[23:12] <CIA-17> ffmpeg: Fixes out of global array read.
[23:12] <CIA-17> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[23:12] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[00:00] --- Wed Feb 29 2012