Ffmpeg-devel-irc

Download

ffmpeg-devel-irc@ffmpeg.org

February 2012

2 participants
60 discussions

ffmpeg-devel.log.20120218
by burek 19 Feb '12

19 Feb '12

[00:00] <CIA-48> ffmpeg: 03Carl Eugen Hoyos 07release/0.10 * r569cb94869 10ffmpeg/cmdutils.c: [00:00] <CIA-48> ffmpeg: Fix ffmpeg -codecs output. [00:00] <CIA-48> ffmpeg: (cherry picked from commit f6492476a63938cc66c51bf61c88407b7749f780) [00:49] <Compn> i gotta say [00:49] <Compn> libav gsoc tasks look bad [00:51] <Compn> vp7? unused rare and stupid on2 crap. hevc/h265 isnt even a spec yet. rtmpe just duplicating the heck out of librtmp? build system to a new project - trying to duplicate autotools ? [00:51] <Compn> hardware accel api, somehow going to merge vaapi and vdpau into one wrapper, and probably do it without carl's help? [00:52] <kierank> aacs is a pretty bad idea fwiw too [00:52] <kierank> hevc is almost finished so parts can be implemented [00:52] <Compn> i mean , merging vaapi would be great [00:52] <Compn> so thats ok with me, ignore that one [00:53] <Compn> or someone said skype used vp7 [00:53] <Compn> can you record videos from skype ? [00:53] <kierank> yes they do [00:53] Action: Compn has no clue [00:56] <iive> the vaapi/vdpau merger could be a great thing or great disaster, depending on what model they prefer. [00:57] <Compn> wasnt a wrapper for both proposed years ago ? [00:57] <iive> and given they probably make it in spite of carl, I'd say they would pick the wrong one. [00:57] <Compn> vaapi could do it already i thought ? [00:58] <Compn> i mean, you could run vdpau under vaapi [00:58] <iive> yes. [00:58] <Compn> except everyone rejected the way vaapi did it, years ago :\ [01:00] <iive> well, we will see. I hope I would be wrong. [01:01] <michaelni> the main issue i see with h265 is the mentor is not really qualified to mentor it. [01:02] <iive> who is the mentor? [01:02] <michaelni> ronald IIRC [01:02] <iive> BBB? [01:02] <michaelni> and IMO dark shikari or loren would be much more qualified libav devels [01:02] <michaelni> yes [01:04] <Compn> imo writing an advanced codec like h265 will be is an incredibly hard task for a student [01:05] <iive> he worked with darkshikari on optimizing vp8 and afaik, he is emplyed by google to work on the codec. [01:05] <Compn> unless you find another kostya... [01:06] <iive> while I would also prefer ds and pengvado, I think he could handle the task. [01:06] <kierank> for various reasons ds will not be mentoring this year [01:06] <kierank> that's almost certain [01:06] <Compn> is he mentoring x264 soc ? [01:06] <Compn> is there an x264 soc ? [01:06] <kierank> doubt it [01:06] <iive> btw, is h265 finalized ? [01:06] <kierank> nearly [01:07] <LordRPI> oh boy... [01:07] <iive> i actually haven't read anything about it. Do you have some easy to read summary of the changes? [01:08] <j-b> kierank: who spoke about aacs? [01:08] <Compn> aacs is on the libav gsoc page [01:08] <j-b> why? [01:08] <Compn> also its kinda rude to just move the article instead of copying it [01:08] <kierank> it's an ffmpeg project [01:08] <j-b> http://git.videolan.org/?p=libaacs.git;a=shortlog [01:08] <kierank> j-b: exactly.. [01:08] <j-b> it works perfectly fine [01:09] <Compn> kierank / j-b : i think i meant the task to connect libaacs to ffmpeg [01:09] <Compn> in order to play/decrypt discs [01:09] <kierank> still silly imho [01:09] <kierank> just sticking random stuff in ffmpeg because you can [01:09] <Compn> whaaaaaaaaaat [01:09] <Compn> people want to encode blurays ? [01:09] <Compn> how is that random ? [01:09] <kierank> decrypt them separately [01:09] <iive> it is actually punishable by acta [01:09] <Compn> or you think they need to rip discs to hd first ? [01:09] <j-b> Compn: connect libaacs to ffmpeg? this is stupid. libbluray does that. [01:10] <kierank> ffmpeg is not a drm library [01:10] <Compn> j-b : well that aacs task was written a year or two ago ... just copied to new one [01:10] <j-b> this is totally stoopid [01:10] <Compn> feel free to delete it off gsoc page [01:11] <Compn> there isnt a mentor for it anyhow [01:11] <j-b> integration ffmpeg of libbluray is a 10hour task, at most [01:11] <j-b> and it gives you libaacs for free [01:11] <Compn> theres already a patch... [01:11] <j-b> it does not give bd+ [01:11] <Compn> i just didnt know how it all worked when i suggested it years ago [01:11] <j-b> which is _important_ [01:11] <kierank> libbluray in ffmpeg does make some sense [01:12] <Compn> ok removed from wiki [01:12] <Compn> everyone happy? :) [01:12] <Compn> i dont want to duplicate libaacs [01:13] <j-b> now, I would suggest libbdplus SoC [01:14] <iive> i thought libblueray can handle it, why else would it need java? [01:14] <kierank> for all the other blu-ray stuff in java [01:14] <iive> like bd+? [01:15] <kierank> you are not making any sense any more [01:15] <kierank> blu-ray uses lots of java components, one of those java components is bd+ [01:15] <j-b> bd+ and bd-j are different [01:15] <j-b> bd-live is part of bd-j [01:15] <j-b> bd+ does not need really need java, while bd-j does. [01:16] <Compn> j-b : its just that we are noobs when it comes to stupid bluray technology [01:16] <Compn> so we ask dumb questions [01:16] Action: Compn owns no bluray equipment or discs [01:16] <Compn> and a libbdplus would probably do best in vlc gsoc :) [01:16] <iive> If I remember what i've read bd+ have some command language that is quite similar to java bytecode, so the implementation would need its virtual machine or something like that. [01:16] <Compn> if you could convince enough of those doom9 guys to mentor anonymously... [01:17] <Compn> stupid sony/mpaa lawyers [01:17] <j-b> iive: indeed. [01:28] <iive> j-b: this is what i thought they use the jvm for. [01:29] <j-b> no, the jvm is mostly used for bd-j and animation [01:33] <iive> if there is bd+ implementation, could jre be used. Or could some of the opensource jvm be modified and used instead? [01:38] <j-b> no [01:38] <j-b> the current bd+ implementation does not use jre [01:47] <iive> so, is the vm simple enough to be done from scratch in one gsoc or is it easy task to modify existing one? [01:48] <j-b> the code is already done [01:48] <j-b> one needs to fix and update it [01:49] Action: Compn hopes one day a list of aacs keys will be committed to mplayer [01:49] <Compn> then anyone can pull it from svn ;) [01:50] <j-b> highly illegal [01:53] <Compn> behind 7 proxies... [01:53] <Compn> :P [01:53] <iive> Compn: you want all mplayer developers arrested and extradited to usa, like megaupload ? [01:53] <Compn> haha [01:53] <Compn> i'm still sad about megaupload :( [01:54] <iive> the fun is just starting. [01:54] <Compn> honestly usa is bankrupt , we dont have enough money for any wars, prison, etc [01:54] <Compn> not sure whats going on right now [01:54] <Compn> 'more of the same' [01:55] <iive> money are easy, you just print more. [01:55] <Compn> poor greece :( [01:55] <Compn> who was first? iceland ? [01:55] <iive> well, their problem is that they can't print euro. [01:55] <Compn> whos next ? [01:57] <michaelni> printing more money is just a hidden way of stealing from the people [01:57] <Compn> not really hidden either [01:57] <iive> yes. it is. [01:57] <Compn> when bread costs $5 a loaf [01:57] <Compn> i know somethings wrong [01:58] <michaelni> well id guess most people consider the police walking into their house and taking their stuff to sell it and the government printing more money as different things [01:58] <michaelni> but its not really different [01:59] <iive> Compn: is that with or without th5 roe ? [01:59] <Compn> iive : thats with all of the corn syrup our tax money can buy! [02:00] <Compn> usa subsidizes corn syrup, then corn syrup companies advertise on tv that corn syrup isnt bad for you :( [02:00] <iive> michaelni: it is actually different thing. in first case you loose the stuff you've already bought. In the second they took the stuff you were going to buy. [02:00] <j-b> Compn: seriously, getting the keys is simple [02:01] <j-b> http://vlc-bluray.whoknowsmy.name/ [02:01] <Compn> oh nice keydb [02:01] <michaelni> iive, sure the details are different [02:02] <Compn> j-b : so just keep the player key and dont need to record the disc key then? [02:02] <michaelni> but when your money is worth nothing you have to sell something if you have something [02:02] <michaelni> so the endresult is the same [02:02] <Compn> yes but there is no one to buy anything you pawn off :) [02:02] <Compn> pawn shops are full of goods [02:03] <j-b> Compn: just use the pk, yes. [02:03] <Compn> j-b : its probably illegal to even link to that site now, in usa [02:03] <Compn> :d [02:03] <j-b> maybe [02:04] <iive> Compn: is there something in usa tha tis legal? [02:05] <Compn> iive : i used to know, but they've changed laws recently [02:05] <Compn> j-b : you think the authors of that site would mind if mplayer put a news entry up about vlc playing bluray ? [02:06] <j-b> no idea. [02:06] <j-b> no idea who they are [02:06] <iive> Compn: this is a sneaky move! [02:06] <j-b> and their KEYDB.cfg certificate is already revoked [02:06] <j-b> Compn: and to be honest, it has nothing to do with VLC [02:07] <Compn> i can see that [02:08] <j-b> iive: back to the topic, I would not recommend anyone to work on BD+ in their names [02:08] <Compn> or people from usa to work on it at all. [02:08] <j-b> libbluray is fine [02:08] <Compn> or new zealand apparently ;\ [02:08] <j-b> libaacs is ok-ish, I guess. [02:08] <j-b> bd+ is not [02:08] <Compn> nz used to be a cool place [02:10] <iive> Compn: nope, they have always been quite authoritarian. But it is calm place, without much distractions and entertainment. And they like it that way. [02:10] <Compn> did they also mistreat the maori, like austrailia did with the aborigines ? [02:11] Action: Compn cant remember his history [02:11] <iive> I never knew. Ask Lobs. [02:18] <durandal_1707> michaelni: sent patch which need to be commited immediately! [02:18] <durandal_1707> 10 days passed and nobody reported issue.... [02:31] <michaelni> durandal_1707, just a seocnd, ill finish the merge and then apply it [02:34] <CIA-48> ffmpeg: 03Ronald S. Bultje 07master * rb4027d9749 10ffmpeg/libavcodec/ (sinewin.h sinewin_tablegen.c sinewin_tablegen.h wmaprodec.c): [02:34] <CIA-48> ffmpeg: wmapro: change max. block size to 13 bits. [02:34] <CIA-48> ffmpeg: WMApro actually support 13-bits block sizes (potentially even up to 14), [02:34] <CIA-48> ffmpeg: and thus we should support that also. If we get block sizes beyond what [02:34] <CIA-48> ffmpeg: the decoder can handle (14 is possible depending on s->decode_flags), [02:34] <CIA-48> ffmpeg: error out instead of crashing. [02:34] <CIA-48> ffmpeg: 03Ronald S. Bultje 07master * r52e4018be4 10ffmpeg/libavcodec/ (flacdec.c golomb.h): [02:34] <CIA-48> ffmpeg: flac: fix infinite loops on all-zero input or end-of-stream. [02:34] <CIA-48> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind [02:34] <CIA-48> ffmpeg: CC: libav-stable(a)libav.org [02:34] <CIA-48> ffmpeg: 03Ronald S. Bultje 07master * r25c42234a9 10ffmpeg/libswscale/output.c: [02:34] <CIA-48> ffmpeg: swscale: make monowhite/black RGB-independent. [02:34] <CIA-48> ffmpeg: Disadvantage is that it no longer allows modifying brightness through [02:34] <CIA-48> ffmpeg: adjustment of the RGB lookup table. Advantage is that now monowhite/black [02:34] <CIA-48> ffmpeg: no longer need to be identified as a RGB format. [02:34] <CIA-48> ffmpeg: 03Kostya Shishkov 07master * r5d9ad5425f 10ffmpeg/libavcodec/proresdsp.c: prores: initialise encoder and decoder parts only when needed [02:34] <CIA-48> ffmpeg: 03Ronald S. Bultje 07master * r1bab6f852c 10ffmpeg/libswscale/x86/swscale_template.c: [02:34] <CIA-48> ffmpeg: swscale: make access to filter data conditional on filter type. [02:34] <CIA-48> ffmpeg: Prevents crashes on 1-tap filter (unscaled). Also rename "bguf" argument [02:34] <CIA-48> ffmpeg: to "vbuf", seems that was a typo. [02:34] <CIA-48> ffmpeg: 03Paul B Mahol 07master * r6d10c5bfc0 10ffmpeg/libavformat/cdxl.c: [02:34] <CIA-48> ffmpeg: cdxl: fix audio for some samples [02:34] <CIA-48> ffmpeg: There may be extra padding at and of chunk. [02:34] <CIA-48> ffmpeg: Signed-off-by: Paul B Mahol <onemda(a)gmail.com> [02:34] <CIA-48> ffmpeg: Signed-off-by: Justin Ruggles <justin.ruggles(a)gmail.com> [02:34] <CIA-48> ffmpeg: 03Paul B Mahol 07master * rf02edc4c06 10ffmpeg/libavformat/apetag.c: [02:34] <CIA-48> ffmpeg: apetag: add proper support for binary tags [02:34] <CIA-48> ffmpeg: export as attachment streams [02:34] <CIA-48> ffmpeg: Signed-off-by: Paul B Mahol <onemda(a)gmail.com> [02:34] <CIA-48> ffmpeg: Signed-off-by: Justin Ruggles <justin.ruggles(a)gmail.com> [02:36] <CIA-48> ffmpeg: 03Paul B Mahol 07master * rbfe7bace8e 10ffmpeg/libavcodec/libx264.c: [02:36] <CIA-48> ffmpeg: libx264: unbreak libx264rgb [02:36] <CIA-48> ffmpeg: Signed-off-by: Paul B Mahol <onemda(a)gmail.com> [02:36] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [03:08] <Compn> j-b : news posted. hopefully will save people some time from ripping blurays to disc before playback [03:08] <Compn> all of the google results show the ripping method :\ [03:08] <j-b> yes [03:09] <ohsix> url? [03:11] <Compn> ohsix : http://www.mplayerhq.hu [03:11] <Compn> well [03:11] <Compn> ohsix : http://www1.mplayerhq.hu [03:12] <Compn> mirrors will update later [03:12] <ohsix> oic, thought i missed more context :D [03:13] <Compn> just me causing trouble [03:14] <Compn> trolling and coming up with bad ideas :) [04:32] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * rc36010c286 10ffmpeg/libavformat/avidec.c: [04:32] <CIA-48> ffmpeg: avidec: remove harmless duplicate code. [04:32] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [04:32] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * rc9ac08277d 10ffmpeg/libavformat/movenc.c: [04:32] <CIA-48> ffmpeg: movenc: Dont crash on aspect=0/0. [04:32] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [11:44] <CIA-48> ffmpeg: 03Stefano Sabatini 07master * r360218713f 10ffmpeg/libavfilter/vf_aspect.c: lavfi/aspect: show log info even in case no argument is provided to setdar/setsar [11:44] <CIA-48> ffmpeg: 03Stefano Sabatini 07master * r80715c3b23 10ffmpeg/libavfilter/vf_aspect.c: lavfi/aspect: remove confusing comments [11:44] <CIA-48> ffmpeg: 03Stefano Sabatini 07master * r4538d66010 10ffmpeg/libavfilter/vf_aspect.c: [11:44] <CIA-48> ffmpeg: lavfi/aspect: set default value to 0:1, as stated in the docs [11:44] <CIA-48> ffmpeg: Fix regression introduced in commit 70ffda3217c58bbbfb8a7e7c58824b8ca6c56128. [12:04] <CIA-48> ffmpeg: 03Reimar Döffinger 07master * r7c8d477299 10ffmpeg/libavformat/oggparseogm.c: (log message trimmed) [12:04] <CIA-48> ffmpeg: Make AAC in Ogg (ogm) work. [12:04] <CIA-48> ffmpeg: This needs the extradata to be extracted. [12:04] <CIA-48> ffmpeg: The approach used is the one MPlayer uses, though it is [12:04] <CIA-48> ffmpeg: unclear whether the 4 bytes extradata that are skipped [12:04] <CIA-48> ffmpeg: should be skipped always or only for AAC. [12:04] <CIA-48> ffmpeg: The AAC parser must be disabled, too, otherwise playback [12:27] <CIA-48> ffmpeg: 03Reimar Döffinger 07master * rad8173eb81 10ffmpeg/libavformat/os_support.c: [12:27] <CIA-48> ffmpeg: Allow other programs to open the same files on Windows. [12:27] <CIA-48> ffmpeg: In order to match Linux behaviour better our [12:27] <CIA-48> ffmpeg: Windows-specific open() replacement should disable [12:27] <CIA-48> ffmpeg: Windows default file locking. [12:27] <CIA-48> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> [14:37] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * rf3bd226de6 10ffmpeg/ffmpeg.c: [14:37] <CIA-48> ffmpeg: ffmpeg: Fix setting flags for codec copy. [14:37] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [15:43] <j-b> http://git.videolan.org/?p=libpostproc.git;a=summary [16:30] <Daemon404> j-b, that looks like the old repo [16:31] <Daemon404> i nuked that one and made an entirely new one [16:31] <Daemon404> as per michaelni's request [16:31] <j-b> Daemon404: this is what you gave me yesteday [16:32] <Daemon404> eh? [16:32] <Daemon404> https://github.com/dwbuiten/postproc/commits/master vs http://git.videolan.org/?p=libpostproc.git;a=summary [16:32] <j-b> https://github.com/dwbuiten/postproc or so [16:32] <Daemon404> look at it [16:32] <Daemon404> they differ. [16:32] <j-b> carp [16:32] <j-b> crap [16:32] <j-b> something went wrong [16:32] <Daemon404> :) [16:33] <Daemon404> there's a reason i dont run ym own git servers [16:33] <Daemon404> pain~ [16:35] <j-b> Daemon404: http://git.videolan.org/?p=libpostproc.git;a=summary [16:36] <j-b> Daemon404: ok? [16:36] <Daemon404> thats looks better :P [16:36] <Daemon404> i really should strip down teh rest of the unused cruft from configure [16:36] <Daemon404> but it works atm [16:38] <j-b> ok [16:42] <CIA-48> ffmpeg: 03Reimar Döffinger 07master * rf23a740303 10ffmpeg/ (6 files in 3 dirs): [16:42] <CIA-48> ffmpeg: Add muxer test based on stream-copy from FATE sample. [16:42] <CIA-48> ffmpeg: This should test muxing Theora into Ogg. [16:42] <CIA-48> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> [20:22] <GA-Flix> hi everyone, i need a bit help with the FF-MPEG license ... [20:22] <GA-Flix> am i allowed to add the ffmpeg.exe file to a project which executes the file hidden (no console window visible, but a tooltip shown that says 'Convert to MP3 using ffmpeg.exe in hidden mode. See http://www.ffmpeg.org/ for further information.') [20:22] <GA-Flix> didn't understand the text on http://www.ffmpeg.org/legal.html ... my english is not the best :( [21:02] <durandal_1707> lol got segv with ffprobe [21:03] <durandal_1707> michaelni: it appears ffprobe segv for random text files [21:13] <durandal_1707> michaelni: 29b9aee49f310fa0eb194f591417aaef0d519e7a introduced segv [21:35] <CIA-48> ffmpeg: 03Clément BSsch 07master * r8688e6acba 10ffmpeg/ffprobe.c: ffprobe: fix crash if the file can't be opened. [21:35] <ubitux> durandal_1707: fixed [21:36] <CIA-48> ffmpeg: 03Reimar Döffinger 07master * r030e96fe55 10ffmpeg/libavformat/oggdec.c: [21:36] <CIA-48> ffmpeg: oggdec: improve warning message. [21:36] <CIA-48> ffmpeg: Print whether it was a keyframe or a non-keyframe that [21:36] <CIA-48> ffmpeg: was incorrectly marked as the other. [21:36] <CIA-48> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> [21:36] <CIA-48> ffmpeg: 03Reimar Döffinger 07master * r76cc0992d8 10ffmpeg/ (libavformat/oggenc.c tests/ref/lavf-fate/ogg_vp3): (log message trimmed) [21:36] <CIA-48> ffmpeg: oggenc: fix condition when not to flush due to keyframe granule. [21:36] <CIA-48> ffmpeg: The previous condition of 0 page size was wrong, [21:36] <CIA-48> ffmpeg: that would disable the mechanism for all frames at [21:36] <CIA-48> ffmpeg: a start of a page, thus some keyframes still would not [21:36] <CIA-48> ffmpeg: get their own granule. [21:36] <CIA-48> ffmpeg: The real problem is that header packets must not be flushed, [22:12] <Compn> j-b : does vlc have rtmpe support? [22:12] <Compn> i see this bug mentions librtmp patch > http://trac.videolan.org/vlc/ticket/4342 [22:12] <Compn> figured it would have been committed :P [22:14] <CIA-48> ffmpeg: 03Reimar Döffinger 07master * r57182b9f0f 10ffmpeg/tests/ (lavf-regression.sh ref/lavf-fate/ogg_vp3): [22:14] <CIA-48> ffmpeg: lavf-fate-ogg_vp3: Try -idct auto instead of -idct simple. [22:14] <CIA-48> ffmpeg: -idct simple for some reason results in different results on [22:14] <CIA-48> ffmpeg: different systems. [22:14] <CIA-48> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> [22:25] <ubitux> Compn: if you need ideas for gsoc, i think supporting more subtitles formats would be nice, as well as porting or introducing a/v filters [22:25] <ubitux> i'm pretty sure libav ppl don't care much about this, so there is no real risk of duplicates [22:25] <kierank> what kind of subtitles [22:26] <ubitux> we need a microdvd decoder at least, and i think any subtitles format supported by mplayer has its place in ffmpeg [22:26] <cbsrobot> I suggest creating DCP would be a nice task aswell [22:27] <ubitux> i'm working on it anyway, but well, time's lacking [22:28] <kierank> cbsrobot: you need a decent jpeg2k encoder for that [22:29] <cbsrobot> it could be done with libopenjpeg [22:30] <cbsrobot> dcp needs 12-bit, xyz colospace j2k files in a mxf container, audio in mxf and subtitle in mxf and create the bundle [22:31] <cbsrobot> I think it's doable as a gsoc task [22:31] <kierank> well is it the right place for ffmpeg to be creating bundles of stuff [22:31] <kierank> since there are other projects that do that already [22:31] <ubitux> what about mkv ordered chapters and stuff? :) [22:31] <kierank> imho ffmpeg/libav should deal with the codecs and mux and that's it [22:32] <cbsrobot> you got apoint, but I see the bunlde a a sort of container - nothing more [22:34] <cbsrobot> what about adding libraw to ffmpeg ? [22:34] <Compn> ubitux : sounds great, do you have a repo i could link to ? [22:35] <durandal_1707> cbsrobot: whould that need sws changes anyway [22:35] <Compn> cbsrobot : dcp ? [22:35] <Compn> dcp muxer ? [22:35] <Compn> cbsrobot : can you put it on wiki page? i have no idea links and info etc... [22:35] <ubitux> Compn: a repo of what? [22:35] <kierank> cbsrobot: at fosdem this was discussed [22:35] <Compn> ubitux : your sub work [22:35] <cbsrobot> I guess I need an account ... [22:36] <kierank> cbsrobot: some kind of library to deal with compressed formats [22:36] <Compn> cbsrobot : mail mike(a)multimedia.cx , or alternatively, if you want me to post it, just paste me some specs and description of project [22:36] <cbsrobot> I just propose to see what other ppl think before adding it to the wiki [22:36] <kierank> uncompressed i mean [22:36] <ubitux> Compn: nothing public ATM, and nothing mature enough [22:36] <Compn> ah [22:36] <ubitux> sorry :) [22:36] <Compn> no worries :) [22:36] <Compn> ubitux : ok, do you want to mentor subtitle project? ;p [22:37] <ubitux> i'm not comfortable enough with the code atm [22:37] <Compn> mentors get ... 500 google dollars :P [22:37] <cbsrobot> libraw is a wrapper for dcraw - to read all the fancy raw files [22:37] <ubitux> i hope i'll be able to help next year, but not this time [22:37] <ubitux> Compn: i don't care about the money :) [22:37] <Compn> lol @ mkv ordered chapters [22:37] <Compn> we are still waiting on playlist support i think [22:38] <Compn> maybe a task to finish playlist support and just get it in there [22:38] <Compn> playlist is blocking few things now :\ [22:38] <cbsrobot> durandal_1707: I guess so [22:38] <ubitux> and actually, i don't like the idea of being paid for this; it makes me feel too responsible :p [22:38] <kierank> Compn: no they don't [22:38] <kierank> the org gets the momey [22:38] <cbsrobot> google just profits [22:39] <Compn> kierank : i have no idea how ti works. [22:39] <Compn> heh [22:39] <Compn> does libass support srt , sub, etc ? [22:39] <Compn> youd think we just pass it all thru libass... [22:40] <Compn> or did libass say he didnt want to do that ? [22:40] <cbsrobot> also >8 bit support for filters would be a nice task [22:40] <cbsrobot> but maybe too easy [22:41] <Compn> could be good qual task [22:41] <Compn> if too easy [22:41] <ubitux> Compn: i'm working on this [22:41] <Compn> ahhh i guess libass cant do srt [22:41] <Compn> ubitux : http://mailman.videolan.org/pipermail/vlc-devel/2011-September/081803.html [22:41] <cbsrobot> I'm not sure - maybe ubitux or saste can tell [22:41] <JEEB> yeah, libass only supports ass IIRC [22:41] <Compn> thats a pain [22:41] <ubitux> it shouldn't support anything else imo [22:41] <JEEB> ubitux, aye [22:41] <saste> Compn: confirm [22:41] <JEEB> conversion from srt to ass isn't exactly hard tho for rendering purposes [22:42] <saste> that's what libass is supposed to do ;) [22:42] <Compn> what [22:42] <Compn> hmm [22:42] <ubitux> Compn: check the subtitles talk from a month or two ago; we discuss this a bit [22:42] <Compn> nah i'll pass :) [22:42] <Compn> i like mplayer can pass srt to libass :) [22:42] <ubitux> then basically, we need to add subtitles support in lavfi [22:42] <ubitux> right now i'm "learning" lavfi with audio, because i have some related stuff [22:43] <ubitux> then i'll work again on the subtitles. [22:43] <Compn> well you guys mostly have wiki accounts, you can add your own ideas ;) [22:43] <Compn> since i'm not up on whos boat is the right color for what feature [22:43] <cbsrobot> saste: libavf >8 bit support - could that be a task ? [22:44] <durandal_1707> I propose following task: RE and write decoder for TAK [22:45] <cbsrobot> kierank: what else was discusset at fosdem ? [22:45] <cbsrobot> and kierank did you ever get hands on the aja oem sdi cards for obe ? [22:45] <saste> cbsrobot: libavf >8 bit support: what do you mean? [22:46] <cbsrobot> I saw lately they have linux drivers aswell [22:46] <cbsrobot> saste: most filters work in 8 bit [22:46] <kierank> cbsrobot: no [22:46] <cbsrobot> would be nice to have em in 16 bit aswell [22:46] <kierank> the aja cards don't provide anything i don't have already i think [22:46] <saste> uhm i want to propose a generic lavfi task [22:46] <Compn> http://ale5000.altervista.org/subtitles.htm [22:46] <Compn> neato [22:47] <saste> something like working on porting filters, e.g. from mplayer [22:47] <kierank> cbsrobot: a rough idea about a library dealing with uncompressed pixel formats (e.g v210, r210 etc) [22:47] <saste> also there are many image formats which we still don't support [22:47] <kierank> because there are so many [22:47] <Compn> saste : still dont like -vf mp huh? :P [22:47] <durandal_1707> kierank: huh? [22:47] <saste> iirc we still don't support compressed GIFs [22:48] <saste> -vf mp is cursed, that's why i want to get rid of it ;-) [22:48] <Compn> getting rid of it wont get projects back together [22:48] <cbsrobot> kierank: I was just wondering, because you only mention the balckmagic cards on the wiki - and they seem to have some flaws [22:48] <kierank> the dveo ones are fine [22:49] <kierank> in fact they are the only ones i trust [22:49] <saste> no but that would make a nice exercise for a student anyway [22:49] <Compn> kierank : you mean create a seperate project for raw codecs ? [22:49] <kierank> Compn: well the point was that this should be merged with swscale [22:49] <kierank> because v210 etc are actually pixel formats [22:49] <kierank> not codecs [22:49] <Compn> wow that bikeshed still going ? [22:50] <durandal_1707> kierank: feel free to improve sws so that can happen [22:50] <kierank> well it means you can use the same simd for the difference variants [22:50] <Compn> the reason it wasnt put in swscale before was ... well i cant remember [22:51] <durandal_1707> Compn: there was flamewar(s) about this [22:51] <cbsrobot> Compn: you have a bikeshed on your boat ? [22:51] <kierank> the only argument about them being codecs i can tell is that they go into containers [22:51] <Compn> durandal_1707 : i remember them [22:51] <Compn> kierank : i agree with putting raw into swscale, good luck! :) [22:52] <Compn> i'd like bayer rgb in there as well, as long as we are talking about that [22:52] <Compn> forgot what bayer format i wanted now, but something... [22:52] <Compn> maybe webcam colorspace [22:52] <Compn> old soc page probably [22:52] <kierank> http://web.archiveorange.com/archive/v/yR2T4OXznFIkMCek3x24 [22:53] <Compn> cbsrobot : yes, but if the boat isnt floating, can it still be considered a boat ? [22:54] <Compn> kierank : michaelni speficially says if theres a patch he will reconsider putting v210 in swscale [22:54] <cbsrobot> Compn: as long as it's integer it can be considered a boat - I'm quite sure. [22:55] <Compn> kierank : mostly i think no one wanted to maintain swscale at the time [22:55] <Compn> so ... easier to drop it in a codec and run! [22:55] <Compn> run, before the attack of the killer bikeshed! [22:55] <Compn> (good plot for nerd movie) [22:57] <kierank> well the reason people read v210 is to convert to another colourspace [22:58] <Compn> great, then they should be able to do that in the fastest possible wayt [22:58] <durandal_1707> quote from michaelni: "its not a pixel format. Its a very poorly compressing codec." ;-) [22:59] <ohsix> flip a coin [22:59] <kierank> yes and the fastest possible way is not to unpack it then repack it [22:59] <ohsix> or defer to one person [23:00] <Compn> then what michaelni said in 2010 still stands [23:00] <Compn> send patch [23:00] <kierank> well yes but we'll do it for all the packed pixel formats [23:00] <Compn> sounds great :) [23:01] <durandal_1707> next day they will come with idea to move everything to swscale... [23:01] <Compn> mplayer can utilize ffmpeg's raw decoders so its easy to add to mplayer... [23:02] <kierank> durandal_1707: seriously, I don't see what the problem is [23:02] <Compn> durandal_1707 : it would speed things up probably [23:02] <Compn> kierank is just looking for someone to argue with [23:02] <Compn> ;) [23:02] <kierank> no i'm not [23:02] <durandal_1707> kierank: good luck with it, swscale is nightmare [23:03] <Compn> werent we just talking about swscale the other day ? [23:03] <Compn> forgot about what [23:03] <Compn> maybe it was a diff file [23:11] <Compn> anyways cbsrobot , kierank , ubitux , saste : thanks, those were all some good ideas for soc :) [23:12] <Compn> and durandal_1707 [23:12] <Compn> i forgot what tak was again [23:13] <ubitux> :) [23:14] <ubitux> 22:44:30 < durandal_1707> I propose following task: RE and write decoder for TAK [23:15] <durandal_1707> or just RE it and i will write decoder [23:17] <Compn> are there samples in the sample repo ? [23:47] Action: Compn added some ideas to page [23:48] <ubitux> maybe we could have a "coverage" gsoc too [23:48] <ubitux> (trying to improve fate coverage on a maximum of stuff) [23:49] <Compn> those are good qualification tasks [23:49] <Compn> adding more fate tests , i mean [23:49] <ubitux> btw, i guess we don't have any real time updates? [23:49] <Compn> ? [23:49] <Compn> whats a real time update ? [23:49] <ubitux> erh, brainfart, i mean real time public coverage html page updates [23:50] <ubitux> just like fate [23:50] <Compn> oh [23:50] <Compn> mike might be working on that [23:50] <Compn> code coverage [23:50] <Compn> or have ideas for it [23:50] <ubitux> i could add the --enable-coverage to one of my instances [23:50] <ubitux> and make a page publicaly available [23:50] <ubitux> shouldn't be much work [23:51] <Compn> you go girl! [23:51] Action: Compn goes afk [23:51] <ubitux> i'm not sure how well it will integrates to fate [23:51] <ubitux> especially since the "build" is dropped at the end [23:52] <ubitux> Compn: is mike on irc? [23:52] <kierank> no [23:52] <ubitux> mike multimedia mike? [00:00] --- Sun Feb 19 2012

1 0

ffmpeg.log.20120218
by burek 19 Feb '12

19 Feb '12

[00:34] <pasteeater> why do you want shared and pic? [01:08] <Austen> should I not? [01:28] <pasteeater> it can make compilation trickier, and you probably don't it, but i'm not sure what you're doing exactly [01:30] <Austen> so, i attempted to build almost all the libraries with disable-shared, but have ffmpeg build with --enable-shared [01:31] <Austen> with fPIC in my CFLAGS, and that caused the shared libraries to fail during make [01:31] <Austen> i was looking at it more, it appears theres not that much of a benifit to having enable-shared; a bigger binary but the libraries have to be there in either case.... [01:32] <Diogo> hi, one question please, first this is possible use ffmpeg for real time encoding, i need to build a file upload in php (send file via POST http) and the server when is getting the video start to encode..any tutorial on the web? like transloadit.. [01:34] <Diogo> i can see a solution http://transloadit.com/demos/video-encode/encode-a-video-in-realtime [01:35] <Diogo> but i don't know how they do this.. [01:35] <Diogo> thanks for your help [03:50] <const_antine> hello [03:50] <const_antine> here's a confusing quote from faq [03:50] <const_antine> "Both Xvid and DivX (version 4+) are implementations of the ISO MPEG-4 standard (note that there are many other coding formats that use this same standard). Thus, use -c:v mpeg4 to encode in these formats." [03:51] <const_antine> isn't mpeg-4 format for media container files? [03:51] <const_antine> and not media itself? [03:56] <const_antine> nvm, wrong [04:18] <entreri> hello, it is possible to take a picture from the command-line using ffmpeg ? [04:20] <Tjoppen> what kind of picture? [04:20] <entreri> with a webcam [04:20] <Tjoppen> -f v4l2 -s 640x480 -i /dev/video0 [04:20] <Tjoppen> or so [04:21] <Tjoppen> and -t 0.01 to limt the number of frames [04:21] <entreri> ok thanks, I forgot the -s [04:21] <Tjoppen> so a single picture should be something like ffmpeg -t 0.01 -f v4l2 -s 640x480 -i /dev/video0 -qscale 2 out.jpeg [04:22] <entreri> and by curiosity, can I take a picture of the terminal (TTY) and/or of xorg ? [04:23] <Tjoppen> -f x11grab -s `xwininfo -root|grep geometry|awk '{print $2}'` -r 30 -i :0.0 [04:23] <Tjoppen> is what I use for desktop capture [04:23] <Tjoppen> -f alsa -i default for mic [04:24] <Tjoppen> I have a script that does all of those. sadly my webcam seems to be broken [04:24] <entreri> wow, I rarely deal with complicated commands like this, that's sick:P [04:24] <entreri> thanks a lot [04:25] <Tjoppen> yeah, it's because the x11grab input doesn't default to the current resolution for some reason [04:25] <Tjoppen> but it works wonderfully, even with multiple screens [04:29] <Tjoppen> ah, one downside: my screens have different resolutions. it doesn't fill in the gaps with black [04:38] <burek> entreri, you can also use -vframes 1 [04:38] <burek> instead of -t [04:46] <entreri> I tried both, but I get errors ; [video4linux2,v4l2 @ 0x9660b60] Cannot find a proper format for codec_id 0, pix_fmt -1 [04:46] <entreri> /dev/video0: input/output error [04:47] <entreri> can't find anything on the web [04:50] <entreri> that's what I did ; sudo -u user ffmpeg -f video4linux2 -s 640x480 -i /dev/video0 -t 0.01 test.jpeg [04:50] <entreri> it does the same as root [05:03] <burek> if you just installed your ffmpeg [05:03] <burek> through apt-get or similar [05:03] <burek> its possible that you don't have a proper codec built in ffmpeg [05:03] <burek> can you please use pastebin.com, to show your command line and its output? [05:07] <entreri> burek: that's right, I just installed it through pacman [05:08] <entreri> should I look for a codec or something ? [05:08] <burek> i dunno [05:08] <burek> i dont see the output [05:12] <entreri> pastebin.com/nLh52ehU [05:13] <burek> the command is missing [05:15] <entreri> the command that generated that is ; ffmpeg -f video4linux2 -s 640x680 -i /dev/video0 out.mpeg [05:15] <entreri> yeah sorry I forgot [05:15] <entreri> you want a new pastebin with the command instead ? [05:16] <burek> try something more common like: ffmpeg -f video4linux2 -s 640x480 -i /dev/video0 -vcodec libx264 -crf 20 out.mp4 [05:18] <entreri> same... [05:18] <entreri> the webcam works with flash also [05:20] <entreri> maybe it needs to be done inside an X server ? [05:20] <entreri> I do the command in TTY1 [05:20] <entreri> ctrl-alt F1 [05:21] <burek> what exactly is your goal [05:22] <entreri> well, taking pictures and video using ffmpeg:P [05:22] <entreri> but it doesn't work and I don't know why [05:23] <burek> it seems your camera sends a video stream which ffmpeg doesn't understand [05:23] <burek> your best bet is to try to compile ffmpeg [05:23] <burek> or grab a static build [05:23] <burek> and see if that works [05:23] <burek> it will be faster [05:24] <burek> https://sites.google.com/site/linuxencoding/builds [05:24] <burek> just download ffmpeg, unpack and run [05:25] <burek> but run it with ./ffmpeg ... [05:26] <entreri> damn [05:26] <entreri> I just found something [05:26] <entreri> my driver is linux-uvc [05:27] <entreri> and it doesn't support v4l2 [05:27] <entreri> you were right about the stream that ffmpeg did not understand:P [05:28] <entreri> well, I can at least record sound [17:45] <francogrex> Hi I am using this command to make a conversion: ffmpeg -i movie.avi -target pal-dvd -aspect 16:9 -sameq movie.mpg [17:45] <francogrex> But I'd like to reduce the size of the mpg a little bit, which parameters can help me do that without compromosing image and sound quality a lot? [17:46] <Mavrik> first of all [17:47] <Mavrik> -sameq doesn't do what you think it does, lose it :) [17:47] <francogrex> so it's useless here? [17:47] <Mavrik> yeah, it basically just ruins your qualiy [17:48] <Mavrik> otherwise, use bitrate parameter to "set" quality [17:48] <francogrex> jeez, good to know [17:48] <Mavrik> -b 2000k [17:48] <Mavrik> or somesuch [17:48] <Mavrik> depending on how big you want your file to be [17:48] <JEEB> sameq is the most misused parameter in ffmpeg I think [17:48] <francogrex> what is the default bitrate? [17:48] <JEEB> 200k or so? [17:48] <JEEB> some quite low figure [17:49] <Mavrik> francogrex, very low, bout 200 or 300k [17:49] <Mavrik> JEEB, yeah, it has a terrible terrible name :\ [17:49] <francogrex> so maybe I should set it to 100k ? [17:49] <JEEB> ... [17:49] <JEEB> > DVD [17:49] <JEEB> > 100kbps [17:49] <JEEB> DVDs are generally 8000k or so [17:50] <sacarasc> Up to 9500 or something. [17:50] <JEEB> something like that with vbv [17:50] <sacarasc> 9800? [17:50] <Mavrik> yeah, standard DVDs are like 4000-900k [17:50] <Mavrik> *9000k [17:50] <francogrex> man, I am using the 200k defaukt, no wonder the image is crap [17:50] <francogrex> but still for this specific one, I need to reduce it [17:51] <francogrex> it's quite a large movie [17:51] <JEEB> you most probably don't want MPEG-2 then [17:51] <JEEB> if you want to compress [17:51] <francogrex> i'm used to mpg, I use dvdauthor afterwards to make the video_ts [17:51] <JEEB> MPEG-2 is pretty much needed for DVDs, and you can't really lower the bitrate much with MPEG-2 unless you like blockfests [17:52] <JEEB> unless your source is a non-moving picture or something [17:52] <francogrex> blockfests ? [17:52] <sacarasc> Lots of blockiness. [17:53] <francogrex> I guessed. I'll give it a shot anyway [17:54] <francogrex> would changing the aspect 16:9 help instead? [17:56] <francogrex> WARNING: The bitrate parameter is set too low. It takes bits/s as argument, not kbits/s [17:56] <francogrex> I typed -b 100k [17:57] <JEEB> basically I guess it's lower than the encoder would like to :P [17:58] <sacarasc> That;s only 12.5kilobytes per second. [17:58] <sacarasc> That's nothing. [17:59] <francogrex> ok I see the default is about bitrate=3233.7kbits/s so 100 is indeed too low [17:59] <francogrex> I would set it ad 2000 [18:00] <francogrex> hmm, now the average is 850... although i asked for 2000 [18:02] <francogrex> sorry was a type I forgot the k [18:02] <francogrex> typing error, now it's aroound 2000, i'll see the final quality [20:17] <GA-Flix> hi everyone [20:18] Action: sacarasc flicks GA-Flix. [20:19] <GA-Flix> am i allowed to add the ffmpeg.exe file to a project which executes the file hidden (no console window visible, but a tooltip shown that says 'Convert to MP3 using ffmpeg.exe in hidden mode. See http://www.ffmpeg.org/ for further information.') [20:20] <GA-Flix> didn't understand the text on http://www.ffmpeg.org/legal.html ... my english is not the best :( [20:22] <Mavrik> GA-Flix, yes you are allowed to do that, just make sure your ffmpeg build is redistributable if you want to include it to your project [20:23] <GA-Flix> i just added the .exe file to the project folder (everyone who wants it can copy it ^^), do you mean that with 'redistributable'? [20:23] <sacarasc> Can't contain things like faad or faac. [20:25] <GA-Flix> 'faad' and 'faac' are unknown things for me :( something with encoder and decoder i suppose (wikipedia) [20:26] <Mavrik> GA-Flix, some libraries that ffmpeg can include have different licenses... when you compile FFMpeg is shows clearly "Licenes: NON-REDISTRIBUTABLE" if that's the case [20:26] <Mavrik> GA-Flix, which ffmpeg.exe build are you using? [20:27] <GA-Flix> static x86 (x32) [20:27] <GA-Flix> http://ffmpeg.zeranoe.com/builds/win32/static/ [20:27] <Freakshow> pasteeater: just fyi: [20:27] <Freakshow> web2.lab2# ffmpeg [20:27] <Freakshow> ffmpeg version 0.9.1.git-b4ec645 Copyright (c) 2000-2012 the FFmpeg developers [20:27] <Freakshow> ffmpeg-devel on freebsd [20:28] <Mavrik> GA-Flix, Zeranoe's builds are redistributable so go ahead :) [20:28] <GA-Flix> ok, thank you very much :) [20:29] <GA-Flix> where can i find such information in the future? [20:30] <Mavrik> you have to check the licences of the programs/libraries you use [20:32] <GA-Flix> each file? the "licenses" folder contains 21 licenses ô.0 [20:33] <Mavrik> file? [20:34] <GA-Flix> "licenses" folder contains 21 files (bzip2.txt, ffmpeg.txt, freetype.txt, ...) [20:35] <GA-Flix> and each file seems to be a license (i haven't had much to do with licenses yet) [20:35] <pasteeater> Freakshow: thats newer than i would have guessed [20:50] <pasteeater> anyone know how to use the select filter to grab an image at specfic percentages of duration? (50%, etc) [22:41] <Freakshow> pasteeater: at least it matches what I'm using on my customer facing product [00:00] --- Sun Feb 19 2012

1 0

ffmpeg-devel.log.20120217
by burek 18 Feb '12

18 Feb '12

[00:29] <CIA-48> ffmpeg: 03Stefano Sabatini 07master * r7bdefc0f12 10ffmpeg/libavfilter/ (version.h vf_overlay.c): [00:29] <CIA-48> ffmpeg: lavfi/overlay: add logic for avoiding overlaying frames with PTS > main frame PTS [00:29] <CIA-48> ffmpeg: Also add debug logging messages for helping tracking down similar [00:29] <CIA-48> ffmpeg: issues. [00:29] <CIA-48> ffmpeg: Fix trac ticket #467. [00:30] <ubitux> saste: hi :) [00:31] <saste> hi [00:34] <pyoor> Hi all. I'm new to the mpeg format and was hoping someone here might be able to answer some questions for me. Viewing an m4a in atomboxstudio I see a number of fields in the mvhd atom similar to what's outlined here (https://developer.apple.com/library/mac/#documentation/QuickTime/QTFF/QTFFC…) Why are only half of these listed in the ISO (jaadec.sourceforge.net/specs/ISO_14496-12_ISOBMFF.pdf)? [00:45] <CIA-48> ffmpeg: 03Justin Ruggles 07master * rf320fb894c 10ffmpeg/ (4 files in 4 dirs): [00:45] <CIA-48> ffmpeg: bethsoftvid: pass palette in side data instead of in a separate packet. [00:45] <CIA-48> ffmpeg: Update FATE reference to account for now non-existent palette packet. [00:45] <CIA-48> ffmpeg: This also fixes the FATE test if frame data is not initialized in [00:45] <CIA-48> ffmpeg: get_buffer(), so update comment in avconv accordingly. [00:45] <CIA-48> ffmpeg: 03Martin Storsjö 07master * rf3a094f2da 10ffmpeg/libavformat/rtsp.c: [00:45] <CIA-48> ffmpeg: sdp: Ignore RTCP packets when autodetecting RTP streams [00:45] <CIA-48> ffmpeg: The rtp demuxer which listens for RTP packets and detects the [00:45] <CIA-48> ffmpeg: RTP payload type will currently get confused if the first packet [00:45] <CIA-48> ffmpeg: received is an RTCP packet. Thus ignore such packets. [00:45] <CIA-48> ffmpeg: Signed-off-by: Martin Storsjö <martin(a)martin.st> [00:45] <CIA-48> ffmpeg: 03Christophe GISQUET 07master * rf3e084909b 10ffmpeg/libavcodec/x86/mpegaudiodec_mmx.c: (log message trimmed) [00:45] <CIA-48> ffmpeg: mpegaudio: replace memcpy by SIMD code [00:45] <CIA-48> ffmpeg: By replacing memcpy with an unrolled loop using the alignment knowledge [00:45] <CIA-48> ffmpeg: it has, some speedup can be obtained. [00:45] <CIA-48> ffmpeg: Before (gcc 4.6.1): ~400 cycles [00:45] <CIA-48> ffmpeg: After: ~370 cycles [00:46] <CIA-48> ffmpeg: Overall, around 2% speed increase when decoding a 2400s mp3 to f32le. [00:46] <CIA-48> ffmpeg: 03Ronald S. Bultje 07master * rae591aeea5 10ffmpeg/libavcodec/vc1dec.c: [00:46] <CIA-48> ffmpeg: 03Justin Ruggles 07master * rc9fdf3241a 10ffmpeg/ (libavformat/bethsoftvid.c tests/ref/fate/bethsoft-vid): [00:46] <CIA-48> ffmpeg: bethsoftvid: synchronize video timestamps with audio sample rate [00:46] <CIA-48> ffmpeg: According to unofficial documentation, the video rate is locked to the audio [00:46] <CIA-48> ffmpeg: sample rate. This results in proper synchronization of audio and video [00:46] <CIA-48> ffmpeg: timestamps from the demuxer. This only works if the first audio packet occurs [00:46] <CIA-48> ffmpeg: before the first video packet or the audio sample rate is the default rate of [00:46] <CIA-48> ffmpeg: 11111 Hz, both of which are true for all samples in our archive. [00:46] <CIA-48> ffmpeg: 03Justin Ruggles 07master * r773ff823da 10ffmpeg/libavformat/bethsoftvid.c: [00:46] <CIA-48> ffmpeg: bethsoftvid: add audio stream only after getting the first audio packet [00:46] <CIA-48> ffmpeg: This avoids initializing a stream with dummy values or when the file does not [00:46] <CIA-48> ffmpeg: contain audio. [00:46] <CIA-48> ffmpeg: Also set duration for audio packets, using the sample rate as the time base. [00:46] <CIA-48> ffmpeg: 03Ronald S. Bultje 07master * rc6643fddba 10ffmpeg/libavcodec/golomb.h: [00:46] <CIA-48> ffmpeg: golomb: avoid infinite loop on all-zero input (or end of buffer). [00:47] <saste> ubitux: do you have comments on the -count* options for ffprobe? [00:48] <ubitux> about the latest merged patch by mateo`? [00:48] <saste> yes [00:48] <Compn> pyoor : good question, did you also see the mp4 atom list on http://wiki.multimedia.cx ? [00:49] <pyoor> Compn: no I haven't [00:49] <Compn> pyoor : not sure why there is a difference, is it possible that apple only uses part of the iso spec? [00:49] <Compn> http://wiki.multimedia.cx/index.php?title=MP4 [00:49] <pyoor> actually you know I have seen that [00:50] <pyoor> but it doesn't list the structure inside the atoms [00:50] <ubitux> saste: won't that silently infinite loop with network streams for instance? [00:50] <ubitux> i guess only in case of specifying the option so not really important [00:50] <saste> yes it will... [00:51] <saste> but that's also the expected behavior [00:51] <ubitux> anyway, i guess it is ok if tested [00:51] <Compn> pyoor : what are you looking for that info for? going to create your own mp4 muxer ? [00:52] <pyoor> Compn: Actually writing a file fuzzer for mp4 and quicktime based files [00:52] <Compn> ah [00:52] <pyoor> as for the apple question, according to wiki "MPEG-4 Part 14 is essentially identical to the QuickTime file format, but formally specifies support for Initial Object Descriptors (IOD) and other MPEG features." [00:52] <pyoor> heh, but that's all I know [00:52] <ubitux> saste: i'll push http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/2012-February/120999.html soon, no objection? [00:52] <pyoor> today's really the first day I've dug into the container so finding information so far has been difficult [00:53] <Compn> i bet running zzuf on some files will cause some crashes in our mp4 demuxer :D [00:53] <pyoor> heh, haven't seen that in a long time [00:53] <pyoor> i develop nearly everything in peach [00:53] <Compn> ah [00:53] <Compn> zzuf seems kinda popular around here [00:54] <pyoor> probably because it requires little intervention [00:54] <pyoor> which is great in prep time [00:54] <pyoor> peach requires you to practically define the entire file format [00:54] <Compn> ah [00:54] <Compn> are you a 'security researcher' or doing it just for funs ? [00:55] <pyoor> i.e. http://www.flinkd.org/projects/peach-pits/peach-pit-for-the-zip-file-format… [00:55] <pyoor> researcher [00:55] <Compn> oh no xml! [00:55] Action: Compn shields his eyes [00:55] <pyoor> heh [00:56] <saste> ubitux: ok if correct, that's a consequence of the av_get_channel_layout() changes I suppose [00:56] <Compn> it might be useful to read the mp4 demuxer in ffmpeg, mplayer and vlc (they all have different mp4 demuxers) [00:56] <Compn> plus mp4box and other mp4 tools [00:56] <Compn> read the source, i mean [00:56] <pyoor> right [00:56] <ubitux> saste: yes, i had some surprises with -af; it was detecting stereo as 2.1 and similar :p [00:56] <pyoor> that was the next step. i thought i got lucky with the iso [00:57] <ubitux> saste: btw, did you receive my mails? [00:57] <Compn> pyoor : you might also want to ask in #libav-devel [00:57] <saste> ubitux: which one? [00:57] <pyoor> Compn: thanks I'll try that [00:57] <ubitux> saste: "libavfilter audio AVFrame" [00:57] <saste> ah yes [00:58] <ubitux> ok :) [01:09] <CIA-48> ffmpeg: 03Matthieu Bouron 07master * r29b9aee49f 10ffmpeg/ (Changelog doc/ffprobe.texi doc/ffprobe.xsd ffprobe.c): [01:09] <CIA-48> ffmpeg: ffprobe: add count_frames and count_packets options [01:09] <CIA-48> ffmpeg: Signed-off-by: Stefano Sabatini <stefasab(a)gmail.com> [01:20] <kierank> you know there's nothing at http://www.ffmpeg.org/schema/ffprobe, right? [01:34] <saste> kierank: yes, iirc the xml schema spec doesn't force the schema url to be "linkable", but I may be wrong [01:34] <saste> anyway we may publish it on ffmpeg.org [01:38] <ubitux> saste: what kind of rework is needed for AVFilterBufferRef? [01:40] <ubitux> :( [06:13] <CIA-48> ffmpeg: 03Zongyao Qu 07master * r1b84cf41e0 10ffmpeg/libavcodec/Makefile: [06:13] <CIA-48> ffmpeg: Makefile: fix proresdec dependancies [06:13] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [08:39] <zamb> Hi everybody. I tried compiling ffmpeg from git about an hour ago, and got this error: [08:39] <zamb> libavcodec/libavcodec.so: undefined reference to `ff_jpeg_fdct_islow_10' [08:39] <zamb> I fixed it by editing "libavcodec/Makefile" and adding "jfdctint.o" [08:39] <zamb> to the "OBJS-$(CONFIG_PRORES_LGPL_DECODER)" line. That is all. Thank you. [12:41] <ubitux> mmh& AVFrame has no channel layout& just channels :( [12:45] <ubitux> wait no channels either? beh. [13:51] <Daemon404> isnt that uses set in codeccontext? [13:52] <Daemon404> usually* [13:59] <ubitux> lavfi returns bufferref based on the avframe props [13:59] <ubitux> so if we want lavfi to do the same as for video, we need all the props in the avframe [13:59] <ubitux> or maybe i'm missing sth [14:01] <ubitux> (the same for audio i meant) [15:57] Action: Daemon404 pokes michaelni [16:30] <michaelni> Daemon404, yes, how can i help ? [16:31] <Daemon404> michaelni, i was going to ask you something about git cherry-pick with regards to keeping both dates, but i decided to base standalone postproc on ffmpeg [16:31] <Daemon404> git filter-brnach is urnning now. takes 1-2 hrs. [16:33] <j-b> hello [16:34] <Daemon404> yes i decided to get off my ass and do it after flying to UK and job hunting for a while. [16:34] <av500> j-b: world! [16:45] <Daemon404> $ echo "scale=4; `grep Rewrite p.log | wc -l` / 37952 * 100" | bc | sed -e 's/$.\{5\}$.*/\1%/' [16:45] <Daemon404> 27.49% [16:45] <Daemon404> god i hate filter-branch. [16:49] <michaelni> Daemon404, you could just git clone, git rm all but libpostproc and be done with it [16:49] <Daemon404> thats a lot of extra history and cruft left in the repo [16:50] <Daemon404> id rather spend some cpu time now [16:50] <Daemon404> btw michaelni, what's your take on the whole +100 postproc api version [16:51] <Daemon404> should standalone postproc keep this? [16:53] <michaelni> i guess it can be droped with the next minor bump [16:53] <michaelni> but it could as well be kept [16:53] Action: Daemon404 is impartial [16:53] <michaelni> doesnt really do any harm [17:01] <CIA-48> ffmpeg: 03Kostya Shishkov 07master * r857d14863d 10ffmpeg/libavcodec/proresdsp.c: prores: initialise encoder and decoder parts only when needed [17:01] <CIA-48> ffmpeg: 03Andrew Wason 07master * ra0e44414df 10ffmpeg/libavformat/movenc.c: (log message trimmed) [17:01] <CIA-48> ffmpeg: movenc: leave st->codec->frame_size as is instead of forcing it to 1. [17:01] <CIA-48> ffmpeg: Muxing pcm audio in MOV using avcodec_encode_audio() was failing [17:01] <CIA-48> ffmpeg: because avcodec_encode_audio() returns an incorrect packet size of 4 [17:01] <CIA-48> ffmpeg: bytes. This can be reproduced by modifying the sample [17:01] <CIA-48> ffmpeg: ffmpeg/doc/examples/muxing.c to encode PCM, see ML patch [17:01] <CIA-48> ffmpeg: muxing-test.diff [17:31] <CIA-48> ffmpeg: 03Reimar Döffinger 07master * r3ce7148dff 10ffmpeg/libavformat/movenc.c: [17:31] <CIA-48> ffmpeg: movenc: fix crashes if a stream is empty. [17:31] <CIA-48> ffmpeg: For some reason this always happens with -f ismv. [17:31] <CIA-48> ffmpeg: See trac issue #996. [17:31] <CIA-48> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> [17:31] <CIA-48> ffmpeg: 03Reimar Döffinger 07master * r9e0f897fe5 10ffmpeg/ (configure tests/lavf-regression.sh tests/ref/lavf/ismv): [17:31] <CIA-48> ffmpeg: Add ismv regression test. [17:31] <CIA-48> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> [17:31] <CIA-48> ffmpeg: 03Reimar Döffinger 07master * rd493170ec9 10ffmpeg/libavformat/matroskadec.c: [17:31] <CIA-48> ffmpeg: matroskadec: introduce resync function. [17:31] <CIA-48> ffmpeg: This allows handling matroska files with errors. [17:31] <CIA-48> ffmpeg: Fixes test4.mkv and test7.mkv from the official Matroska test suite. [17:31] <CIA-48> ffmpeg: These are also trac issues #544 and #545. [17:31] <CIA-48> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> [18:09] <lionaneesh> Hey [18:09] <lionaneesh> michaelni, There? [18:10] <durandal_1707> michaelni is always busy [18:12] <Compn> hes here often tho [18:13] <Compn> just wait around lionaneesh [18:13] <lionaneesh> Compn, Yeah i am. [18:18] <michaelni> lionaneesh, how can i help ? [18:58] <Daemon404> michaelni, FINALLY i got a nice-looking repo [18:58] <Daemon404> after some intense git-fu [19:00] <Daemon404> all the commit dates are in-tact and [19:00] <Daemon404> $ diff -urN ffmpeg/libpostproc/ postproc/libpostproc/ | wc -l [19:00] <Daemon404> 0 [19:00] <j-b> nice [19:01] <Daemon404> and git log --pretty=one | wc -l -> 185 [19:01] <Daemon404> nice and sane... [19:02] <Daemon404> http://pastebin.com/raw.php?i=gr9SUP3M <-- nothing sane here tho. [19:53] <Daemon404> j-b, just testing cross-compiule for powerpc (i.e. with psotproc's altivec file), and ill push to a new rpeo. [20:05] <Daemon404> yup works :o [20:10] <Daemon404> j-b / michaelni : https://github.com/dwbuiten/postproc [20:44] <j-b> Daemon404: cool [21:25] <pasteeater> moron question: i know nothing of fate/emulation. how useful would a i7 machine be for fate? it's always on and mostly idle until it has to encode several large batches DV on occasion. [21:34] <Compn> fate is just building ffmpeg using scripts [21:34] <CIA-48> ffmpeg: 03Carl Eugen Hoyos 07master * r44f5b26212 10ffmpeg/libavcodec/cdxl.c: Make CDXL palette opaque. [21:34] <Compn> and cross-compiling, if you have those libs installed [21:34] <Compn> or you could also run virtual machines and compile in that enviroment [21:34] <Compn> itd be very useful [21:35] <Compn> not sure what systems need coverage tho, best ask someone who knows [21:35] <beastd> + exercising the build binaries (with samples) [21:35] <Compn> yeah, testing (md5sum output) of all samples in the fate test [21:36] <pasteeater> that might be a good start. [21:40] <michaelni> what we need for fate ? ... everything thats not there yet and "build only" means just compile / that is we have no hardware to run it on [21:41] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * rf9caec0cf9 10ffmpeg/libavcodec/x86/h264_deblock.asm: [21:41] <CIA-48> ffmpeg: h264: change deblock_h_chroma_8_mmxext() to prevent valgrind confusion. [21:41] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [21:44] <beastd> pasteeater: just some quick thoughts. some bsd variants are missing, some microsoft windows version and cygwin and os/2 is missing completely [22:27] <durandal_1707> getting this: [avi @ 0x29202570]Failed to find timestamp -1456 in index 0 .. 49 [22:48] <durandal_1707> michaelni: ffmpeg summer of code 2012 wiki page is gone [22:52] <michaelni> durandal_1707, if you have a avi that has a seeking bug, give me the file and ill fix it [22:54] <durandal_1707> michaelni: this is created file with ffmpeg [22:54] <durandal_1707> michaelni: sample generated from vsynth [22:54] <michaelni> which ? [22:56] <durandal_1707> michaelni: y41p, r210 [22:57] <durandal_1707> cljr [22:59] <durandal_1707> v308 [22:59] <michaelni> it happens when you seek to before or after the file [23:00] <durandal_1707> it happens if you use left/right with ffplay to seek [23:00] <michaelni> yes [23:01] <durandal_1707> it should not give these error, because there is no such timestamp anyway [23:02] <ohsix> it's told to seek back that far, what's it going to do [23:02] <ohsix> the person seeking should limit the range if it's going to seek to one that doesn't exist :D [23:04] <CIA-48> ffmpeg: 03Paul B Mahol 07master * r545e887902 10ffmpeg/libavcodec/y41penc.c: [23:04] <CIA-48> ffmpeg: y41penc: switch to encode2() [23:04] <CIA-48> ffmpeg: Signed-off-by: Paul B Mahol <onemda(a)gmail.com> [23:04] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [23:04] <CIA-48> ffmpeg: 03Paul B Mahol 07master * r48ad7f822d 10ffmpeg/libavcodec/r210enc.c: [23:04] <CIA-48> ffmpeg: r210enc: switch to encode2() [23:04] <CIA-48> ffmpeg: Signed-off-by: Paul B Mahol <onemda(a)gmail.com> [23:04] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [23:55] <CIA-48> ffmpeg: 03Carl Eugen Hoyos 07master * rf6492476a6 10ffmpeg/cmdutils.c: Fix ffmpeg -codecs output. [00:00] --- Sat Feb 18 2012

1 0

ffmpeg.log.20120217
by burek 18 Feb '12

18 Feb '12

[00:01] <burzum> something else im not sure about is how to create an mp3 with dynamic bitrate using ffmpeg, any hints? [00:02] <Mavrik> um, I think you have to use -aq instead of -ab [00:02] <burzum> also whats the difference between mp3lame and the regular mp3 implementation in ffmpeg? [00:02] <Mavrik> so use a "quality factor" instead of bitrate [00:02] <burzum> -aq works for vorbis but not mp3, the file came out as 192kbps file [00:02] <Mavrik> burzum, hmmm... afaik libmp3lame is the only way to encode mp3 in ffmpeg [00:02] <Mavrik> the other one is just a decoder [00:02] <burzum> Mavrik but im ending up with a mp3 file that plays [00:03] <Mavrik> burzum, ah... you'll have to check on google for "ffmpeg libmp3lame vbr" then, I don't use that particular encoder [00:03] <Mavrik> yeah, because libmp3lame encodes it. [00:03] <Mavrik> check the mapping at the end of output [00:03] <burzum> $params = "-i $in -ab 320k -ar 44100 -f mp3 $out"; [00:03] <Mavrik> -f sets container format, not the codec [00:03] <burzum> [12] => Stream #0.0: Audio: libmp3lame, 44100 Hz, 2 channels, s16, 320 kb/s [00:04] <burzum> oh my fault [00:04] <burzum> --enable-libmp3lame [00:05] <burzum> i missed it the first time i was looking for it [00:05] <Mavrik> that's just a parameter with which ffmpeg has been compiled, has nothing to do with that you're currently encoding [00:08] <VooDooNOFX_> why does ffmpeg insist on outputting via stderr instead of stdout? [00:09] <sacarasc> To annoy you! [00:09] <VooDooNOFX_> heh, no kidding! [00:09] <relaxed> so you can pipe the output [00:10] <relaxed> you can redirect stderr with 2>ffmpeg.log [00:12] <VooDooNOFX_> yeah, it's just more annoying when video encoding process and ffmpeg headers are both on the same output. [00:13] <relaxed> I don't follow [00:13] <VooDooNOFX_> while encoding a video, you get the configuration information printed out everytime [00:13] <VooDooNOFX_> I know I can do -v quiet, but then I don't even get error messages [00:15] <VooDooNOFX_> i'm using ffprobe -print_format json to grab frame and stream information out of a file [00:15] <relaxed> and? [00:16] <VooDooNOFX_> if I run it with a bad command, or on a file that doesn't exist, or anything goes wrong, I get an empty {} string instead of a useful error message when also using the -v quiet flag [00:16] <VooDooNOFX_> otherwise, without -v quiet, I get the ffmpeg configuration information, followed by by json string. [00:17] <VooDooNOFX_> if i'm asking for print_format, shouldn't ffmpeg also output the error messages and the like using that format as well, perhaps even including configuration params. [00:17] <relaxed> check the exit status [00:17] <VooDooNOFX_> yeah, I can check exit status to "that" it failed. I'm more interested in seeing "why". [00:18] <relaxed> ffprobe blah returns: "blah: No such file or directory" [00:19] <VooDooNOFX_> try 'ffprobe -v quiet blah -print_format json' [00:20] <VooDooNOFX_> or ffprobe -v error -print_format json blah [00:21] <relaxed> it returns: blah: No such file or directory\n{\n} [00:21] <VooDooNOFX_> or even much worse: ffprobe -v error -print_format xml blah [00:22] <relaxed> if this is for a script you should test the input file exists [00:22] <relaxed> what do expect to happen? [00:22] <VooDooNOFX_> any number of things can cause ffprobe/ffmpeg to not process a file. missing codecs, bad params, etc [00:23] <relaxed> WHAT DO YOU EXPECT TO HAPPEN? [00:23] <VooDooNOFX_> I would expect that asking print_format json would output { 'error': True, 'error_message': 'blah: No such file or directory' } [00:23] <VooDooNOFX_> formatted output [00:23] <relaxed> the error is right before the JSON output [00:24] <VooDooNOFX_> try the xml option [00:25] <relaxed> yes [00:27] <VooDooNOFX_> what is the intention of print_format then? [00:28] <relaxed> it's self-explanatory [00:30] <relaxed> it's not -print_ffprobe_errors_in json :P [00:30] <VooDooNOFX_> the answer for me was I just saw the -show_errors flag ;) [00:31] <VooDooNOFX_> just not as intui [00:31] <VooDooNOFX_> intuitive as I thought. heh. I'm kind of slow like that. [00:31] <VooDooNOFX_> Is there a -show_all option by any chance? [00:33] <pyoor> Hi all. I'm new to the mpeg format and was hoping someone here might be able to answer some questions for me. Viewing an m4a in atomboxstudio I see a number of fields in the mvhd atom similar to what's outlined here (https://developer.apple.com/library/mac/#documentation/QuickTime/QTFF/QTFFC…) Why are only half of these listed in the ISO (jaadec.sourceforge.net/specs/ISO_14496-12_ISOBMFF.pdf)? [00:34] <pyoor> ahh sorry, just realized this might be better suited for ffmpeg-devel [00:46] <OsmanB> hello , can i send webcam stream thru rtmp on ffmpeg C++ ? i didnt see any sample ? [00:46] <OsmanB> if(avformat_open_input(&pFormatCtx, "rtmp://82.222.191.3/webtv/webtv_wowza1" , NULL, NULL)!=0) <= i do this for reading [00:46] <OsmanB> how can i do output to rtmp ? [00:47] <OsmanB> live out [02:41] <Freakshow> -f flv rtmp rtmp://your.server.whatever:port/path/to/stream [02:41] <Freakshow> and um... really? http://pastebin.com/cnQ99H5k [03:22] <T3mp3sT> OK so I have calculated the bit streams to make a 1GB file. One Q thought on my cmd: [03:23] <T3mp3sT> ffmpeg -i Drive.2011.BluRay.1080p.x264.DTS-LTT.avi -pass 1 -vcodec libx264 -preset veryslow -b:v 1226.72134943k -f matroska -sn -an -y /dev/null && ffmpeg -i Drive.2011.BluRay.1080p.x264.DTS-LTT.avi -pass 2 -vcodec libx264 -preset veryslow -b:v 1226.72134943k -acodec libfaac -ab 128k -ac 2 -sn drive2.mkv [03:24] <T3mp3sT> On the second pass should I be using the avi output or the motroska as defined in pass 1 [03:26] <T3mp3sT> As you see the input is an avi and I tell pass one to dump to motroska. In pass 2 I am still using the input file as avi but the output is mkv [03:27] <relaxed> you can overwrite the 1st pass output [03:27] <T3mp3sT> relaxed could you elaborate? [03:28] <relaxed> for the second pass use -y drive2.mkv [03:28] <relaxed> also, +1 for floating point bitrate [03:28] <T3mp3sT> Thx [03:32] <relaxed> oh, omit the -y as your first pass is sent to /dev/null. [03:33] <relaxed> but yes, matroska is a better, more modern, container than avi [03:34] <MindSpark> hello [03:34] <T3mp3sT> so you are saying: ffmpeg -i movie.avi -pass 1 -vcodec libx264 -preset veryslow -b:v 1226.72134943k -f matroska -sn -an /dev/null && ffmpeg -i movie.avi -pass 2 -vcodec libx264 -preset veryslow -b:v 1226.72134943k -acodec libfaac -ab 128k -ac 2 -sn -y drive2.mkv . Not where I am using the avi and the motroska (pass 1) and output.mkv (pass 2) [03:34] <relaxed> VooDooNOFX_: glad you found a solution [03:35] <relaxed> T3mp3sT: looks good, you're using matroska for both passes. [03:35] <MindSpark> I am trying to make ffserver read a rtmp feed and stream to rtp. For some reason, when I play using vlc or mplayer from my local computer, I get timeouts, as if the file is empty. But I can see the encoding being done by ffserver [03:35] <T3mp3sT> relaxed nice. Thx for your help [03:36] <MindSpark> can anyone help with that? [03:37] <T3mp3sT> relaxed calculating all the bit rates was a pain in the butt to get a 1GB file. Are there any scrips around to do this. If not Ill write one. [03:38] <MindSpark> I need to understand the whole rtp/rtsp/unicast/multicast/rdp concept [03:38] <MindSpark> can someone tell me what all that does and how ffserver utilizes them? [03:44] <T3mp3sT> Seems that on Debian kernel 3.2 is really handling ffmpeg well. I have a 12% speed increase. On Debian [03:45] <T3mp3sT> Think it is dealing with the threading in a "smarter" way...not sure. [03:47] <relaxed> T3mp3sT: add -threads 0 to use all your cores [03:47] <T3mp3sT> I thought that was default? [03:48] <relaxed> I think it is but you never know how old someone's build is [03:49] <T3mp3sT> using the newest gits for x264, ffmpeg and all the goodies you can think of. Pulled them today. [03:53] <relaxed> here's an easy way to calculate the 2pass bitrate: bc -l <<<((($size_in_kilobytes - ($length_in_seconds * ($audio_bitrate / 8)) ) / $length_in_seconds) * 8)" [03:54] <relaxed> where $audio_bitrate is in kbs, like 192 [03:54] <T3mp3sT> Dang, is Linux actually using Turbo Boost starting with this kernel? Since I am using a very nice water cooling system my CPU is at 21C. Normally Boost is dictated by core temps (thus normally useless for this kind of stuff). AFAIK Boost is was not even supported in Linux without weird hacks? Might explain the speed increase. [03:54] <relaxed> missing quote, bc -l <<<"((($size_in_kilobytes - ($length_in_seconds * ($audio_bitrate / 8)) ) / $length_in_seconds) * 8)" [05:05] <Logos01> Howdo, folks. I have an flv file I'm trying to strip the audio out of, but every time I do it introduces distortion at the peaks. (Sounds like the speakers are maxed out even when at 50% volume, to more clearly illustrate my meaning.) [05:06] <Logos01> Is there a pointer of directing of what I could do to address this? I'm already making sure I'm keeping the same ab rate. [05:07] <sacarasc> Logos01: ffmpeg -i foo.flv -vn -acodec copy output.mka [05:07] <sacarasc> \o/ [05:10] <Logos01> Hrm. [05:10] <Logos01> I need these to be mp3 format at the endstate. [05:10] <sacarasc> What is it in the flv? If it's already mp3, just change mka to mp3 and you should be good. [05:11] <Logos01> mp2 [05:11] <Logos01> aac actually, looking more closely. [05:15] <Logos01> This distortion is severe, too. [05:15] <Logos01> Every peak is just torn up. [05:26] <pasteeater> according to what player? [05:26] <Logos01> mplayer [05:26] <Logos01> And Audacity. [05:26] <pasteeater> what about ffplay? [05:26] <pasteeater> use a pastebin site to show your ffmpeg command and the complete console output [05:28] <darkstarbyte> How would I copy 1 video stream 2 audio streams and 2 subtitles streams that are in one video? [05:31] <darkstarbyte> I keep getting an error [05:31] <darkstarbyte> codec id 94210 not found [07:17] <mjb_> Greetings. Not sure if this is the best place to ask, but a recent (Feb. 15) build of ffmpeg is giving me grief about timestamps in a .flv I'm trying to read via the CLI. [07:17] <mjb_> I get a gajillion messages like "DTS 1169265, next:1178010980 st:1 invalid droping" (that's the very last one) and then I get "[flv @ 0000000001C78C80] pts < dts in stream 1" and "av_interleaved_write_frame(): Invalid argument". [07:17] <mjb_> The file plays fine in VLC, but I can't do anything with it in ffmpeg because of this error. It seems to be related to a patch from Feb. 7. I am wondering if the timestamps are actually being parsed correctly or if there's a way to get around this. [07:24] <mjb_> http://pastebin.com/kVdcPpes [07:25] <aidalgol> While re-encoding an MPEG, I got the message "[mp2 @ 0xbffae0] incomplete frame" and then "Error while decoding stream." [07:26] <aidalgol> It appears to have aborted at that point but I can't be sure. Is this a fatal error? [07:35] <aidalgol> OK, I just ran it again, and it failed at the exact frame. [07:36] <aidalgol> Is there any way to make ffmpeg keep going past bad frames? [14:01] <sym-> hi, I'm trying to use ffmpeg.exe to determine duration and frame size for a video file [14:01] <sym-> but am unable to redirect ffmpeg's console output to a text file [14:02] <sacarasc> sym-: Try using something like mediainfo. It's awesome at this kind of job. [14:02] <sacarasc> But you'd have to redirect stderr to stdout to a file. [14:04] <sym-> thank you sacarasc, I see mediainfo is a GUI program? I'm need this information for a ruby script, so it would be more convenient to have it as text though [14:04] <sacarasc> There are GUI and CLI versions. [14:04] <sym-> ah... [14:05] <ubitux> just run ffprobe [14:06] <sym-> ffprobe has the exact same problem unfortunately [14:06] <sym-> sacarasc, thank you very much! mediainfo console version gives me what I need [14:07] <ubitux> ffprobe -show_format -print_format json input > out.json [14:07] <ubitux> this works here [14:07] <sym-> wow, that worked for me too [14:08] <ubitux> you also have xml, compact and various other outputs [14:08] <ubitux> the dump format you wanted to get is printed on stderr, thus ff.. 2> out should work [14:09] <sym-> what does the 2 stand for ? [14:09] <ubitux> file descriptor 2, stderr [14:10] <sym-> wow. I had no idea. :-| thanks [14:11] <ubitux> ffprobe -show_format input.avi 2>/dev/null | grep -E '(duration|size)' [14:11] <ubitux> this works too. [14:14] <sym-> thank you so much sacarasc and ubitux, this saves me a lot of poking about [14:36] <matthiaskrgr> how do I convert several jpgs to a mp4 movie? there were several tutorials on the net but most of them seemed to be outdated :( [14:38] <bgmarete> Hi. With regard to the `vob' format, what option(s) force the generation of a GOP header on each GOP? [14:41] <ubitux> matthiaskrgr: http://ffmpeg.org/ffmpeg.html#Video-and-Audio-file-format-conversion ffmpeg -f image2 -i foo-%03d.jpeg [...] ? [14:41] <ubitux> bgmarete: -g 1 maybe? [14:42] <bgmarete> ubitux: Thanks but that can't be right... That would mean an I-Frame every two frames, which is different from what I want.. [14:43] <ubitux> ok, i don't know then [14:47] <matthiaskrgr> ubitux: hmm, the images are named like current.jpg.1329385713 [14:48] <matthiaskrgr> do I use %011 then ? [14:48] <matthiaskrgr> *%011d [14:53] <matthiaskrgr> current.jpg.* fails as well [15:00] <ubitux> current.jpg.%d ? [15:03] <matthiaskrgr> http://pastebin.com/R3bEgSWT [15:10] <matthiaskrgr> ... the awkward moment when you decide to use "convert" instead of "ffmpeg".. :/ [15:26] <matthiaskrgr> ok, got it working with convert and mencoder :p [16:59] <Freakshow> just in case I missed a reply to this yesterday [16:59] <Freakshow> http://pastebin.com/cnQ99H5k [17:00] <Freakshow> it's curious that I have a version newer on this bsd box and yet configure is complaining that it can't find anything to use [17:01] <Freakshow> enabled libmp3lame && require "libmp3lame >= 3.98.4" lame/lame.h lame_set_VBR_quality -lmp3lame [17:01] <Freakshow> oh sorry... [17:01] <Freakshow> that's indicating the reverse [17:20] <Freakshow> so more over... are there any rundowns on installing for bsd? I see CentOS and Ubuntu [17:20] <Freakshow> in the wiki [17:21] <Freakshow> I guess I could look at the macos on ramdisk setup [17:21] <Freakshow> any thoughts? [17:53] <relaxed> Freakshow: why don't you install ffmpeg-devel from /usr/ports ? [18:06] <Freakshow> relaxed: would that get me latest? [18:07] <Freakshow> I'm trying to test forcing -rtsp_transport tcp and it doesn't seem like it's available until 0.9.1 [18:13] <Freakshow> relaxed: I'll ping you once I'm in the office... gotta get on the road [19:31] <Freakshow> relaxed: the version that's in devel is too old to do what I need... plus I'll never get input from folks here if I'm running anything close to 0.7.1 [19:31] <Freakshow> :( [20:31] <pasteeater> Freakshow: unfortunately nobody has volunteered to make a BSD ffmpeg guide. [20:32] <pasteeater> ...and i'm a *BSD noob. [20:35] <brontosaurusrex> pasteeater: what, ubuntu is not good enough for ya? [20:40] <pasteeater> brontosaurusrex: i have 6 Ubuntu VMs. should i get more? [20:40] <brontosaurusrex> pasteeater: :) bo [20:40] <brontosaurusrex> no* [20:43] <relaxed> Freakshow: install the lastest binutils and gcc from ports, then point ffmpeg's configure to the new gcc and `as` [20:44] <relaxed> Freakshow: Are your ports up-to-date? This looks current to me. http://www.freshports.org/multimedia/ffmpeg-devel/ [20:52] Action: pasteeater votes relaxed as *BSD guide volunteer [21:24] <Freakshow> ugh! relaxed: something is clearly wrong with my ports tree then [21:25] <Freakshow> ours showed a much, much older version [21:25] <Freakshow> I'll look into it [21:25] <Freakshow> thanks for bringing that to my attention [21:43] <Mattias> I'm trying to use ffserver, I got the sample config file and haven't changed much. when I try to do fmpeg -i videofile.mkv http://localhost:8090/feed1.ffm I get this: [tcp @ 0x155bdc0] TCP connection to localhost:8090 failed: Connection refused <-- and a HTTP 503 error [21:43] <Mattias> on server side I see: Fri Feb 17 21:40:31 2012 127.0.0.1 - - [GET] "/feed1.ffm HTTP/1.1" 503 308 [21:44] <Mattias> I can access the stat page fine [21:50] <burek> Mattias, probably access problems [21:50] <burek> check file permissions [21:50] <burek> oralso, try with an empty config (as much as possible empty), and uncomment things slowly, until you get 503 again [21:51] <burek> -oralso +or [21:58] <Mattias> burek: thanks, somehow I got it to work by commenting stuff [22:02] <burek> :beer: :) [22:09] <Mattias> burek: you don't happen to have awesome settings for lan streaming? ^.^ so I kind of get the same hd quality the video has that I want to stream. [22:09] <Mattias> I've tried increasing things a lot but doesn't seem to help much beyond a certain point, lots of small "boxes" can be seen [22:10] <Freakshow> pasteeater / relaxed: do either of you know which version that the current devel branch is going to give me? [22:23] <burek> Mattias, to be honest [22:23] <burek> I couldn't manage to config ffserver] [22:23] <burek> so I temporarily switched to vlc [22:23] <burek> but I'll check back ffserver soon to test if things are any better now [22:26] <Mattias> I've tried experimenting a little now, using avi, video and audio gets out of sync, using mpeg, I get loads of blocks. using asf, sound hardly works.. I wonder if I should just crank up bitrate over 10000 [22:26] <Mattias> with mpeg [22:31] <Mattias> yeah, gotta experiment some more with mpeg, getting good quality now but I probably cranked it up too much :) [22:37] <Mattias> great, got it good now :) [22:47] <pasteeater> Freakshow: i do not know. sorry. [23:00] <Freakshow> no worries [23:57] <Austen> anybody know why this error shows up when building ffmpeg? [23:57] <Austen> libopenjpeg.a(cio.o): relocation R_X86_64_32 against `a local symbol' can not be used when making a shared object; recompile with -fPIC [23:57] <Austen> ive turned on -fPIC and enable-pic [23:58] <Austen> and with libopenjpeg as well... [00:00] --- Sat Feb 18 2012

1 0

ffmpeg-devel.log.20120216
by burek 17 Feb '12

17 Feb '12

[01:46] <CIA-48> ffmpeg: 03Kostya Shishkov 07master * r1a265f6187 10ffmpeg/ (8 files in 3 dirs): prores encoder [01:46] <CIA-48> ffmpeg: 03Martin Storsjö 07master * r735be9cdfb 10ffmpeg/libavformat/rtspdec.c: [01:46] <CIA-48> ffmpeg: rtsp: Make rtsp_demuxer_class static [01:46] <CIA-48> ffmpeg: Signed-off-by: Martin Storsjö <martin(a)martin.st> [01:47] <CIA-48> ffmpeg: 03Martin Storsjö 07master * ra082ac4125 10ffmpeg/libavcodec/mpegvideo_common.h: [01:47] <CIA-48> ffmpeg: mpegvideo: Remove a leftover function declaration [01:47] <CIA-48> ffmpeg: Signed-off-by: Martin Storsjö <martin(a)martin.st> [01:47] <CIA-48> ffmpeg: 03Martin Storsjö 07master * r167f3b8de7 10ffmpeg/libavformat/ (28 files): [01:47] <CIA-48> ffmpeg: libavformat: Add an ff_ prefix to some lavf internal symbols [01:47] <CIA-48> ffmpeg: Prefix the functions/tables brktimegm, pcm_read_seek, [01:47] <CIA-48> ffmpeg: dv_offset_reset, voc_get_packet, codec_movaudio_tags, [01:47] <CIA-48> ffmpeg: codec_movvideo_tags. [01:47] <CIA-48> ffmpeg: After this, lavf has no global symbols without the proper prefix. [01:47] <CIA-48> ffmpeg: Signed-off-by: Martin Storsjö <martin(a)martin.st> [01:47] <CIA-48> ffmpeg: 03Martin Storsjö 07master * rddce8953a5 10ffmpeg/libavcodec/ (18 files): [01:47] <CIA-48> ffmpeg: h263: Add ff_ prefix to nonstatic symbols [01:47] <CIA-48> ffmpeg: Signed-off-by: Martin Storsjö <martin(a)martin.st> [01:47] <CIA-48> ffmpeg: 03Martin Storsjö 07master * re96b4a53df 10ffmpeg/libavcodec/ (36 files): [01:47] <CIA-48> ffmpeg: vlc/rl: Add ff_ prefix to the nonstatic symbols [01:47] <CIA-48> ffmpeg: Signed-off-by: Martin Storsjö <martin(a)martin.st> [01:47] <CIA-48> ffmpeg: 03Martin Storsjö 07master * r9cf0841ef3 10ffmpeg/libavcodec/ (75 files in 9 dirs): [01:47] <CIA-48> ffmpeg: dsputil: Add ff_ prefix to the dsputil*_init* functions [01:47] <CIA-48> ffmpeg: Signed-off-by: Martin Storsjö <martin(a)martin.st> [01:47] <CIA-48> ffmpeg: 03Martin Storsjö 07master * r99560a4caa 10ffmpeg/libavcodec/ (7 files): [01:47] <CIA-48> ffmpeg: libavcodec: Add ff_ prefix to some nonstatic symbols [01:47] <CIA-48> ffmpeg: Prefix the functions atrac_generate_tables, atrac_iqmf, dct_quantize_c. [01:47] <CIA-48> ffmpeg: Signed-off-by: Martin Storsjö <martin(a)martin.st> [01:47] <CIA-48> ffmpeg: 03Martin Storsjö 07master * r3e2efacdd8 10ffmpeg/libavcodec/ (dct-test.c dsputil.c dsputil.h jfdctfst.c mpegvideo_enc.c): [01:47] <CIA-48> ffmpeg: libavcodec: Prefix fdct_ifast, fdct_ifast248 [01:47] <CIA-48> ffmpeg: Signed-off-by: Martin Storsjö <martin(a)martin.st> [01:47] <CIA-48> ffmpeg: 03Martin Storsjö 07master * r873c89e2a6 10ffmpeg/libavcodec/ (dsputil.c x86/mpegvideo_mmx.c x86/mpegvideo_mmx_template.c): [01:47] <CIA-48> ffmpeg: dsputil: Add ff_ prefix to inv_zigzag_direct16 [01:47] <CIA-48> ffmpeg: Signed-off-by: Martin Storsjö <martin(a)martin.st> [01:47] <CIA-48> ffmpeg: 03Kostya Shishkov 07master * r8835c2c829 10ffmpeg/libavcodec/ (Makefile proresdata.c proresdata.h proresdec.c): prores: move data shared between decoder and encoder to common file [01:47] <CIA-48> ffmpeg: 03Kostya Shishkov 07master * r3ec623c22f 10ffmpeg/libavcodec/proresdec.c: [01:47] <CIA-48> ffmpeg: prores: fix multithreaded decoding case when slice quantisers are not the same [01:47] <CIA-48> ffmpeg: Since quantisation matrices are stored in context, decoding slices with [01:47] <CIA-48> ffmpeg: different quantisers in parallel leads to unpredictable content of [01:47] <CIA-48> ffmpeg: aforementioned matrices and wrong output picture thereof. [01:47] <CIA-48> (57 lines omitted) [02:26] <Zeranoe> Any updates on 'error: 'wcscpy_instead_use_StringCbCopyW_or_StringCchCopyW' undeclared (first use in this function)'? paste is here: http://pastebin.com/BmRH7HVt do I need to submit it as a bug? [03:23] <CIA-48> ffmpeg: 03Pilotware 07master * r7606825350 10ffmpeg/libavcodec/dnxhd_parser.c: [03:23] <CIA-48> ffmpeg: Handle interlaced formats (two coding units together make one full frame) [03:23] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [03:28] <Loof> Zeranoe: Hey, is there a 'stable' build of 0.10 for (x64) windows? The constant builds are nice, but I'd rather not track a moving target [03:30] <Loof> Or maybe I'm missing the point and there's some other place I should look at to determine which ref has the most big fixes :D [03:34] <Loof> Hmm [04:17] <CIA-48> ffmpeg: 03Mashiat Sarker Shakkhar 07master * r3ce50bf51d 10ffmpeg/libavcodec/wmalosslessdec.c: Cosmetics: fix some whitespace errors [04:17] <CIA-48> ffmpeg: 03Mashiat Sarker Shakkhar 07master * r6345184fc9 10ffmpeg/libavcodec/wmalosslessdec.c: Support 24-bit decoding [04:17] <CIA-48> ffmpeg: 03Mashiat Sarker Shakkhar 07master * r59df25effd 10ffmpeg/libavcodec/wmalosslessdec.c: Move some variable declarations to comply with C90 [04:17] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * rdc08e7ea0f 10ffmpeg/libavcodec/wmalosslessdec.c: [04:17] <CIA-48> ffmpeg: wmall: remove ;; [04:17] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [04:17] <CIA-48> ffmpeg: 03Mashiat Sarker Shakkhar 07master * rd1ea26f640 10ffmpeg/libavcodec/wmalosslessdec.c: [04:17] <CIA-48> ffmpeg: Do not try to read residue if ave_mean <= 1 [04:17] <CIA-48> ffmpeg: Otherwise, we end up with with log(0) or log(1). av_ceil_log2 simply [04:17] <CIA-48> ffmpeg: assumes the argument is non-zero and returns wrong result when it is. [04:17] <CIA-48> ffmpeg: (Not that there is a proper way of returning an undefined value.) [04:17] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * r28ba1027ec 10ffmpeg/: (log message trimmed) [04:17] <CIA-48> ffmpeg: Merge remote-tracking branch 'shariman/wmall' [04:17] <CIA-48> ffmpeg: * shariman/wmall: [04:17] <CIA-48> ffmpeg: Do not try to read residue if ave_mean <= 1 [04:17] <CIA-48> ffmpeg: Move some variable declarations to comply with C90 [04:17] <CIA-48> ffmpeg: Cosmetics: fix some whitespace errors [04:17] <CIA-48> ffmpeg: Support 24-bit decoding [05:41] <Compn> getting another prores encoder? :p [05:41] <Compn> ehe [06:38] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * r0f155c8601 10ffmpeg/libavcodec/dct-test.c: [06:38] <CIA-48> ffmpeg: fate: trying to fix "libavcodec/dct-test.o:(.rodata+0xdc): undefined reference to fdct_altivec" [06:38] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [06:38] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * rb92ecda2c9 10ffmpeg/libavcodec/snow.c: [06:38] <CIA-48> ffmpeg: snow: zero scratchbuf. [06:38] <CIA-48> ffmpeg: This should fix some valgrind warnings. [06:38] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [18:11] <Zeranoe> How does FFmpeg utilize pulseaudio and is it practical in a Windows setting? [18:11] <av500> it doesnt [18:12] <av500> in fact it does [18:12] <av500> there is a PA indev [18:15] <iive> pulse audio is horrible and ridden with bugs that are not fixed for years. avoid it when possible. [18:17] <Compn> and it wasnt designed for video syncing with audio [18:17] <Compn> which is hilarious... [18:17] <av500> Compn: I think the only design requirement was to be marginally better than the crap before [18:18] <av500> e.g. arts [18:19] <iive> I would rather blame the implementation of the api. and i'm not sure it is better than arts... [18:19] <av500> I did not say the implementation was up to the spec :) [18:21] <iive> oh, btw, does somebody know if there is official pulseaudio api documentation? [18:23] <Zeranoe> So --enable-libpulse is not practical? [18:25] <iive> avoid when possible. Sometimes you have no other option. [18:26] <av500> Zeranoe: as said, the only pulse stuff is that ffmpeg can read from pulse [18:26] <av500> read the git log for more details [18:38] <Compn> Zeranoe : what did you want to use pulse for ? [18:38] <Compn> ffmpeg has dshow input on win32 i think ? [18:38] <Compn> or did you want some kind of windows audio server ? [18:39] <Compn> in which case, esound or nas might do it [18:39] <Compn> i dont know if PA works on windows [18:39] <Compn> and if ffmpeg can use PA on windows [18:41] <av500> why would one want PA on windows? [18:42] <av500> on windows every app can open the equivalent of /dev/dsp and play sound [18:42] <av500> it's just linux that refuses to do this [18:42] <Compn> alsa allowed this iirc :P [18:42] <Compn> or oss [18:42] <Compn> one of them [18:42] <Compn> maybe both with the right opts [18:42] <av500> oss depended on the sound card [18:43] <Compn> oh thats it [18:43] <av500> and back then, mixing in kernel was deemed too cpu heavy or do [18:43] <av500> so [18:43] <Compn> heh [18:43] <av500> or not kernel-worthy [18:43] <av500> so now we have the mess [18:43] <Compn> who veto'd that in the kernel anyhow [18:43] <Compn> everyone ? linus? [18:43] <av500> no idea [18:44] <av500> that might have been pre-internet even :) [18:44] <Zeranoe> I saw that FFmpeg supports it, and it can be compiled for Windows, I was just checking how FFmpeg uses it. [18:44] <av500> ffmpeg does not use it [18:44] <av500> ffmpeg users might use it [18:46] <Zeranoe> av500: I just noticed that there was support for the lib [18:47] <Compn> you wanted to put support in win32 builds ? :P [18:48] <Zeranoe> Compn: Yes, I was considering it, but I probably wont now. [19:10] <iive> oss4 have internal software mixer. kernel people don't like it, so they push it to user land, but this cannot sustain low latency and smoothness at the same time... [19:11] <iive> so the userland program needs realtime priority... that brings more troubles than letting the internal software mixer. [19:14] <av500> iive: full ack [19:17] <michaelni> iive, just curious, who exactly of the kernel people? [19:18] <iive> sorry, don't remember... [19:18] <michaelni> np, isnt important .... [19:25] <Compn> michaelni : you should bring it back up on kernel list, see if anyone agrees :) [19:39] <Daemon404> audio on linux is lol. [19:40] <av500> yes :( [19:41] <Daemon404> http://blogs.adobe.com/penguinswf/2007/05/welcome_to_the_jungle.html <-- reminds me of this [19:41] <Daemon404> by mike melanson [19:41] <av500> I found this one today: http://insanecoding.blogspot.com/2009/06/state-of-sound-in-linux-not-so-sor… [19:41] <nevcairiel> and it hasnt changed a bit since 2007, except there is another blob for oss4 :P [19:42] <Daemon404> nevcairiel, :D [19:42] <av500> blob? [19:42] <av500> ah [19:42] <nevcairiel> on the jungle map :) [19:42] <Daemon404> i've always used windows for audio playback [19:42] <Daemon404> i used to periodically try linux [19:42] Action: av500 was thinking binary blob [19:42] <Daemon404> and lol, and go back. [20:23] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * r67f17e7202 10ffmpeg/libavcodec/proresdec2.c: [20:23] <CIA-48> ffmpeg: proresdec2: Fix 16 vs. 8bit compiler warnings. [20:23] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [20:23] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * r5a03a3d998 10ffmpeg/libavcodec/proresdec2.c: [20:23] <CIA-48> ffmpeg: proresdec2: use ff_dsputil_init() avoid deprecated ff_ less variant. [20:23] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [20:23] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * r8b18288c08 10ffmpeg/libavcodec/qtrleenc.c: [20:23] <CIA-48> ffmpeg: qtrleenc: fix uninitialized variable warnings [20:23] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [22:08] <Freakshow> who maintains the freebsd port of ffmpeg [22:09] <Freakshow> http://pastebin.com/4x7bjK3U [22:11] <iive> are you sure this is ffmpeg related? the error is about opencv. [22:12] <Freakshow> you know... this is true, perhaps it's a portinstall issue... opencv is there... the ffmpeg makefile is calling for opencv-core, somewhere along the way things are getting confused [22:19] <funman> ffmpeg probably depends on opencv.. [22:22] <Freakshow> that is a certainty [22:22] <Freakshow> but for whatever reason, portinstall/portupgrade isn't able to get past that point in the install [22:23] <Freakshow> is this trac worthy? [22:24] <funman> freebsd trac, yeah [22:24] <Freakshow> ah, you live up to your name [22:25] <Freakshow> ;) [22:25] <Freakshow> I will do the needful [22:25] <funman> Freakshow: cd /usr/ports/graphics/opencv-core; make install [22:25] <Freakshow> thanks! [22:26] <Freakshow> well, no need to clutter the channel with that stuff... but I've tried that [22:27] <michaelni> ffmpeg straight from git compiles & works fine on freebsd [22:27] <michaelni> with no ports & stuff ... [22:29] <michaelni> we have a freebsd box that tests git ffmpeg regularly (see fate.ffmpeg.org) [22:31] <Freakshow> thanks michaelni, I've found it necessary to update to head in my test environment... bsd 8.1 [22:31] <Freakshow> so I'll be giving that a shot shortly [22:31] <Freakshow> and disregarding the current ports version [23:04] <funman> michaelni: for #999 see http://git.videolan.org/?p=vlc.git;a=commitdiff;h=bb7509af051d8384b9266bcdb… [23:08] <funman> michaelni: gave myself an account and wrote that on the ticket [23:09] <michaelni> funman, thx :) [23:09] <michaelni> i just wanted to copy and paste ... [23:10] <funman> #define DSHOW_NO_STRSAFE before including dshow should work [23:10] <funman> hmm not sure [23:10] <funman> let me try to make a patch [23:11] <funman> Unknown option "--enable-dshow". [23:14] <michaelni> maybe its --enable-indev=dshow [23:18] <funman> michaelni: oops don't apply the patch it's not enough [23:20] <funman> now it's correct (builds libavdevice.a) [23:22] <durandal_1707> vf_deshake cant use ff_dsputil_init() [23:27] <michaelni> funman, is there a patch somewhere that i should apply ? [23:27] <michaelni> durandal_1707, yes :/ [23:27] <funman> michaelni: i added it inline to the ticket [23:27] <michaelni> ohh ok [23:28] <funman> waiting to see if trac fucked up the patch format :) [23:29] <michaelni> it seems to have applied with no complaints [23:30] <funman> {{{ }}} seems to do the job, cool [23:30] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * rad21c05172 10ffmpeg/libavcodec/diracdec.c: [23:30] <CIA-48> ffmpeg: diracdec: fix use of deprecated funnctions [23:30] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [23:30] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * r934abf6ff8 10ffmpeg/libavcodec/g729dec.c: [23:30] <CIA-48> ffmpeg: g729dec: fix use of deprecated functions. [23:30] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [23:30] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * r22b125f912 10ffmpeg/libavcodec/g729dec.c: [23:30] <CIA-48> ffmpeg: g729dec: fix 'warning: assignment from incompatible pointer type' [23:30] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [23:30] <CIA-48> ffmpeg: 03Rafaël Carré 07master * r05ee0db1b2 10ffmpeg/libavdevice/ (dshow_filter.c dshow_pin.c): (log message trimmed) [23:30] <CIA-48> ffmpeg: dshow_filter: fix compilation with mingw-w64 [23:30] <CIA-48> ffmpeg: Closes ticket #999 [23:30] <CIA-48> ffmpeg: NO_DSHOW_STRSAFE asks dshow.h header to not use secure string function [23:30] <CIA-48> ffmpeg: replacements. [23:30] <CIA-48> ffmpeg: Using secure replacements would break mingw.org compatibility as they don't [23:30] <CIA-48> ffmpeg: declare/define those functions. [23:32] <durandal_1707> hmm url_feof will always return 0 [23:33] <durandal_1707> 3bd624b4ee355328382295af411f5343b1329f43 introduced that funny behaviour [23:34] <michaelni> hmm, it really shouldnt do that ... [23:35] <durandal_1707> lol, fill_buffer() will set eof_reached back to 1 [23:36] <michaelni> of course [23:37] <michaelni> also a assert() placed there triggers on the first test of fate so its not always returning 0 [23:37] <michaelni> :) [23:40] <durandal_1707> so this is for case if data is appended when file was played? [23:41] <durandal_1707> not all demuxer use this function, should they? [23:45] <michaelni> probably yes [23:46] <durandal_1707> lol, but libav devs ditched this function [00:00] --- Fri Feb 17 2012

1 0

ffmpeg.log.20120216
by burek 17 Feb '12

17 Feb '12

[00:11] <quebre> hello [00:11] <quebre> i want to convert RAW format that i've recorded using Fraps to 720p quality so i can put it on youtube [00:11] <quebre> i tried few times but it's always not full screen and it's always ugly [00:12] <quebre> so the last syntax i have tried is: ffmpeg -i bf3-2012-02-13-21-53-53-84.avi -vcodec libx264 -vpre 720p bf3-4.mp4 [00:12] <quebre> but it says: [00:12] <quebre> File for preset '720p' not found [00:12] <quebre> can someone point me to good way of encoding RAW to 720p ? [00:14] <quebre> Noone? :) [00:14] <quebre> oh.. come on... it can't be THAT hard? [00:19] <vadim> quebre: try '-fpre' followed with the full path to the presset file [00:20] <quebre> hmm [00:20] <quebre> and if i dont have the preset file? where i can get good preset for 720p ? [00:23] <relaxed> quebre: ffmpeg -i bf3-2012-02-13-21-53-53-84.avi -vcodec libx264 -preset slow -crf 18 bf3-4.mp4 [00:24] <quebre> relaxed: Error while opening encoder for output stream #0:1 - maybe incorrect parameters such as bit_rate, rate, width or height [00:25] <relaxed> pastebin.com the full output [00:25] <quebre> ok [00:26] <quebre> http://pastebin.com/qYwfWwDe [00:28] <quebre> getting such errors since two days [00:28] <quebre> no idea what's wrong [00:29] <relaxed> quebre: ffmpeg -i bf3-2012-02-13-21-53-53-84.avi -vcodec libx264 -preset slow -crf 18 -acodec libfaac -ac 2 -ab 128k bf3-4.mp4 [00:29] <quebre> lets try [00:29] <pasteeater> the important message is: libaacplus doesn't support this output format! [00:30] <relaxed> seems like that should have been the last message: "ERROR: libaacplus doesn't support this output format!" [00:30] <pasteeater> so use settings that work with libaacplus or use a different aac encoder. such as libvo-aacenc or libfaac, both of which your ffmpeg supports [00:30] <quebre> relaxed: now it works [00:30] <quebre> but what changed ? [00:30] <quebre> mhm [00:30] <relaxed> pasteeater just explained why [00:30] <quebre> yes just noticed [00:31] <quebre> so it will be now GOOD quality ? [00:31] <quebre> i mean video [00:31] <quebre> dont care much for audio.. [00:31] <pasteeater> take a look. you probably won't see a difference between source and output with relaxed's example [00:32] <quebre> great ! [00:32] <relaxed> you can encode a 10 second sample if you add "-t 10" to the command [00:32] <quebre> oh [00:32] <quebre> coz im encoding 4gb portion [00:32] <quebre> would take looong to test it [00:34] <quebre> looks identical [00:34] <quebre> :} [00:34] <pasteeater> good, because youtube will poop on it when they re-encode it [00:35] <quebre> what do you mean ? [00:35] <quebre> they will make it look ugly ? [00:37] <quebre> i wonder what soft youtube is using to re-encode [00:37] <quebre> maybe it's ffmpeg ;p [00:37] <pasteeater> i mean that they re-encode anything you give it so it's best to give them something nice looking to begin with [00:39] <quebre> ah [01:44] <quebre> pasteeater: are you here ? [02:42] <relaxed> CodeH3x: pastebin.com the output of "ffmpeg -i video.mp4 -i input.mkv" [02:59] <mininessie> i need a command to capture multiple screencaps of a video at the same time meaning when i use the command it will every 2 seconds take a capture of said video without having to manually change the command every 2 seconds [03:01] <mininessie> aynone here [03:01] <smj> what happened to -vf pp? it says "No such filter: pp" is there a new way to use it? [03:02] <mininessie> smj: can you help me [03:04] <smj> maybe [03:05] <mininessie> okay did you see what i said earlier [03:05] <smj> what does multiple screencaps of a video at the same time mean? [03:08] <mininessie> i mean the video i want i need multiple screen caps of it but i do not want to manually type in a command and change the time in the video the capture is taken and run the command once and ever 2 seconds takes a capture so at 2 seconds capture one 4 seconds capture 2 and so on until video is done [03:08] <pixeldonor> i'm trying to compile ffmpeg with mp3lame using the --enable-libmp3lame. i've downloaded and compiled mp3lame, but it doesn't seem to be able to find it when i use ./configure. [03:08] <pixeldonor> any way to tell it where to look? [03:12] <smj> mininessie: I'll try that [03:17] <smj> I guess he got the help somewhere else [03:23] <VooDooNOFX_> is there a better way to add metadata tags to a video file other than ffmpeg -i inputfile -metadata title="Movie Title" -metadata year="2010" outputfile ? [03:48] <Randy_T> Is it possible use deblock with a zones option with x264. For example, "-x264opts zones=0,1000,b=2,deblock=-1:-1". The problem here is that ffmpeg separates x264 options with ":" colon and the x264 zones option separates it's own options with "," comma. [03:49] <relaxed> VooDooNOFX_: inputfile doesn't have any metadata set? [03:49] <VooDooNOFX_> relaxed, it has some, but i'm trying to bake in more to the file [03:50] <relaxed> Randy_T: use -deblock [03:51] <relaxed> VooDooNOFX_: then I guess that's the only way [03:52] <Randy_T> I want to use that deblock option in that particular zone. if I substitute a comma there than it causes a problem with the zone option, and if i leave it as-is, it's an ffmpeg error [03:53] <relaxed> if you need that type of control I would use x264 directly [03:53] <relaxed> did you try quoting the colon? \: [03:54] <relaxed> or -x264opts 'zones=0,1000,b=2,deblock=-1:-1' [03:55] <Randy_T> Thanx, but both those didnt work. [03:55] <Randy_T> Maybe somethign with profiles.. [04:13] <jshanab_win7> I have a large application that uses both live555 and libavcodec. As I am researching adding RTMP protocols to my server I ran across mention that ffmpeg also does streaminces in the two streaming libraries,. Can soomeone tell me the difference between to 2 libs, Can I maybe eliminate live555 from my app? [04:17] <smj> Randy_T: does x264opts zones=0,1000,b=2,deblock=-1;-1 do the trick? [04:19] <Randy_T> hallelujah, it worked! [04:19] <smj> :) [04:22] <Randy_T> I was gona try to make an ffmpeg preset, but i was guessing ffmpeg just takes the options in the preset and plops them on the command line anyway. Didn't think to try a semi-colon, Thank you. [04:34] <Randy_T> fyi. It still needs to be in quotes too. so '-x264opts zones=0,1000,b=2,deblock=-1;-1' [04:34] <Randy_T> oops. i mean quotes for 'zones=0,1000,b=2,deblock=-1;-1 [07:05] <darkstarbyte> What do I have to install to convert wmv files? [07:16] <overflow_0f8b> guessing, wmv codec? [09:10] <zap0> darkstarbyte, ffmpeg [11:46] <CodeH3x> hey [11:46] <CodeH3x> anyone here? [12:02] <jensverwiebe> hi relaxed , thx for your supply to help looking into ffv1 prob [12:03] <jensverwiebe> relaxed: turned out it is just a breaking change making use of latest version in the procuction pieline nessecary [12:04] <jensverwiebe> relaxed: all fine now, just near all players around are a bit outdated in ffmpeg regards, so sonŽt support ffv1 transparency yet [12:04] <jensverwiebe> cyaŽs folks, go on with great work :) [12:31] <Guest23090> hi. I just found http://www.ogg-converter.net which charges for a frontend to (a statically linked) ffmpeg without attributing ffmpeg. [12:31] <Guest23090> This seems to be in conflict with the LGPL. Is there an established procedure for such cases? [12:32] <Mavrik> hm [12:32] <Mavrik> it seems hall of shame page is offline [12:37] <raphman> http://web.archive.org/web/20101003203602/http://ffmpeg.org/shame.html [12:39] <raphman> hmm, ok, I'll have a look at the bug tracker. [16:11] <huglester> Guys, I get: Error initializing filter 'movie' with args 'watermark_video.png' [16:11] <huglester> Segmentation fault [16:12] <huglester> but If I replace with .jpg iamge (same image just saved as jpg) - it works fine [16:51] <MindSpark> hello, can anyone help me out with good oss to convert/transcode rtsp to rtmp? [17:12] <gingerling> hello [17:12] <gingerling> trying to convert a gif animation from commons to movie format to work in pitivi [17:13] <gingerling> anyone know how to do this. I am a gui user really [17:17] <gingerling> i found a ton of terminal manuals but dont understand what they mean, can anyone spare a sec to teach me? [17:21] <MindSpark> gingerling, ffmpeg takes an input files specified by -i [17:21] <MindSpark> and an output file [17:21] <MindSpark> you can add options to it [17:21] <MindSpark> if you're under *nix, simply do man ffmpeg [17:22] <gingerling> i have seen [17:22] <gingerling> http://www.linuxquestions.org/questions/linux-software-2/converting-animate… [17:22] <gingerling> i have messed arround a bit [17:22] <gingerling> but just get the first image in the gif [17:22] <gingerling> people talk about frame rate [17:22] <gingerling> i have changed this in gui but dosnt seem to help (using winff) [17:31] <MindSpark> sorry dc [17:32] <gingerling> np [17:33] <MindSpark> so you tried something very basic such as ffmpeg -i file.gif file.avi ? [17:35] <gingerling> i dont understand how to use this in terminal, sorry, i know its a pain. I am a real gui user [17:35] <gingerling> i tried using winff, which is the gui ffmpeg, and tried all the obvious things [17:35] <gingerling> but it just outputs the first frame of the giff [17:43] <MindSpark> then try to cli [17:43] <MindSpark> commandline [17:43] <MindSpark> type the command I told you [17:46] <gingerling> putting my file name inthere? [17:49] <gingerling> dosnt do anything, lots of no such file or directorys [17:56] <T3mp3sT> Hi, was wondering if there is something wrong with this cmd. ffmpeg -i videoIn.avi -vcodec copy -acodec libfaac -ab 128k -ac 2 -sn videoOut.mkv [17:56] <T3mp3sT> This use to work for me all the time. [18:02] <KyranBe> T3mp3sT, it would help if you could specify what is not working [18:02] <T3mp3sT> KyranBe sure, one sec. My bad. [18:04] <T3mp3sT> Stream mapping: [18:04] <T3mp3sT> Stream #0:0 -> #0:0 (copy) [18:04] <T3mp3sT> Stream #0:1 -> #0:1 (ac3 -> libfaac) [18:04] <T3mp3sT> Press [matroska @ 0x30640e0] Can't write packet with unknown timestamp [18:04] <T3mp3sT> av_interleaved_write_frame(): Invalid argument [18:05] <T3mp3sT> KyranBe thats the error I am getting. Weird. I recompile everything again from git last night but it did not help. [18:07] <T3mp3sT> whats the stuff about "unknown timestamp" ? [18:07] <KyranBe> T3mp3sT I'm no expert, so I'm afraid I can't help you out with this, but now at least someone more knowledgeable than me can see what's wrong [18:08] <T3mp3sT> KyranBe appreciate you attempting to help me out with this issue. [18:08] <KyranBe> you could try to demux the audio, convert the audio stream separately and then remux everything in an mkv [18:09] <T3mp3sT> KyranBe to avoid all that I just did the -sn to get rid of any fluffy streams. But Ill try anything. [18:11] <T3mp3sT> hmmm.... might might have an idea what this is about. I use a SSD and in fstab one of the statements is noatime. This is commonly done with SSDs to prevent ware. Maybe ffmpeg is using time in some weird way. [18:12] <KyranBe> I doubt it, I have noatime for all my drives [18:12] <T3mp3sT> KyranBe good to know. [18:14] <KyranBe> I get the same error here btw [18:15] <KyranBe> even with acodec copy [18:15] <T3mp3sT> @@ [18:15] <KyranBe> instead of the mkv container use the mp4 container [18:16] <KyranBe> then remux the mp4 file to an mkv file [18:16] <KyranBe> that works [18:16] <KyranBe> (for me at least) [18:17] <T3mp3sT> Ill give it a try. [18:18] <T3mp3sT> So "-vcodec copy" does now work for you?? [18:18] <KyranBe> yeah [18:18] <KyranBe> I did this [18:19] <KyranBe> ffmpeg -i input.avi -vcodec copy -acodec faac -ab 128k -ac 2 temp.mp4 [18:19] <KyranBe> then [18:19] <KyranBe> ffmpeg -i temp.mp4 -vcodec copy -acodec copy out.mkv [18:20] <KyranBe> those two should produce the same output, but don't give the error [18:20] <KyranBe> it's just a workaround [18:20] <T3mp3sT> lol, sure why not. Your only changing the container around in the last cmd there. Could just drop in a && between them [18:20] <T3mp3sT> a workaround that works [18:21] <T3mp3sT> KyranBe thx you have been more then helpful. [18:22] <KyranBe> no problem [18:24] <T3mp3sT> Nice works just fine now. [18:43] <MindSpark> hello, can anyone help me out with good oss to convert/transcode rtsp to rtmp? [19:38] <burek> MindSpark, [19:38] <burek> can you please use pastebin.com, to show your command line and its output? [19:40] <MindSpark> burek, I have no cl and no output, I was asking about what I could do if I have an rtmp url which I want to convert to rtsp [19:41] <burek> I'm really not familiar with rtsp.. if it's a container, use -f if it's a vcodec use -vcodec [19:41] <burek> for ex. [19:41] <burek> ffmpeg -i input.rtmp -f rtsp output.rtsp [19:41] <burek> or [19:41] <burek> ffmpeg -i input.rtmp -vcodec rtsp output.rtsp [19:44] <MindSpark> burek, rtmp is not a container as far as I understand. It's a protocol [19:44] <MindSpark> I mean rtsp [19:44] <MindSpark> both [19:44] <MindSpark> lol [19:44] <burek> oh :D [19:44] <burek> ^^ [19:45] <burek> well, take a look at ffserver then [19:45] <burek> also this [19:46] <MindSpark> burek, thanks! will look into that! [19:46] <burek> :beer: :) [19:47] <MindSpark> haha, my pleasure, you just come pick it up [19:47] <MindSpark> Egypt ;) [19:47] <burek> ^^ :) I will one day :) I will :) [19:48] <burek> It's just across 1 sea :D [19:49] <MindSpark> :) [20:04] <slackyman> hi [20:04] <slackyman> I have a little problem cross compiling ffmpeg [20:04] <slackyman> with mingw-64 [20:05] <slackyman> mingw64 on Ubuntu 11.10 comes without winpthreads [20:05] <slackyman> if I compile pthreads I get an error [20:07] <huglester> guys, I'm getting the: " Could not find codec parameters (Video: png)" .. how do I enable png support? I want to have watermark (png) on my images [20:08] <relaxed> pastebin.com your command and all output [20:08] <huglester> I tried recompiling to newest version.. added: --enable-decoder=png but still no luck [20:09] <relaxed> you shouldn't need to do that [20:10] <huglester> relaxed: maybe you have any other ideas on what is needed to be done? [20:10] <huglester> if I replace png iamge with jpg image - everything works [20:11] <relaxed> how about you pastebin.com your command and all output [20:14] <pasteeater> png decoder and encoder requires zlib [20:14] <pasteeater> but we can only guess until you supply command and output as relaxed asked several times [20:15] <burek> apt-get install libpng12-dev [20:16] <huglester> hm.. wait I'll just tried toisntall zlib dev [20:16] <huglester> recompiling ffmpeg [20:25] <huglester> thanks relaxed, pasteeater, burek. I have installed zlib1g-dev (on debian squeeze) and now it worsk:)! [20:26] <huglester> thanks for kind support :) [20:37] <slackyman> can I use mingw64 zeranoe build integrating it with my mingw64 pre-existent tool-chain? [20:53] <tiglionabbit> I want to use ffmpeg with libfaac on ubuntu. I'm pretty sure I have all the relevant packages installed, but still I get "Unknown encoder 'libfaac'" http://friendpaste.com/2CWsETlwpNrGze42uL30DB [20:53] <tiglionabbit> it works on my mac though, where I compiled ffmpeg using brew [20:55] <pasteeater> tiglionabbit: what ubuntu version? [20:55] <tiglionabbit> oneiric [20:56] <pasteeater> see option C here: http://ubuntuforums.org/showthread.php?t=1117283 [20:56] <pasteeater> (or A if you don't mind compiling) [20:57] <tiglionabbit> also, I noticed that if I just say -acodec aac, it suggests I use libvo_aacenc, but if I do that, this happens: http://friendpaste.com/42H2NmDBxjWq6AzqY9M8gm [20:58] <pasteeater> i suppose it doesn't like 6 channel audio. [20:59] <tiglionabbit> I tried passing -channels 2 but it didn't help [21:00] <pasteeater> the option is -ac, not -channels [21:00] <tiglionabbit> on my mac it's -channels [21:00] <pasteeater> ok [21:00] <tiglionabbit> ok I passed -ac 2 and it segfaulted [21:00] <pasteeater> you're using ffmpeg libav, not ffmpeg from FFmpeg by the way [21:01] <tiglionabbit> ok [21:01] <pasteeater> *ffmpeg from libav [21:01] <tiglionabbit> I'll install the medibuntu one then [21:02] <pasteeater> it shouldn't segfault. i wonder if recent ffmpeg from FFmpeg does the same [21:02] <tiglionabbit> still segfaults on that command [21:03] <tiglionabbit> http://friendpaste.com/Zq4PP8rZgQtilE3zYIuKW [21:03] <tiglionabbit> but with that new package, I now have libfaac [21:03] <tiglionabbit> so I'm good =] [21:04] <pasteeater> alternatively you could copy the audio instead of re-encoding it [21:04] <tiglionabbit> yeah, but I was just using this file as an example. Not all of em have aac audio [21:05] <pasteeater> if you have the time you could follow option A which will install recent ffmpeg, and test the command. if it seg faults i'd like to know. [21:05] <tiglionabbit> ok [21:05] <pasteeater> ...i forgot that guies doesn't install libvo-aacenc, but i can help you with that if you are interested [21:06] <pasteeater> or simply provide a small sample that causes the issue [21:06] <pasteeater> and i'll test [21:06] <tiglionabbit> ok [21:36] <pasteeater> ffmpeg works fine here. i'll blame ubuntu/libav. [21:38] <vadim_> pasteeater: what is the theme? ) [21:39] <pasteeater> the theme? sorry, i don't understand your question. [21:41] <vadim_> pasteeater: you told "ffmpeg works fine here. i'll blame ubuntu/libav."... what are you talking about? :) [21:42] <pasteeater> libvo_aacenc seg faulting when going from 6 channel input to 2 channel output as shown by tiglionabbit about 30 minutes ago [21:43] <pasteeater> libavffmpeg from ubuntu repo seg faulted, but ffmpeg from git does not, so good-enough-for-me [21:44] <vadim_> pasteeater: ah, i see... anyway, just came in, and do not have the entire log... [22:07] <mgeary> hey folks [22:08] <mgeary> i've got a conversion running (batched via a bash script), and it's maxing out the CPU. When i log into the screen session, it appears to be unresponsive. The AAC i'm converting is ~90 minutes. It's been running for ~20 minutes. Is there any chance it's still converting, or do i kill it? [22:11] <mgeary> killed it [22:14] <relaxed> mgeary: you don't redirect ffmpeg's output anywhere? [22:15] <mgeary> just to stdout [22:15] <mgeary> but it stopped updating [22:15] <relaxed> which aac encoder were you using? [22:16] <relaxed> hmm, screen session [22:17] <relaxed> there is a key combo in screen that will freeze the terminal. [22:18] <relaxed> iirc, ctrl+q will fix it [22:19] <tiglionabbit> how do you scale down video without breaking the aspect ratio? Is there a way to set a max height and max width? [22:19] <relaxed> tiglionabbit: -vf scale=720:-1 [22:20] <tiglionabbit> whoa where is that described? [22:20] <relaxed> the man page and website [22:20] <mgeary> relaxed: i just killed the process and my shell script moved on to the next file [22:24] <tiglionabbit> so scale=720 -- what does that do? Do I really have to do something like this example to preserve the aspect ratio but make the video smaller than a certain size? scale='min(500\, iw*3/2):-1' [22:25] <relaxed> scales the width to 720 and the height to whatever is need to maintain the original aspect ratio [22:25] <relaxed> scale=720:-1 [22:25] <tiglionabbit> relaxed: why can't I specify both the max width and the max height? [22:26] <relaxed> that should be possible [22:26] <tiglionabbit> and not scale it up if it's already small [22:27] <mgeary> hrm. another conversion is pegging the CPU and not updating the stdout [22:28] <relaxed> mgeary: can you pastebin the command running and output? [22:28] <mgeary> the command is just: ffmpeg -i foo.aac foo.aac.mp4 [22:29] <mgeary> the output is: http://pastebin.com/0SmmJhC9 [22:31] <relaxed> mgeary: SVN-r26402, ffmpeg hasn't used svn for some time now. [22:31] <relaxed> you should really build the latest version from git along with libvo-aacenc for aac encoding. [22:31] <mgeary> hrm. i just downloaded this and compiled on Feb 10 [22:32] <mgeary> yeah, i'm pretty sure i used git to get it [22:32] <relaxed> compiling an old version from svn still leaves you with an old version [22:33] <mgeary> hrm. maybe i'm confusing my local machine with this one on this server [22:33] <relaxed> my last build shows: git-2012-01-25-247d30a [22:34] <relaxed> tiglionabbit: I think you'll have to script it. [22:51] <tiglionabbit> hm, when I try to use -vf scale options it says Error opening filters! [22:54] <relaxed> tiglionabbit: command? [22:57] <Freakshow> I think I may have asked this before... but [22:57] <tiglionabbit> ffmpeg -i output.mp4 -vf scale=200:10 output-squished.mp4 [22:58] <tiglionabbit> was just trying to test it [22:58] <Freakshow> nm... I'm not going to bring it up... older version [23:00] <mgeary> how big a deal is it if i compile x264 without asm? [23:00] <relaxed> huge [23:01] <mgeary> ok. looks like i need to manually upgrade yasm :S [23:01] <mgeary> this server must be on a pretty old CentOS release :( [23:01] <Freakshow> yeah, that's pretty big [23:02] <relaxed> mgeary: servers usually have old version of everything [23:02] <mgeary> i know, story of my life [23:02] <mgeary> CentOS release 5.7 (Final) [23:02] <mgeary> sigh [23:04] <relaxed> grab the yasm source, compile with the prefix pointed to $HOME/yasm-temp, then: export PATH=$HOME/yasm-temp/bin:$PATH [23:04] <mgeary> already done, though i just compiled to /usr/local [23:10] <pasteeater> mgeary: you might find this use(ful|less) http://ffmpeg.org/trac/ffmpeg/wiki/CentosCompilationGuide [23:10] <mgeary> oh, nice [23:10] <mgeary> i'm "make"-ing ffmpeg now, but if i hit a snag, i'll use that [23:11] <mgeary> i'm realizing i forgot ogg support, but that might be okay [23:12] <pasteeater> libogg is requited if you compile libtheora and libvorbis, IIRC [23:23] <mgeary> hrm. now i'm getting this: [23:23] <mgeary> ffmpeg: error while loading shared libraries: libx264.so.120: cannot open shared object file: No such file or directory [23:23] <mgeary> but it's there in /usr/local/lib/ [23:23] <Mavrik> you're missing x264 and/or you forgot to run sudo ldconfig [23:24] <mgeary> when/where do i run ldconfig? i didn't do that [23:24] <mgeary> i do have x264, but i don't seem to have ldconfig in my path [23:25] <relaxed> sudo /sbin/ldconfig [23:25] <Mavrik> if you're on linux it refreshes the shared library loader database [23:25] <mgeary> what directory should i be in ? ffmpeg's? [23:25] <mgeary> and do i need to recompile after doing that? [23:26] <relaxed> nope [23:26] <mgeary> then something is wrong [23:26] <mgeary> ran it, got no output, still getting that error [23:26] <Mavrik> then your library path is broken [23:26] <mgeary> in my env? [23:27] <mgeary> or in the configure of ffmpeg? [23:27] <Mavrik> in your env [23:27] <relaxed> mgeary: do you have the file /etc/ld.so.conf [23:27] <mgeary> shouldn't my configure have broken? [23:27] <Mavrik> what you're seeing is the shared library loader error [23:27] <Mavrik> not ffmpeg configuration [23:27] <mgeary> relaxed: i do [23:27] <Mavrik> meaning - when loading ffmpeg binary, the shared loader can't find libx264.so.120 [23:28] <Mavrik> do ld.so.conf or LD_LIBRARY_PATH is wrong [23:28] <relaxed> mgeary: sudo echo "/usr/local/lib" >> /etc/ld.so.conf; sudo ldconfig [23:28] <Mavrik> or something else (use ldd to see which libs aren't found) [23:32] <mgeary> bingo. thanks, guys [23:34] <mgeary> hrm. shoot. now i'm getting this: [23:34] <mgeary> http://dpaste.com/704152/ [23:34] <mgeary> "libfaac doesn't support this output format" [23:35] <Mavrik> IIRC libfaac has problems with mono output [23:35] <mgeary> simple fix? [23:36] <Mavrik> switch to stereo. [23:36] <mgeary> -ac 2 ? [23:37] <mgeary> sure enough [23:37] <mgeary> of course, that's bulking up my file unnecessarily, isn't it? [23:38] <Mavrik> yeah but you don't really have much choice :) [23:38] <Mavrik> you can try using internal experimental aac encoder, but that one was pretty broken last time I tried it [23:41] <mgeary> should i worry about "could not find codec parameters" messages? [23:43] <Mavrik> yep. :) [23:44] <mgeary> how do i get the info i need to fix whatever is going on there? [23:46] <Mavrik> it's usually all in the output info [23:46] <Mavrik> just look few lines up :) [23:46] <mgeary> hrm. 0 channels [23:47] <mgeary> ah [23:47] <mgeary> source file is zero byes [23:47] <mgeary> that'll do it, i suppose :) [23:54] <burzum> hi [23:55] <burzum> is there a way to autodetect how many channels (audio files only) a file has and set the output file to have the same channels? [23:56] <Mavrik> that should be done automatically when you don't pass channel number to encoder [23:57] <burzum> ok cool thanks [00:00] --- Fri Feb 17 2012

1 0

ffmpeg-devel.log.20120215
by burek 16 Feb '12

16 Feb '12

[00:18] <ubitux> saste: you were right, i just using ts2str() instead of ts2timestr() in silencedetect [00:23] <CIA-48> ffmpeg: 03Clément BSsch 07master * rd3b06399ff 10ffmpeg/libavfilter/af_silencedetect.c: lavfi/silencedetect: use av_ts2timestr() macro. [00:44] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * re7dbfa59f2 10ffmpeg/libswscale/x86/swscale_mmx.c: [00:44] <CIA-48> ffmpeg: swscale: enable some more SIMD functions. [00:44] <CIA-48> ffmpeg: They no longer just segfault. [00:44] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [00:44] <CIA-48> ffmpeg: 03Paul B Mahol 07master * r83e2e9315a 10ffmpeg/libavcodec/ffv1.c: [00:44] <CIA-48> ffmpeg: ffv1: PIX_FMT_GRAY8 support [00:44] <CIA-48> ffmpeg: Signed-off-by: Paul B Mahol <onemda(a)gmail.com> [00:44] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [00:47] <ubitux> i may have found a way to do a "build only" [00:48] <ubitux> makeopts="FATE=" might do the trick [00:48] <ubitux> since make fate FATE= doesn't run anything [00:50] <ubitux> 'going to give it a try [02:14] <CIA-48> ffmpeg: 03Diego Biurrun 07master * r9d03cb9fc5 10ffmpeg/libswscale/ (swscale_internal.h x86/swscale_mmx.c): swscale: Remove some write-only variables related to alpha handling. [02:14] <CIA-48> ffmpeg: 03Paul B Mahol 07master * r1c77a5307f 10ffmpeg/libavformat/smjpegdec.c: [02:14] <CIA-48> ffmpeg: smjpegdec: implement seeking [02:14] <CIA-48> ffmpeg: Signed-off-by: Paul B Mahol <onemda(a)gmail.com> [02:14] <CIA-48> ffmpeg: Signed-off-by: Diego Biurrun <diego(a)biurrun.de> [02:14] <CIA-48> ffmpeg: 03Martin Storsjö 07master * rdf6050188c 10ffmpeg/libavformat/movenchint.c: [02:14] <CIA-48> ffmpeg: movenc: Use defines instead of hardcoded numbers for RTCP types [02:14] <CIA-48> ffmpeg: Signed-off-by: Martin Storsjö <martin(a)martin.st> [02:14] <CIA-48> ffmpeg: 03Diego Biurrun 07master * rc3b57d6e76 10ffmpeg/ (4 files in 2 dirs): [02:14] <CIA-48> ffmpeg: librtmp: Add "lib" prefix to librtmp URLProtocol declarations. [02:14] <CIA-48> ffmpeg: This allows easily differentiating between both implementations within the build [02:14] <CIA-48> ffmpeg: system and combining the native implementation for plain RTMP with librtmp for [02:14] <CIA-48> ffmpeg: the RTMPE, RTMPS, RTMPT, RTMPTE protocol variants. [02:14] <CIA-48> ffmpeg: 03Janne Grunau 07master * rbf61ef2316 10ffmpeg/libavcodec/rv34.h: [02:14] <CIA-48> ffmpeg: rv34: use uint16_t for RV34DecContext.deblock_coefs [02:14] <CIA-48> ffmpeg: It is used as bitfield with 16 entries. [02:14] <CIA-48> ffmpeg: 03Janne Grunau 07master * r29330721b0 10ffmpeg/libavcodec/rv34.c: [02:14] <CIA-48> ffmpeg: rv34: use AVERROR return values in ff_rv34_decode_frame() [02:14] <CIA-48> ffmpeg: Also adds an error message. [02:14] <CIA-48> ffmpeg: 03Janne Grunau 07master * rb54b40b33c 10ffmpeg/libavcodec/rv40.c: rv40: prevent undefined signed overflow in rv40_loop_filter() [02:14] <CIA-48> ffmpeg: 03Janne Grunau 07master * r2bd730010d 10ffmpeg/libavcodec/rv34.c: [02:14] <CIA-48> ffmpeg: rv34: handle size changes during frame multithreading [02:14] <CIA-48> ffmpeg: Factors all context dynamic memory handling to its own functions. [02:17] <Aaron_M> what's the right way to make interlacing info (whether or not a stream is interlaced) available inside a codec's parser? [06:05] <red_wolf> hi. I want to try the ass filter, but while "./configure ... --enable-libass ..." shows ass in "Enabled filters", "./ffmpeg -filters" does not list ass in its supported filters. why? [06:08] <relaxed> red_wolf: user questions belong in #ffmpeg [06:08] <red_wolf> sorry. [10:27] <CIA-48> ffmpeg: 03Clément BSsch 07master * r95317821ea 10ffmpeg/libavutil/audioconvert.c: audioconvert: consistent use of FF_ARRAY_ELEMS for channel_layout_map. [10:27] <CIA-48> ffmpeg: 03Clément BSsch 07master * rcba4e2cbbc 10ffmpeg/libavfilter/af_pan.c: [10:27] <CIA-48> ffmpeg: pan: fix uninitialized channel_id variable. [10:27] <CIA-48> ffmpeg: Fix broken parsing with pan=2:FL=FR:FR=FL and similar. [12:10] <ubitux> devices are handled exactly the same as formats? [12:11] <ubitux> found an interesting doxy, nevermind [13:36] <ubitux> ### Compiler Error in file /home/ux/fate/ffmpeg/libavfilter/libmpcodecs/vf_fspp.c during Code_Expansion phase: [13:36] <ubitux> ### ASM constraint <o> not supported [13:36] <ubitux> mmh. [13:36] <ubitux> seems vf_fspp is back again. [13:37] <ubitux> michaelni: i added an open64 instance [13:37] <ubitux> but seems to fail as states the message above [16:25] <Compn> ffmpeg -i http://93.91.54.5:8001 -acodec copy -vcodec copy blah.ts [16:25] <Compn> that work for anyone ? [16:28] <Compn> i get [mpegts @ 0248AB00] Application provided invalid, non monotonically increasing dts to muxer in strea [16:28] <Compn> but i'm also using older ffmpeg [16:34] <^sandro^Zzz> hey has anyone succeeded in grabbing a udp://:port input using juts x264 ? [17:14] <Compn> anyone remember how to fix the "non monotonically increasing dts to muxer" problem ? [17:30] <michaelni> ubitux, does it work if you disable libmpcodecs or libavfilter ? [17:32] <ubitux> how can i disable mpcodecs? [17:33] <ubitux> oh, disable-filter=mp should do. [17:33] <Compn> ubitux : what asm is it using in open64 ? [17:34] <Compn> why not yasm/nasm ? [17:34] <ubitux> it's inline asm, i don't know what's the assembler [17:34] <Compn> the assembler is failing [17:35] <Compn> what version, why not use yasm/nasm [17:35] <ubitux> can yasm/nasm be used for inline asm? [17:35] <Compn> oh, i dunno :D [17:35] <ubitux> :) [17:36] <Compn> ubitux : btw someone said --disable-optimizations build was broken, want to add a autobuild for that ? [17:36] <ubitux> sure, ok [17:38] <ubitux> added [17:55] <ubitux> Compn: 'seems it worked [17:57] <^sandro^Zzz> can anyone help me [17:57] <^sandro^Zzz> with ffmpeg [17:57] <^sandro^Zzz> [mp2 @ 0x2356360] Header missing [17:57] <^sandro^Zzz> Error while decoding stream #0:3 [17:57] <^sandro^Zzz> [udp @ 0x23078e0] circular_buffer: OVERRUN [17:57] <^sandro^Zzz> PES packet size mismatch.0 size= 46kB time=00:00:00.43 bitrate= 877.3kbits/s dup=1304 drop=0 [17:57] <^sandro^Zzz> [mpegts @ 0x23073a0] PES packet size mismatch [17:59] <ubitux> arg i don't know what's wrong with open64, it gets killed for random reasons... [17:59] <ubitux> and i don't see any particular reason for [18:07] <ubitux> line 110: Warning: Variable atom in fucntion read_kuki_chunk might be used uninitialized [18:08] <ubitux> nice, a compiler with typo in error messages [18:08] <ubitux> :D [18:13] <Compn> ubitux : guess we dont need to autobuild it then, thanks for testing :) [18:13] <Compn> lol [18:13] <ubitux> well, it doesn't hurt [18:31] <michaelni> ubitux, if all else fails you could try disable asm or optimizations for open64 [19:14] <ubitux> michaelni: i think the main issue now is that it gets killed randomly (sig9) [19:14] <ubitux> the memory usage isn't heavy at all, no swap, nothing in the logs [19:14] <ubitux> i have no idea what's happening [20:00] <^sandro^Zzz> hello everyone [20:00] <^sandro^Zzz> anyone have any idea how to capture a UDP sat channel and output to .ts [20:01] <^sandro^Zzz> im doing ffmpeg -i udp://@:5000 but [20:01] <^sandro^Zzz> it keeps on giving me buffer overun errors [20:12] <CIA-48> ffmpeg: 03Nicolas George 07master * rbd5080b1b0 10ffmpeg/libavfilter/af_pan.c: af_pan: comment a tricky piece of code. [21:39] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * r60991ad6ae 10ffmpeg/ffmpeg.c: [21:39] <CIA-48> ffmpeg: ffmpeg: Fix image allocation. [21:39] <CIA-48> ffmpeg: This probably fixes some of the use of uninitialized issues valgrind shows in fate. [21:39] <CIA-48> ffmpeg: It might also fix other issues. [21:39] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [21:39] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * r25893ad6c9 10ffmpeg/libavcodec/ffv1.c: [21:39] <CIA-48> ffmpeg: ffv1: Warn the user if transparency is stored. [21:39] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [21:47] <pasteeater> beastd: did you ever mail your poster idea to ffmpeg-devel mailing list? [21:48] <beastd> pasteeater: yes, i did. in reply to the CLT announcement. maybe i should have changed the subject to attract more readers. [21:50] <pasteeater> what was the subject? [21:50] <beastd> maybe i'll bump it later. or else just work further on it on my own. somehow many people i cannot really handle sketches ^^ [21:51] <pasteeater> the poster size should be A0 which is 1189x841 mm [21:53] <beastd> oh as you say it. i think i forgot to mention it should be A0 and the minimum recommended resolution [22:05] <^sandro^Zzz> hello.. anyone very familiar with UDP steaming [22:05] <^sandro^Zzz> raw mpeg2 data [22:09] <^sandro^Zzz> http://pastebin.com/ELPhg7VB < ---- anyone know what is wrong.. other encoders seem to be receiving the signal properly.. i think its just ffmpeg on this one.. im stuck for over a week now [22:11] <pasteeater> ^sandro^Zzz: you asked this at #ffmpeg, which is the right place, but you don't need to paste the same question in here too [22:14] <^sandro^Zzz> ok [22:14] <^sandro^Zzz> not sure if it is a developemnt issue [22:14] <^sandro^Zzz> i think its a bug [22:14] <vadim> ^sandro^Zzz: yes, it is... and a quite old one... [22:15] <vadim> dear developers... maybe you will find some time to solve this http://ffmpeg.org/trac/ffmpeg/ticket/609 [22:28] <^sandro^Zzz> did the fifo thing [22:28] <^sandro^Zzz> http://pastebin.com/vrAGvUN8 [22:28] <^sandro^Zzz> but nwo i get that [22:29] <^sandro^Zzz> its complaining about something dup [22:29] <^sandro^Zzz> duplicate frames [22:29] <^sandro^Zzz> or something like that. [22:29] <^sandro^Zzz> thats libx264 [22:29] <^sandro^Zzz> the frames are the same in a udp stream aren't tehy.. being sent from a device [22:31] <vadim> ^sandro^Zzz: I confirm that the workarround described in ticket 609 does not work... any way it's a open bug [22:32] <^sandro^Zzz> well that puts everything on hold [22:33] <^sandro^Zzz> any other way to capture the UDP://@:port traffic and transcode it to h264 [22:33] <^sandro^Zzz> well i want it to .ts file [22:33] <^sandro^Zzz> but at a lower bit rate [22:36] <^sandro^Zzz> is there another capture / encoder that can be sued [22:36] <^sandro^Zzz> that is good and stable ? [22:36] <vadim> try VLC [00:00] --- Thu Feb 16 2012

1 0

ffmpeg.log.20120215
by burek 16 Feb '12

16 Feb '12

[00:06] <jensverwiebe> relaxed: getting late here, will be back tomorrow ( +- 10:00 GMT ) [05:28] <__vincent> Is it possible to catcatanate to a video file only video data [05:28] <__vincent> no header [05:29] <relaxed> sure with something like raw yuv [05:30] <__vincent> i got very corrupt video files [05:30] <__vincent> and they are very important [05:30] <relaxed> mp4? [05:30] <__vincent> yes [05:31] <__vincent> battery run out from digital video camera and it did not finnish writing the header [05:31] <__vincent> moov atom its called [05:32] <__vincent> but i do have working files [05:32] <__vincent> am just trying to copy the header some way [05:32] <relaxed> what does "working" mean? [05:33] <__vincent> they play [05:34] <relaxed> oh, you mean copy the headers from the working file to the corrupt? [05:35] <relaxed> or the corrupt file play? [05:35] <relaxed> files* [05:36] <__vincent> yes [05:36] <__vincent> something like that [05:37] <__vincent> the header may have to be also modfied slightly i guess [05:37] <__vincent> but they are same resolution [05:37] <__vincent> same camera etc [05:59] <relaxed> __vincent: this might work http://www.grauonline.de/cmsimple2_6/en/?Solutions:HD_Video_Repair_Utility [06:09] <relaxed> red_wolf: where did you install ffmpeg? [06:10] <red_wolf> I have not install it. I run it at the place where ffmpeg is built. [06:11] <relaxed> with ./ffmpeg ? [06:12] <__vincent> yes i've seen that [06:12] <__vincent> its commercial [06:12] <red_wolf> yes. should i install it? [06:12] <__vincent> only 50mbs the free version [06:14] <relaxed> __vincent: I've never used it but if it works and the video is important it may be worth purchasing. [06:19] <red_wolf> relaxed: Now I install ffmpeg to /usr/bin, but the result is the same: ass filter does not show. [06:19] <relaxed> with make install? [06:21] <red_wolf> ah. I see it. I installed it with dpkg -i ffmpeg.deb but i forgot to install libavfilter-xx.deb. [06:21] <red_wolf> thank you. [07:28] <buhman> I have about 100G raw PNG frames I want to encode in vp8; will the defaults for the webm container work fine by default without loosing much fidelity, or will I need to do some sort of multi-pass encode? (if possible) [07:29] <buhman> each frame is about 4-8 megabytes [07:31] <zap0> buhman, you really want vp8? [07:33] <buhman> zap0: do I? [07:34] <zap0> its your choice, just trying to figure out if that is a requirement or a suggestion [07:34] <buhman> zap0: suggestion [07:34] <zap0> there is a PNG codec. [07:35] <buhman> what I guess I really want is it to be mostly-lossless video encoding, but I would expect to spend a very long time (multiple weeks if neccessary) calculating crazy amounts of b-frames [07:35] <zap0> which would more-of-less make a movie from your PNG without any loss [07:35] <buhman> bringing the bitrate down to ~<= 10000mbit/s is one goal [07:37] <buhman> basically I want to bring the average bitrate down to close to 10mbit/s, while loosing as little fidelity as possible; even if this means spending a _very_ long time encoding [07:38] <relaxed> h264 sounds like a good fit [07:41] <buhman> isn't that patented weirdly or something? [07:48] <theholyduck> buhman, not really [07:48] <theholyduck> vp8 wont do what you ask, h264 wil [07:48] <theholyduck> buhman, whats your usage scenario? [07:49] <theholyduck> i.e what are you going to do with the video afterwards? [07:49] <theholyduck> buhman, h264 is liscence and royalty free for pretty much any concivable application you might have [07:49] <theholyduck> buhman, and x264 can do the job you're asking [07:50] <theholyduck> also, second question, is it 10mbit/s constant you want (streaming or something) or 10mbit/s average over the entire video? [07:50] <theholyduck> that said, i wouldnt be suprised of x264 could do it at way less than that bitrate [07:50] <buhman> theholyduck: just a vague target [07:50] <theholyduck> buhman, well, what os are you running? [07:51] <buhman> GNU/Linux [07:51] <theholyduck> then you're going to have to build a propper x264 yourself [07:51] <buhman> "proper"? [07:51] <theholyduck> up to date [07:52] <theholyduck> http://ubuntuforums.org/showthread.php?t=786095 can probably be adapted [07:52] <buhman> is there a minimum x264 version/revision I should have? [07:52] <theholyduck> just the latest [07:53] <theholyduck> they keep adding more and more better stuff to it [07:53] <theholyduck> but, once you have a up to date ffmpeg with up to date x264 [07:53] <buhman> so ffmpeg from git too? [07:54] <theholyduck> buhman, just more or less follow the guide i linked [07:54] <theholyduck> if you're not on ubuntu, some of the dependency names might be a bit different [07:54] <theholyduck> -devel instead of -dev [07:54] <theholyduck> and you might have to make install instead of running checkinstall [07:54] <theholyduck> but the basic process is the same [07:56] <theholyduck> something like ffmpeg -vcodec libx264 -preset slow -crf 16 + input and output filenames and stuff [07:56] <theholyduck> will give you a visually lossless h264 file [07:56] <theholyduck> and in most cases, it should be well under 10mbit/s [07:56] <theholyduck> if not, increase the crf nr [08:00] <buhman> I thought I might mention that these frames are 4,096 x 1,744 ; is that an appropriate "crf" in that case? [08:01] <theholyduck> should work [08:01] <theholyduck> crf is a quality number [08:01] <buhman> that's what I assumed [08:02] <theholyduck> but bitrate could be pretty high the [08:02] <theholyduck> n [08:02] <buhman> shouldn't I be double/triple-passing these? [08:02] <theholyduck> naw [08:02] <buhman> what does that accomplish? [08:02] <theholyduck> you doubble pass when you have a target bitrate/filesize [08:02] <theholyduck> crf is a target quality [08:02] <buhman> oh [08:02] <theholyduck> 16 is visually lossless in all circumstances [08:03] <theholyduck> i can never tell crf 16 from the source [08:03] <theholyduck> which might be overkill here [08:03] <theholyduck> you could go for something higher [08:03] <theholyduck> 0 is bit for bit lossless [08:03] <theholyduck> it goes from there to 51 [08:03] <theholyduck> if 16 is too large, you could try 18 and see if its still good enough [08:03] <theholyduck> and then 22, and so on [08:03] <buhman> so why wouldn't I want 0 then? [08:03] <theholyduck> buhman, because bit for bit lossless [08:03] <theholyduck> is huge [08:04] <theholyduck> if you want the file for editing or something [08:04] <theholyduck> then you could bit 0 [08:04] <buhman> naa [08:04] <theholyduck> but its not going to be anywhere NEAR 10mbit/s [08:04] <theholyduck> i doubt 16 is either actually :p [08:04] <theholyduck> depends on content [08:04] <theholyduck> buhman, but basicly, start at 16, encode say a 100 frame sample [08:05] <theholyduck> if you decide you can accept worse visual quality [08:05] <theholyduck> go to 18 [08:05] <theholyduck> and so on [08:05] <theholyduck> until you find your sweetspot [08:05] <buhman> well png compressed them these 4,096x1,744 frames to 4-8mbytes each, so if that's any indication of content.. [08:05] <theholyduck> buhman, well like [08:05] <theholyduck> changes for instance [08:05] <theholyduck> buhman, i mean, i once made a 1080p clip that was a couple of hundred kb large [08:05] <theholyduck> with x264 [08:05] <theholyduck> for 15 seconds [08:05] <theholyduck> in crf 16 [08:05] <buhman> heh; that's awesome [08:06] <theholyduck> because there was basicly no changes [08:06] <theholyduck> and only 256 colurs :P [08:06] <theholyduck> was a openttdcoop recording [08:06] <theholyduck> buhman, basicly, the less changes there are between frames [08:06] <theholyduck> the more of it can be reused [08:06] <theholyduck> x264 is VERY efficient with static content [08:08] <buhman> that was what I was hoping for [08:09] <buhman> (reducing the size of these semi-losslessly by only saving the difference between frames and not the entire frames themselves) [08:36] <buhman> theholyduck: this should be sufficent, yes? http://ix.io/29G [09:04] <theholyduck> buhman, probably [09:45] <jensverwiebe> moin moin [09:46] <jensverwiebe> hi relaxed [09:49] <jensverwiebe> anyone checked in between possible broken ffv1 ( on OSx ) with pix_fmt rgb32 ? [11:09] <buhman> I'm trying to encode from a series of images [11:10] <buhman> when I do "-f image2 -i %08d.png" as suggested in the manpage; I'm told "[image2 @ 0xa1a3aa0] Could not find codec parameters (Video: png)" [11:12] <kriegerod> you must lack decoder of png [11:12] <kriegerod> check `ffmpeg -loglevel debug` for your configure line [11:12] <kriegerod> and recompile adding --enable-decoder=png [11:12] <kriegerod> should help [11:12] <buhman> --prefix=/usr --enable-gpl --enable-libfaac --enable-libmp3lame --enable-libtheora --enable-libvorbis --enable-libx264 --enable-nonfree --enable-postproc --enable-x11grab --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-version3 --enable-libvpx [11:12] <buhman> ok [11:13] <kriegerod> try first ffprobe some-of.png [11:13] <kriegerod> to check my idea [11:14] <buhman> Stream #0:0: Video: png, rgb24, 4096x1744, 25 tbr, 25 tbn, 25 tbc [11:15] <kriegerod> hm, then i waas wrong [11:16] <kriegerod> give us full cmdline [11:17] <buhman> ffmpeg -f image2 -i %08d.png -i sintel-master-51.flac -r 24 -s 4096x1744 -vcodec libx264 -acodec flac -preset veryslow -crf 16 -threads 0 sintel.mkv [11:18] <jensverwiebe> if i do: ./ffmpeg -pix_fmts , shouldnŽt it show also rgb32 fmt ? [11:19] <buhman> it doesn't say rgb32 here either [11:19] <buhman> but that's only rgb24 anyway [11:20] <jensverwiebe> buhman: ah, ty, wonder if this is the reason for my corrupted output [11:20] <buhman> I think "rgb32" is rgba/argb [11:21] <buhman> it says it's 4 bytes per sample which makes sense... [11:21] <jensverwiebe> buhman: yes, but -pix_fmt rgb32 should be a valid flag, no ? [11:21] <buhman> well not if rgba is the same thing [11:22] <jensverwiebe> buhman: hmm, k, must lookup source then, in versions < 0.10 rgb32 worked with ffv1 [11:23] Action: jensverwiebe brb, coffee, zif ;) [11:26] <jensverwiebe> bk [11:29] <jensverwiebe> hmmm [11:29] <buhman> jensverwiebe: recompilingwith --enable-decoder=png didn't change anything [11:30] <jensverwiebe> buhman: ;) [11:30] <jensverwiebe> kriegerod: buhman: jensverwiebe: recompilingwith --enable-decoder=png didn't change anything [11:31] <jensverwiebe> buhman: iám still worried why: ./ffmpeg -i /Volumes/SystemHD/Users/jensverwiebe/Downloads/ffmpeg_input_rgb32.mov -pix_fmt rgb32 -vcodec ffv1 -an output.mo [11:31] <jensverwiebe> worked in 0.87 but not in 0.10 [11:31] <jensverwiebe> donŽt see a changed define etc. [11:45] <kriegerod> buhman: i tried with my files, succeed. Post your configure string, version, sample files, uncut command output with full debug on ffmpeg-user maillist [11:48] <buhman> a mailing-list worthy question? [11:48] <buhman> that seems like an accomplishment [11:50] <kriegerod> could somebody explain why this code fails when FILTER_STRING is not "scale", but just "fifo" or "copy"? The point is that pixel format conversion takes place, but i don't understand why avfilter doesnt sort out it automatically. https://gist.github.com/1834778 [11:54] <buhman> kriegerod: heh; I just figured it out: 00000000.png was empty [11:55] <jensverwiebe> k, found culprit in ffv1, commenting out s->transparency= 1; for rgb32 fixes output, must find now what breaks there [11:55] <Doxin> I recorded a video at 1fps, but i'd like the playback to be at 30fps, is there any easy way to acomplish this? [11:58] <buhman> yey; I was trying to encode with the same options I had above [11:59] <buhman> it failed after a while saying "x264 [error]: malloc of size 30085632 failed" [11:59] <buhman> erm that machine only has 2G ram [12:00] <buhman> how can I make ffmpeg swap or something [12:00] <buhman> I don't care how slow it is; I just want it done _eventually_ [12:09] <jensverwiebe> oki, pinned down ffv1 garbage with rgb32 to transparency handling [12:16] <jensverwiebe> no dev here ? [12:40] <jensverwiebe> any developer here ? [12:57] <jensverwiebe> buhman: did you try compile with : --enable-memalign-hack [13:38] <Doxin> I'm trying to change the framerate of a video file. i don't want to drop frames, i dont want to add frames. just make it run faster. but when i try this ffmpeg downsamples and adds artifacts like a surreal painter with tourettes . how do i change the framerate without re-encoding the whole file? [13:41] <renderTom> what is the format of your source video? [13:42] <Doxin> renderTom: ogv [13:42] <renderTom> sorry I am not familiar with this format? Is it RAW video or compressed? [13:43] <Doxin> renderTom: probably compressed. [13:43] <renderTom> any ways, I am pretty sure that you can change FPS to ONLY files that are in RAW format [13:43] <Doxin> renderTom: and why is that? [13:43] <renderTom> like ANIMATION or PROREZ formats [13:43] <Tjoppen> use -r [13:43] <Tjoppen> and decent encoder settings [13:44] <Doxin> Tjoppen: how? how do i tell ffmpeg to not change compression settings and such? [13:44] <renderTom> becouse compressed videos have certain GOP (Group of Pictures): IPPBPPI and so on [13:44] <Tjoppen> oh, you want to remux and change framerate? hm, not sure if that's possible. depends on the container [13:44] <renderTom> only files with intra I frames can change FPS without compresion [13:45] <Tjoppen> renderTom: stop spewing nonsense [13:45] <Doxin> renderTom: so? i just want the video player to play the frames faster. as in, it's 1fps now. and it needs to run 30x faster. [13:45] <renderTom> -r will output file with FPS you set, like -r 25 [13:45] <Tjoppen> a sufficiently smart tool can change the relevant pieces of timing information [13:45] <Doxin> renderTom: it will also duplicate frames [13:45] <Doxin> renderTom: which i don't want [13:45] <renderTom> -r will definitely duplicate frames [13:46] <Doxin> i DO NOT want more frames [13:46] <Doxin> i want the existing frames shown shorter. [13:46] <Tjoppen> try something like "ffmpeg -r 25 in.ogv [encoder settings] out.whatever" [13:46] <Tjoppen> err, -r 25 -i in.ogv [13:46] <Tjoppen> instead of -r after -i. that used to work, not sure if that's still the case [13:47] <renderTom> -r 25 -i INPUT FILE will not work if input file is compressed [13:47] <Doxin> Tjoppen: Option framerate not found. [13:47] <Tjoppen> hm, right. there might be a filter to do this though [13:47] <Doxin> why the fudge is this so damn hard? only one damn number needs to be changed in the file. [13:47] <renderTom> that's because you are using compreesed video, that's why you get Option framerate not found. [13:48] <Tjoppen> like -vf settb [13:51] <Tjoppen> Doxin: try adding -vf settb=1/25 [13:54] <Doxin> Tjoppen: mkay [13:54] <Doxin> Tjoppen: before or after the input file? [13:54] <Tjoppen> after [13:55] <Doxin> seems to work [13:55] <Doxin> still duping frames, but *way* less than last time :P [13:58] <Doxin> Tjoppen: still running at 1fps tough :I [14:07] <Tjoppen> *shrugs* [14:07] <Tjoppen> maybe -r 25 as well? [14:08] <Tjoppen> or -vf settb=1 -r 25 [14:08] <Tjoppen> or extract the raw video essence to a file, the do -r before -i on it [14:09] <Tjoppen> what video codec is it? [14:15] <Doxin> Tjoppen: i have *no* clue [14:22] <Doxin> Tjoppen: still running at 1fps [14:26] <jensverwiebe> michaelni: if you have 5 minutes for me , would be nice any time today [14:34] <Tjoppen> Doxin: I'm out of ideas [14:35] <Doxin> well damn. [14:35] <Doxin> Tjoppen: thanks for the help anyways :P [15:35] <Freakshow> any idea of who maintains the freebsd port? [16:31] <pron> Freakshow: http://www.freshports.org/multimedia/ffmpeg/ , Maintained by: mm(a)FreeBSD.org [16:33] <^sandro^Zzz> hey has anyone succeeded in grabbing a udp://:port input using juts x264 ? [16:48] <burek> me [16:49] <burek> ^sandro^Zzz, what's the problem [17:05] <CodeH3x> Hi everyone [17:06] <CodeH3x> Does anyone know how to convert mkv with (.ass) subtitles inside the track to .mp4? [17:07] <CodeH3x> ffmpeg -scodec copy -b 196k -i video.mkv outvid.mp4 does not work [17:08] <hi117> >ASS are not widely supported in mp4 [17:10] <CodeH3x> yeah, i can't find any solution online either [17:10] <hi117> you can put it in there somehow, just dont expect it to play back [17:11] <relaxed> CodeH3x: there is an ass filter for ffmpeg if you want hardsubs. [17:12] <hi117> but its hardsubs, youll lose quality because of reencode [17:12] <hi117> also mp4 might support SRT subtitles [17:13] <CodeH3x> it's okay if i lose some quality [17:14] <hi117> ya, then hardsub reencode [17:14] <CodeH3x> how do i enable or install hardsub re encode? [17:15] <relaxed> it would be easier with mplayer [17:16] <CodeH3x> but if i use hardsub reencoding then i'll need to extra the .ass from the mkv file first, then hardsub it into mp4 ... [17:17] <hi117> go to #mplayer and ask, mencoder is better with hardsubbing [17:17] <CodeH3x> Oh ok, thanks hi117 [17:17] <hi117> np [17:45] <drno_> Howdy again folks. [17:45] <drno_> Been playing around with RTMP streaming, and have solved most of my problems. Found something new, that others have seemed to encounter as well. [17:45] <drno_> Using the segment code added on 12/19, I can't split a file. I just get the [17:46] <drno_> Output file #0 does not contain any stream [17:46] <drno_> error message... has anyone else ran into that using the segment muxer, and if so, hvae you found a workaround? I'm trying to download a release from 12/12 (before the 12/19 change) and test. [17:58] <CodeH3x> hey guys [17:58] <CodeH3x> i'm getting an error installing libass [17:58] <CodeH3x> fribidi-0.10.9-1.el6.rf.x86_64 [17:58] <CodeH3x> checking for FRIBIDI... configure: error: Package requirements (fribidi >= 0.19.0) were not met: [17:58] <CodeH3x> No package 'fribidi' found [17:59] <CodeH3x> ops sorry, i guess i need fribidi 0.19.0 [17:59] <sacarasc> Or higher. [17:59] <CodeH3x> i should read the error messages carefully [17:59] <CodeH3x> do you know where i can find it? centos dag installs 0.10 [18:04] <CodeH3x> i'm getting an error uncompressing this [18:04] <CodeH3x> gzip: stdin: not in gzip format [18:04] <CodeH3x> tar: Child returned status 1 [18:04] <CodeH3x> tar: Error is not recoverable: exiting now [18:05] <CodeH3x> tar zxf fribidi-0.19.2-2-x86_64.pkg.tar.xz [18:05] <CodeH3x> tar zxvf fribidi-0.19.2-2-x86_64.pkg.tar.xz [18:05] <CodeH3x> does anyone know how to uncompress this? [18:10] <ubitux> with J flag [18:10] <ubitux> tar xf should work anyway [18:13] <CodeH3x> i'm sorry guys, the .xz was for slackware and i'm using centos [18:13] <CodeH3x> that's why it wasn't working [18:22] <CodeH3x> I just installed fribidi-0.19.2 yet still i'm getting an error when install libass [18:22] <CodeH3x> checking for FRIBIDI... configure: error: Package requirements (fribidi >= 0.19.0) were not met: [18:22] <CodeH3x> is it because i have 0.19.2 ..? [18:23] <CodeH3x> well nope that's not it [18:23] <CodeH3x> although i have it install it doesn't show up in rpm -qa fribidi [18:23] <CodeH3x> when >which fribidi [18:24] <CodeH3x> "/usr/local/bin/fribidi [18:29] <CodeH3x> i installed fribidi using the rpm package [18:29] <CodeH3x> rpm -qa fribidi [18:30] <CodeH3x> fribidi-0.19.2-2.puias6.x86_64 [18:30] <CodeH3x> yet still i'm getting an error when configuring libass [18:30] <CodeH3x> any ideas? [18:31] <hexmaster> Hi was going to change the sound format of a video. Whats wrong with this cmd? [18:31] <hexmaster> [matroska @ 0x35d9260] Can't write packet with unknown timestamp [18:31] <hexmaster> av_interleaved_write_frame(): Invalid argument [18:31] <hexmaster> ffmpeg -i American.Graffiti.avi -vcodec copy -acodec libfaac -ab 128k -ac 2 -sn American.Graffiti.mkv [18:32] <hexmaster> Its as basic as it gets. I am copying the video and and changing the sound and container. Also dumping into mkv. [18:32] <hexmaster> ffmpeg version N-32984-g950930b [18:33] <hexmaster> about 10 days old or so. [18:33] <CodeH3x> use this [18:33] <CodeH3x> : ffmpeg -i <filename>.avi -vcodec ffv1 -acodec pcm_s16le <filename>.mkv [18:33] <CodeH3x> http://archivematica.org/wiki/index.php?title=AVI_to_MKV_using_FFmpeg [18:41] <hexmaster> CodeH3x There are a bunch of streams that I am dumping with the -sn cmd. Also the speakers my media player (not a PC) are junky little 1.5 watt things. What format is pcm_s16le going to convert this to. Normally to convert stuff I just do a ffmpeg -i in.avi -acodec copy -vcodec copy out.mkv . Have things changed? [18:47] <hexmaster> maybe just a bad day I pulled the git. Ill recompile and see if it works. [19:00] <Zilly> teratorn, hey - it was just a quality thing. It encodes the x264 in good quality with no qscale, but I need to set the qscale to around 10 for the ogg to be in good quality. [19:00] <teratorn> Zilly: ah, that's nice to hear [19:03] <Zilly> Is it normal for libtheora to take three times longer to encode than x264? [19:14] <jensverwiebe> michaelni: there ? [19:16] <michaelni> jensverwiebe, yes [19:16] <michaelni> how can i help ? [19:17] <jensverwiebe> michaelni: i found a glitch in ffv1 codec after >= version 0.10 [19:18] <jensverwiebe> michaelni: looks as s->transparency for RGB32 pix_fmt causes troube [19:19] <jensverwiebe> michaelni: thus an export with ffv1 -pix_fmt rgb32 is faulty on windows and OSX veridied [19:19] <jensverwiebe> verified [19:19] <jensverwiebe> michaelni: if i comment out s->transparency= 1; in line 940 ( release) in ffv1.c all is fine [19:20] <jensverwiebe> problem seems to be tiles [19:21] <jensverwiebe> michaelni: before i made blender builds as mainatiner with ffmpeg 0.87 there it was fine , but i saw this feature was intruduced later [19:21] <michaelni> let me check [19:22] <michaelni> on linux -pix_fmt rgb32 seems working [19:22] <jensverwiebe> michaelni: yes [19:23] <jensverwiebe> michaelni: it also works with huffyuv, issue is limited to ffv1 [19:24] <michaelni> so what exactly fails on windows? encoding? decoding? [19:24] <michaelni> encoding with latest ffv1 + decoding with old ? [19:24] <jensverwiebe> michaelni: i can speak for OSX here atm: the output video is garbage only [19:25] <michaelni> ok so OSX [19:25] <jensverwiebe> michaelni: i tested o.10 and actual git [19:25] <jensverwiebe> michaelni: but problem seems on windows too [19:25] <michaelni> what do you use to encode and what to decode, same libavcodec/ffmpeg version ? [19:26] <jensverwiebe> michaelni: not exactly: to watch i use VideoLAn or QucktimePlayer [19:26] <kriegerod> could somebody share URL scheme to give to ffmpeg to publish onto wowza? i have been said that url is rtmp://x.x.x.x/blah/?token=xxxxxxxxxxxxxx and that stream is is blah_blah. I tried rtmp://x.x.x.x/blah/blah_blah/?token=xxxxxxxxxxxxxx and get "Server error: Invalid credentials supplied" [19:27] <jensverwiebe> michaelni: looks as if i use matched versions ( as in blender export and reimport ) this works again [19:27] <michaelni> then its probably that your decoder simply doesnt support the alpha channel [19:27] <michaelni> try 0rgb instead of rgb32 on the encoider side [19:27] <jensverwiebe> michaelni: if you say you had breaking changes thats o.k. for me [19:27] <michaelni> this should work with old decoders [19:28] <jensverwiebe> michaelni: had problems to use =rgb even with newest git [19:28] <jensverwiebe> 0rgb [19:29] <michaelni> problems ? [19:29] <jensverwiebe> michaelni: ffmpeg did not accept 0rgb as flag [19:29] <michaelni> -vcodec ffv1 -pix_fmt 0rgb [19:29] <michaelni> works here [19:29] <jensverwiebe> michaelni: sec, lemme recheck [19:31] <jensverwiebe> michaelni: ah, stupid me, did 0rgb32 [19:31] <jensverwiebe> michaelni: seems o.k. [19:32] <jensverwiebe> michaelni: o.k, so it was just a breaking change [19:32] <jensverwiebe> thx [19:32] <michaelni> np [19:32] <jensverwiebe> michaelni: cya, we have release today ;) [19:34] <michaelni> jensverwiebe, cya and good luck with your release! [19:35] <jensverwiebe> michaelni: thanks, i push atm guys to remove QTKit and relay complete on ffmpeg ;) [19:35] <michaelni> :) [19:35] <jensverwiebe> with your latest additions will be nice to satisfy all needs :9 thx for great work on ffmpeg ( all guys indeed ) [19:41] <jensverwiebe> michaelni: btw, ic now: was misguided by case PIX_FMT_0RGB32, so set wrong flag [19:59] <Freakshow> member:kriegerod: how are you authenticating on the wms side? I'm presuming ?token= is a passwd hash of some kind? [19:59] <Freakshow> kriegerod: see above... stupid c&p [20:17] <smj> does ffmpeg do interpolation/motionblur when reducing FPS? [20:19] <smj> to prevent choppy motion and this kind of effects [20:19] <smj> http://www.youtube.com/watch?v=jQDjJRYmeWg [20:19] <CodeH3x> Does anyone know where I can find fribidi-0.19.0 or higher for CentOS 6 64bit? [20:20] <CodeH3x> having a hard time here... unfortunately [20:33] <Mavrik> smj, no. [20:43] <smj> is there any plans? [20:47] <kriegerod> for my issue, answer is to use librtmp, and url rtmp://x.x.x.x/blah/?token=xxxxxxxxxxxxxx/channel_name [21:30] <KyranBe> Hi [21:30] <KyranBe> I'm trying to mux in subtitles into an avi file [21:31] <KyranBe> but it always fails with a message saying "Too large number of skiped frames" (sic) [21:38] <magic_1> hi guys, not sure if any of you can help me, after re-installing my ffmpeg (using the following link http://ubuntuforums.org/showpost.php?p=9868359&postcount=1289) i get the following error PHP Warning: PHP Startup: Unable to load dynamic library '/usr/lib64/php5/20090626/ffmpeg.so' [21:40] <pasteeater> magic_1: are you trying to use php-ffmpeg? [21:41] <CodeH3x> i don't see ffmpeg-php being installed in that tutorial [21:41] <CodeH3x> magic_1: seems like you don't have ffmpeg-php installed [21:41] <pasteeater> it's not, and it's not a FFmpeg project and is not supported here. [21:41] <pasteeater> easiest method is to use php5-ffmpeg from the repository [21:42] <CodeH3x> Oh i see, i thought ffmpeg-php was part of ffmpeg [21:42] <pasteeater> that's a common misconception [21:42] <CodeH3x> Oh, i guess it's because they are using ffmpeg in ffmpeg-php [21:43] <CodeH3x> A quick question, has anyone install mkvtoolnix before? [21:44] <pasteeater> looks like someone forked it and is actually maintaining it: https://github.com/char0n/ffmpeg-php [21:44] <pasteeater> CodeH3x: what distro? [21:55] <magic_1> yep [21:55] <magic_1> apologies pasteeater: was making some coffee [21:57] <magic_1> and CodeH3x [22:03] <^sandro^Zzz> Hello, I am looking for someoen who is very familiar with ffmpeg and UDP streaming [22:06] <vadim> ^sandro^Zzz: let me know if you find someone... [22:09] <^sandro^Zzz> http://pastebin.com/ELPhg7VB < ---- anyone know what is wrong.. other encoders seem to be receiving the signal properly.. i think its just ffmpeg on this one.. im stuck for over a week now [22:11] <vadim> ^sandro^Zzz: do you mean the 'circular_buffer: OVERRUN' error?.. [22:13] <^sandro^Zzz> yes [22:13] <^sandro^Zzz> it wont play.. gives me errors.. not sure why [22:16] <vadim> ^sandro^Zzz: this is a reported bug, and a quite old one... http://ffmpeg.org/trac/ffmpeg/ticket/609 [22:30] <magic_1> hehe sorted it out [22:31] <magic_1> seems sudo apt-get install php5-ffmpeg [22:31] <magic_1> sorted the issue out [22:31] <magic_1> how about that hehe [00:00] --- Thu Feb 16 2012

1 0

ffmpeg-devel.log.20120214
by burek 15 Feb '12

15 Feb '12

[00:12] <CIA-48> ffmpeg: 03Clément BSsch 07master * rc214e98634 10ffmpeg/ffmpeg.c: [00:12] <CIA-48> ffmpeg: ffmpeg: remove pointless avcodec_get_frame_defaults(). [00:12] <CIA-48> ffmpeg: filtered_frame is overwritten just below anyway. [00:13] <CIA-48> ffmpeg: 03Clément BSsch 07master * rb986590810 10ffmpeg/ffmpeg.c: [00:13] <CIA-48> ffmpeg: ffmpeg: remove unneeded ost->picref check. [00:13] <CIA-48> ffmpeg: If there ost->picref is NULL, it will likely crash anyway a few lines [00:13] <CIA-48> ffmpeg: below by dereferencing it in order to access video attribute. [00:13] <CIA-48> ffmpeg: 03Clément BSsch 07master * r73413e6bca 10ffmpeg/ffmpeg.c: ffmpeg: raise ENOMEM on avfilter_graph_alloc() failure. [00:58] <ubitux> first run with undef done [01:01] <ubitux> i updated the configs to use --valgrind, i hope it will work; we should be fixed in a few hours [01:45] <durandal_1707> michaelni: should avcodec_align_dimension2() have entries for yuva fmts? [01:45] <durandal_1707> oh, yet another bug [01:46] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * rdda713f870 10ffmpeg/libavcodec/iff.c: [01:46] <CIA-48> ffmpeg: iff: Fix decode_ham_plane32() buf_size. [01:46] <CIA-48> ffmpeg: The wrong variable was passed into decode_ham_plane32() [01:46] <CIA-48> ffmpeg: Fixes: Ticket922 [01:46] <CIA-48> ffmpeg: Reviewed-by: Paul B Mahol <onemda(a)gmail.com> [01:46] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [01:46] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * r387783749f 10ffmpeg/libavcodec/jpegls.c: [01:46] <CIA-48> ffmpeg: jpeglsdec: fix bpp & limit. [01:46] <CIA-48> ffmpeg: Fixes: Tikcet969 [01:46] <CIA-48> ffmpeg: Thanks-to: ITU for the freely available spec. [01:46] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [01:46] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * r00ab9cdae1 10ffmpeg/libavcodec/jpeglsdec.c: [01:46] <CIA-48> ffmpeg: jpeglsdec: Prevent out of array write. [01:46] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [01:57] <michaelni> durandal_1707, yes [01:58] <durandal_1707> michaelni: patch already send [01:58] <michaelni> yep just saw it a second after saying yes :) [02:00] <TheFluff> the input to a bitstream filter must be in an aligned and properly padded buffer, right? [02:18] <CIA-48> ffmpeg: 03Diego Biurrun 07master * rd1c4ec7f5e 10ffmpeg/ (configure libavformat/allformats.c): build: Set correct dependencies for rtmp* protocols implemented by librtmp. [02:18] <CIA-48> ffmpeg: 03Diego Biurrun 07master * r75f11901b5 10ffmpeg/libavcodec/ps2/ (dsputil_mmi.c mmi.h): [02:18] <CIA-48> ffmpeg: mips: Move MMI function declarations to a header. [02:18] <CIA-48> ffmpeg: This fixes compilation with -Werror=missing-prototypes. [02:18] <CIA-48> ffmpeg: 03Diego Biurrun 07master * r0bf184e59c 10ffmpeg/libavcodec/ (8 files in 3 dirs): [02:18] <CIA-48> ffmpeg: Move PS2 MMI code below the mips subdirectory, where it belongs. [02:18] <CIA-48> ffmpeg: Also give a more suitable name to the MMI-optimized IDCT; [02:18] <CIA-48> ffmpeg: it is not PS2-specific, as the name currently suggests. [02:18] <CIA-48> ffmpeg: 03Ronald S. Bultje 07master * rd63b7d8c37 10ffmpeg/libswscale/output.c: swscale: fix incorrect chroma bias in yuv2rgb48_1_c(). [02:18] <CIA-48> ffmpeg: 03Ronald S. Bultje 07master * rc0b87359b2 10ffmpeg/libswscale/ (output.c swscale.c): [02:18] <CIA-48> ffmpeg: swscale: handle gray16 as a "planar" YUV format (Y-only, of course). [02:18] <CIA-48> ffmpeg: This allows removing any gray16-specific code, which is essentially [02:18] <CIA-48> ffmpeg: identical to the per-plane code in yuv2plane*(). [02:19] <CIA-48> ffmpeg: 03Ronald S. Bultje 07master * r3cac475655 10ffmpeg/libswscale/output.c: swscale: move YUV2PACKED16WRAPPER() macro down to where it is used. [02:19] <CIA-48> ffmpeg: 03Martin Storsjö 07master * rddf422803b 10ffmpeg/libavformat/mpegtsenc.c: [02:19] <CIA-48> ffmpeg: mpegtsenc: Add an AVOption for forcing a new PAT/PMT/SDT to be written [02:19] <CIA-48> ffmpeg: When segmenting the output from the mpegts muxer, one can [02:19] <CIA-48> ffmpeg: now set this option when cutting to a new segment, to make sure [02:19] <CIA-48> ffmpeg: the next segment starts with PAT/PMT/SDT. [02:19] <CIA-48> ffmpeg: Signed-off-by: Martin Storsjö <martin(a)martin.st> [02:19] <CIA-48> ffmpeg: 03Martin Storsjö 07master * r5d561514b7 10ffmpeg/libavformat/movenc.c: [02:20] <^sandro^Zzz> hello everyone [02:20] <^sandro^Zzz> having a problem with ffmpeg [02:20] <^sandro^Zzz> and udp capture [02:21] <^sandro^Zzz> anyone see an error liek this : [libx264 @ 0x1781bc0] Input picture width (704) is greater than stride (0) [02:21] <^sandro^Zzz> Video encoding failed [02:22] <durandal_1707> never seen that one, how to reproduce? [02:22] <^sandro^Zzz> i have been stuck on it for 3 days [02:22] <^sandro^Zzz> http://pastebin.com/7azatTzc [02:24] <durandal_1707> that is message from x264 [02:25] <^sandro^Zzz> yup. [02:25] <^sandro^Zzz> looks that way [02:29] <durandal_1707> x264-snapshot-20080222-2245 ??? [02:33] <pasteeater> ^sandro^Zzz: are you sure you want "-f mpeg2video" there? [02:34] <^sandro^Zzz> how do you check x264 version? [02:34] <^sandro^Zzz> i even tried very old ones :( [02:34] <pasteeater> x264 --help [02:35] <^sandro^Zzz> root@bit14:/mnt/encoder/segmenter# x264 -help [02:35] <^sandro^Zzz> x264 core:58 svn-736 [02:36] <durandal_1707> ^sandro^Zzz: why you are using so old version, you will get only problems using such old version with latest ffmpeg (assuming both are build with shared libs, if they are static you are wasting time using older binaries....) [02:37] <pasteeater> 1. user questions are usually asked at #ffmpeg, 2. only recent ffmpeg/x264 is supported. [02:37] <^sandro^Zzz> i hear you. i tried the new ones too [02:37] <michaelni> TheFluff, i suspect it would probably work without [02:37] <^sandro^Zzz> same errors [02:39] <pasteeater> and i doubt ffmpeg 0.10 is actually using x264 from 2008 as it requires >= 118 in configure [02:41] <^sandro^Zzz> ok im going to attempt the latest version again [02:41] <^sandro^Zzz> to see what happens.. i have a feeling x264 bug [02:41] <^sandro^Zzz> or ffmpeg latest bug [02:42] <durandal_1707> ^sandro^Zzz: is your ffmpeg static or shared? [02:43] <^sandro^Zzz> um.. well i did enable shared [02:43] <^sandro^Zzz> same as h264 [02:43] <^sandro^Zzz> what's the diff ? [02:44] <^sandro^Zzz> so static has all libraries [02:44] <^sandro^Zzz> and shared in few libraries [02:44] <^sandro^Zzz> so it should be compiled static [02:44] <TheFluff> michaelni: okay, thanks [02:45] <TheFluff> I made it aligned and padded anyway just to be safe [03:01] <durandal_1707> michaelni: you merged too little from qatar [03:23] <michaelni> durandal_1707, ? [03:23] <durandal_1707> michaelni: you merged bugs [03:24] <durandal_1707> now i need to wait one day for that bugs to get fixed :( [03:26] <CIA-48> ffmpeg: 03Paul B Mahol 07master * r148d1f4799 10ffmpeg/libavcodec/utils.c: [03:26] <CIA-48> ffmpeg: lavc: add missing PIX_FMT_YUVA444P [03:26] <CIA-48> ffmpeg: Signed-off-by: Paul B Mahol <onemda(a)gmail.com> [03:26] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [03:27] <michaelni> durandal_1707, next merge done, just waiting for compile & fate [03:27] <^sandro^Zzz> can somone help me with this error [03:27] <^sandro^Zzz> http://pastebin.com/sWsJQCne [03:27] <^sandro^Zzz> not sure what i have done wrong there [03:28] <durandal_1707> for start, finaly start explaining what you are doing [03:30] <durandal_1707> michaelni: are better wait, stuff is getting ugly [03:32] <michaelni> durandal_1707, thats not a english sentance, what do you mean ? [03:34] <Compn> michaelni : durandal_1707 means that qatar is multiplying bugs upon bugs and you should wait before merging more of them [03:34] <Compn> and that you already merged some bug [03:35] <Compn> which means you probably need to read libav-cvslog or libav-devel to see what the bug is [03:35] <michaelni> libav is adding bugs since a year, no news here [03:35] <Compn> lol [03:35] <durandal_1707> we should probably stop auto merging but there is not enough man-power [03:36] <michaelni> durandal_1707, can you elaborate on what bug is added by the last 2 commits? [03:36] <michaelni> i didnt really look ... [03:37] <Compn> i dont think we auto merge... [03:37] <durandal_1707> there is one patch on ml not yet commited... [03:37] <Compn> michaelni merges manually i thought ? [03:38] <michaelni> i merge & test but of course i am a human and i do miss things sometimes [03:40] <durandal_1707> well if "fix crashes in yuv2yuvX on x86-32" is calling for a additional merge go for it ... [03:41] <michaelni> it sounds like it should be merged so ill do it ... [03:43] <CIA-48> ffmpeg: 03Aneesh Dogra 07master * r186dcbcb50 10ffmpeg/tests/ (fate/image.mak ref/fate/sunraster-8bit_gray-raw): [03:43] <CIA-48> ffmpeg: sunrast: Add fate test for gray8. [03:43] <CIA-48> ffmpeg: Signed-off-by: Janne Grunau <janne-libav(a)jannau.net> [03:43] <CIA-48> ffmpeg: 03Ronald S. Bultje 07master * r771bab7f57 10ffmpeg/libswscale/x86/output.asm: [03:43] <CIA-48> ffmpeg: swscale: fix crashes in yuv2yuvX on x86-32. [03:43] <CIA-48> ffmpeg: They were introduced in an earlier commit that introduced use of named [03:43] <CIA-48> ffmpeg: arguments. One cause was a typo, a second cause appears to be a bug in [03:43] <CIA-48> ffmpeg: x264asm that I work around by not using named arguments. [03:43] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * r1fbd905bbb 10ffmpeg/: [03:43] <CIA-48> ffmpeg: Merge remote-tracking branch 'qatar/master' [03:43] <CIA-48> ffmpeg: * qatar/master: [03:43] <CIA-48> ffmpeg: swscale: fix crashes in yuv2yuvX on x86-32. [03:43] <CIA-48> ffmpeg: sunrast: Add fate test for gray8. [03:43] <CIA-48> ffmpeg: Merged-by: Michael Niedermayer <michaelni(a)gmx.at> [03:52] <durandal_1707> michaelni: why ffmpeg use unscaled converters and libav use bicubic ones? [03:53] <durandal_1707> the mentioned bug happens only in last case [03:53] <michaelni> which exact case ? [03:54] <durandal_1707> with pal8 inputs, try chronomaster samples from fate [03:57] <michaelni> why should bicubic be used when its unscaled ? [03:58] <michaelni> maybe iam missing some detail ... [03:58] <durandal_1707> it is bicubic with mmx2 [03:58] <durandal_1707> as reported by mplayer2 [03:58] <durandal_1707> which one is faster? [04:01] <Compn> it used to be on a scale i thought [04:01] <Compn> the first ones were faster, last one like lanczos was slowest [04:02] <Compn> durandal_1707 : btw you need to use more context in your sentences. its hard to guess what you are talking about half of the time [04:02] <Compn> michaelni needs bugreports, urls, mailing list thread urls, etc [04:02] <Compn> samples, command lines, patches, etc [04:18] <durandal_1707> strange recompiled ffmpeg to latest version and it no longer use unscaled variant [07:24] <ubitux> any idea how to make this work? extra_conf="'--valgrind=valgrind --undef-value-errors=no' --disable-stripping" [07:24] <ubitux> it seems to not like the option split [07:29] <ubitux> 'going back to manual target_exec for this one [07:38] <funman> why mixing ' and " ? [07:39] <ubitux> because "valgrind --undef-value-errors=no" is the --valgrind option, and --disable-stripping is another extra conf option [07:39] <funman> valgrind=valgrind.sh maybe? [07:41] <ubitux> heh indeed that could do the trick [07:42] <ubitux> thanks for the idea [07:51] <trupheenix> can anyone tell me how i can cut a video based on keyframes so that I can send relevant portions of it for video streaming applications? either with ffmpeg or using a simple formula. [07:53] <funman> ffmpeg -i input -t 10 -codec copy output [07:53] <funman> done, 10 seconds of video [07:54] <trupheenix> funman, hmmm... i want to send fragments of the video. [07:54] <trupheenix> funman, so this has to contain the keyframe data. [07:54] <funman> -ss 5 => from 5 to 10 [07:54] <trupheenix> funman, how do i determine this? [07:54] <funman> err, to 15 [07:55] <funman> i'd hope ffmpeg does it itself and cuts on keyframes [08:29] <funman> vlcforce build nb-x86_64-win64 [10:05] <aClam> http://git.videolan.org/?p=ffmpeg.git;a=commit;f=libavcodec/h264_refs.c;h=e… Where is exactly is "issue 2679"? Does it mean issue 2679 on libav issue tracker? http://roundup.libav.org/issue2679 [10:08] <ubitux> aClam: this is from a libav merge, when issue* it's related to the roundup yes, otherwise it should be their bugzilla [10:08] <ubitux> with ffmpeg, we use "ticket", to refer to the trac [10:08] <aClam> Okay, thanks [11:51] <trupheenix> any one give me pointers on how i can make video fragments using ffmpeg? is there a way to do this from my code itself? [11:59] <trupheenix> how can i determine the keyframe interval of a video? [12:06] <CIA-48> ffmpeg: 03Carl Eugen Hoyos 07master * r186923758d 10ffmpeg/libswscale/x86/ (input.asm output.asm): Fix compilation with old yasm that does not support AVX. [15:58] <ubitux> is there any reason not to have a hard dep on libavfilter in ffmpeg? [15:59] <ubitux> so we could use the filters for resampling, or volume stuff [15:59] <Daemon404> nothing NEEDS it afaik [15:59] <Daemon404> (yet) [15:59] <ubitux> well, there is duplicated stuff we could get rid of it [16:00] <ubitux> i think lavfi lacks the audio drifting stuff [16:00] <ubitux> but it seems to be the only missing feature [16:00] <ubitux> and it would simplify and unify a lot of things [17:00] <Zeranoe> I'm running into issues when compiling version 0.10 of FFmpeg. Here is the pastebin: http://pastebin.com/BmRH7HVt Does anyone know a quick fix or is this bug material? [17:02] <ubitux> error: 'wcscpy_instead_use_StringCbCopyW_or_StringCchCopyW' undeclared (first use in this function) [17:03] <Daemon404> does mingw really say this? [17:03] <ubitux> looks like a cpp messup [17:03] <Daemon404> or are upu contaminating it with MSVC headers? [17:03] <Daemon404> you* [17:03] <Zeranoe> Daemon404: This is a pure MinGW-w64 cross toolchain [17:03] <ubitux> it looks like there is a #define strcpy StringCbCopyW or something [17:04] <ubitux> which basically breaks the safety check in ffmpeg [17:04] <ubitux> hi saste! :) [17:04] <saste> hi ubitux [17:05] <ubitux> michaelni: open64, path scale and sunc compiler tests which of these are free (price at least) and can be used on a linux box? [17:05] <ubitux> saste: i have a few questions again for you if you have some time for me [17:06] <saste> ubitux: i can reply now if i have a quick reply for them [17:06] <ubitux> mmh, ok well [17:06] <ubitux> so i rework your ffmpeg af patch [17:06] <ubitux> it works for simple filters such as volume [17:07] <ubitux> but as soon as i play with sample rate or something, everything is messed up [17:07] <Zeranoe> I need to head out here but i will be back later, if anyone can find a fix to this I would really appreciate it. [17:08] <saste> ubitux: which kind of "mess"? crash or something? [17:08] <ubitux> sounds is creepy [17:08] <ubitux> basically like it's trying to play a resampled stuff with the untouched original settings [17:08] <saste> uhm yeah [17:08] <ubitux> this is the C main part, which is a lot similar to the video one: http://pastie.org/3381673 [17:09] <ubitux> i thought i copied every new settings [17:09] <ubitux> but it seems the wrong way of doing it [17:09] <ubitux> (also, i don't see anything in the spirit in your original patch) [17:09] <saste> i'm quite sure it was working at some point, as I had most tests working but with output bitwise differences [17:11] <ubitux> i don't see anything which updates the output stream settings in your patch [17:11] <ubitux> so i believe i'm doing something wrong or just forgot something [17:11] <saste> filtered_frame->pts = av_rescale_q(ref->pts, ist_pts_tb, AV_TIME_BASE_Q); [17:12] <saste> looks wrong [17:12] <saste> rescaling is managed differently in case of audio [17:12] <saste> at least iirc [17:12] <ubitux> how differently? (this is the same as the video) [17:13] <saste> how was it in the original code? [17:13] <saste> also i'm not sure it is correct to set the codec context for each frame [17:14] <ubitux> the original code doesn't seem to touch the frame at all [17:14] <saste> with some codecs you can't change the params on the fly and expect it to work, but maybe we have the same bug in the video path as well [17:14] <ubitux> so i wonder when it is supposed to get teh output settings [17:15] <saste> anyway output params don't change in the filterchain right now, and i suppose we have a normalization mechanism on the input [17:15] <saste> the output params should be available at the end of the filtergraph after the configuration stage [17:15] <saste> and you assume they won't change during the filtering stage [17:17] <ubitux> the original patch only has /* ocodec->sample_rate = ist->output_audio_filter->inputs[0]->sample_rate; */ [17:17] <ubitux> at the end of the configuration stage [17:17] <saste> uhm.. is it commented? [17:17] <ubitux> channel layout seems updated though mmh ok [17:17] <ubitux> saste: yes [17:17] <ubitux> at least according to the patch you sent :P [17:18] <ubitux> i'm going to try to change the settings at this point [17:18] <michaelni> ubitux, i dont know, i thought there are "free" versions available for all 3 [17:18] <saste> well i had some private communication with mina with regards to that [17:18] <michaelni> also see http://en.wikipedia.org/wiki/List_of_compilers [17:18] <saste> just lemme dig it in the archive... [17:18] <ubitux> michaelni, saste: ok, thanks :) [17:18] <michaelni> there might be more there that might be able to compile ffmpeg [17:20] <ubitux> btw, while both of you are here, michaelni, saste: is there any reason not to make a harddep on libavfilter in ffmpeg? [17:21] <ubitux> so we could use it for the resampling, and volume stuff at least [17:21] <michaelni> id be fine with a harddep [17:21] <Daemon404> this reminds me.... [17:21] Action: Daemon404 goes to submit bugs to clang abotu miscompiling ffmpeg [17:21] <saste> ubitux: i remember at some point ffmpeg could be compiled with no lavfi, i don't know if that's still the case [17:21] <ubitux> though, i don't know how the audio drift could be implemented in libavfilter [17:21] <saste> i'm happy leaving michaelni to decide about it [17:22] <ubitux> saste: i added a --disable-avfilter in fate, and it seems to build ffmpeg fine [17:22] Action: michaelni is not really maintaining or testing ffmpeg without lavfi [17:22] <michaelni> ive no real oppinion on a harddep either way [17:23] <michaelni> but i think someone should "maintain" this combination if we support it [17:23] <ubitux> actually, it looks quite a burden to maintain both ffmpeg+internal and ffmpeg+lavfi [17:24] <michaelni> also theres the question if anyone uses it ? [17:24] <ubitux> atm i'm trying to add the -af, and messing with do_audio_out to avoid resampling but keep audio drift is quite complicaetd [17:24] <Daemon404> ubitux, there's an old saying: eat your own dogfood [17:25] <Daemon404> that is, use your own apis instead of custom internal ones [17:25] <ubitux> the thing is, we need to do both atm [17:25] <ubitux> but libavfilter is not up to the ffmpeg internals [17:25] <ubitux> at the moment [17:26] <ubitux> i would be happy to have a -ar doing -af aresample=... and similar [17:38] <ubitux> saste: it seems to work with the configuration [17:38] <ubitux> thanks :) [17:39] <saste> ubitux: did you have to uncomment the sample_rate setting thing? [17:40] <ubitux> yes [17:41] <ubitux> and i need to do something similar for the channel layout [18:04] <CIA-48> ffmpeg: 03Stefano Sabatini 07master * r5d12b83ff1 10ffmpeg/ (doc/ffmpeg.texi ffmpeg.c): [18:04] <CIA-48> ffmpeg: ffmpeg.c: implement -debug_ts option [18:04] <CIA-48> ffmpeg: Mostly useful for debugging purposes. [18:04] <CIA-48> ffmpeg: 03Stefano Sabatini 07master * r175ab5f9a9 10ffmpeg/libavfilter/ (af_ashowinfo.c vf_showinfo.c): lavfi/*showinfo: use av_ts2* macros in showinfo and ashowinfo [18:04] <CIA-48> ffmpeg: 03Stefano Sabatini 07master * r1e6408a18e 10ffmpeg/libavfilter/vf_showinfo.c: [18:04] <CIA-48> ffmpeg: lavfi/showinfo: only print checksum for the existing planes [18:04] <CIA-48> ffmpeg: Make the output a bit nicer. [18:04] <CIA-48> ffmpeg: 03Stefano Sabatini 07master * rf89f9bdadd 10ffmpeg/libavfilter/af_ashowinfo.c: [18:04] <CIA-48> ffmpeg: lavfi/ashowinfo: only print checksum for the existing planes [18:04] <CIA-48> ffmpeg: Make the output a bit nicer. [18:04] <ubitux> saste: it's too bad i can't really use the av_ts2* macro in silencedetect [18:05] <saste> ubitux: why? [18:05] <ubitux> because of the precision [18:05] <ubitux> 1 second is not enough [18:08] <saste> up to 6 digits are used [18:08] <ubitux> anyway, -af aresample works, but changing layout filters like pan don't, even if i set channel layouts and channels correctly in configure_audio_filters [18:08] <saste> but it's using %g format so trailing 0s are not shown [18:09] <saste> (which is a bit ugly for vertical indenting purposes) [18:09] <ubitux> mmh i may have messed in the usage then, 'will look again [19:26] <lionaneesh> Hey guys , I just stopped by to thank you all for allowing me do some FFMPEG tasks [19:26] <lionaneesh> I did 2 ffmpeg tasks i guess (for GCI) [19:26] <lionaneesh> and today they have announced the grand prize winners >> http://google-opensource.blogspot.in/2012/02/google-code-in-2011-grand-priz… [19:27] <lionaneesh> Bye guys [19:27] <lionaneesh> Thanks a lot [20:34] <Compn> wbs : why mentor rtmpe support ? [20:34] <Compn> i mean [20:34] <Compn> thats a lot of work [20:34] <Compn> to duplicate librtmp [21:20] <wbs> Compn: well, if someone is interested in doing it, I might help mentoring it. but I'm not personally totally convinced that it's necessary either [21:22] <michaelni> wbs, IMHO if theres a problem or shortcomming with librtmp then it should be reported & fixed in librtmp [21:22] <michaelni> only if that fails should other pathes be considered ... [21:23] <michaelni> i mean just because you find a student to write a new kernel doesnt mean that it would be good idea [21:23] <wbs> that's a valid point [21:28] <Compn> well i dont mind people working on things. just trying not to waste effort on duplication . [21:36] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * r2bc610ddc5 10ffmpeg/libswscale/output.c: [21:36] <CIA-48> ffmpeg: swscale: yuv2422_1_c_template: fix rounding [21:36] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [21:36] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * rf03a029939 10ffmpeg/libswscale/output.c: [21:37] <CIA-48> ffmpeg: swscale: Add cliping to yuv2422_1_c_template() [21:37] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [21:37] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * rc0c32a3704 10ffmpeg/libswscale/output.c: [21:37] <CIA-48> ffmpeg: swscale: Fix rounding in yuv2rgb_1_c_template() [21:37] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [21:37] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * r94bc672489 10ffmpeg/libswscale/output.c: [21:37] <CIA-48> ffmpeg: swscale: yuv2422_2_c_template() add cliping. [21:37] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [21:37] <CIA-48> ffmpeg: 03Michael Niedermayer 07master * ref96bf294d 10ffmpeg/libswscale/output.c: [21:37] <CIA-48> ffmpeg: swscale: Fix rounding in yuv2mono_1_c_template() [21:37] <CIA-48> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [22:04] <ubitux> michaelni: i see you're doing some swscale fixes, did you see http://fate.ffmpeg.org/report.cgi?time=20120214190603&slot=x86_64-archlinux… ? [22:04] <ubitux> they seem related to swscale somehow [22:05] <ubitux> at least a bunch of them [00:00] --- Wed Feb 15 2012

1 0

ffmpeg.log.20120214
by burek 15 Feb '12

15 Feb '12

[00:18] <drno_> Hiya folks [00:20] <drno_> Question that I'm hoping someone may have experience with... I've got a video stream that I'm encoding from raw images (tv tuner), and writing to a flash media server (red5). I've verified that the stream is good, that the encoding is working (h.264 stored in FLV container), and that all is getting to RED5 (viewed the stream with a flash viewer). [00:20] <drno_> So, up to that point, I know it's functional. [00:20] <drno_> I want to use ffmpeg to capture that stream and save it to a file. [00:20] <drno_> So, I've run the following command: ffmpeg -i "rtmp://localhost/oflaDemo/streamTest live=1" -an -vcodec copy -y test.flv -loglevel verbose [00:21] <drno_> It displays the metadata output, then hangs. I see "Ping's", but nothing gets written. If I hit CTRL-C twice, it exists, giving messages stating that no data was written. [00:22] <drno_> Any ideas on what I could try? [00:25] <Prudhvi> Hi, is there a RPM repo for latest ffmpeg [00:25] <Hyperi> What's wrong with source? [00:25] <Hyperi> You get the best out of it [00:25] <pasteeater> Prudhvi: i don't know of any, but you can follow this: https://ffmpeg.org/trac/ffmpeg/wiki/CentosCompilationGuide [00:25] <Hyperi> + some of the libraries need manually enabled [00:25] <Prudhvi> pasteeater: perfect :) Thanks [00:26] <pasteeater> "you're the best..around. nobody is going to bring you down" [00:27] <pasteeater> drno_: are you using ffmpeg to feed to the video to red5? [00:28] <drno_> pasteeater: yes indeed. [00:28] <pasteeater> why not make two outputs? one to red5 and one to a file [00:29] <drno_> My understanding is that two outputs mean two encodes. [00:29] <drno_> one my box can handle -- two might make it a tad unhappy. [00:30] <drno_> Or is there some way to mux the output without re-encidng? [00:30] <drno_> encoding, rather. [01:10] <MrKeuner> hello, how can I rip bunch of vob files from out of a dvd into a single mpeg2 file without any conversion? [01:12] <sacarasc> If you're lucky cat *.vob > blah.mpeg [01:14] <MrKeuner> sacarasc, how will I know I am lucky? [01:14] <sacarasc> If it works. [01:16] <MrKeuner> sacarasc, I need to know that file is healthy, some application's playing the file does not necessarily prove this [01:16] <MrKeuner> thanks for trying to help [01:17] <sacarasc> If you use ffmpeg to do it, you're basically doing the same thing anyway. [01:20] <MrKeuner> I understand that. I just need to be sure that the output file is healthy. I would trus t ffmpeg more than I trust cat for this operation [01:20] <Zeranoe> Does anyone know where I can get a source archive of libnut? I see it is made by MPlayer/FFmpeg so it should be around here somewhere? [01:29] <pasteeater> Zeranoe: git://git.ffmpeg.org/nut [01:29] <pasteeater> 'git clone git://git.ffmpeg.org/nut' to be specific [01:31] <Zeranoe> pasteeater: Thanks a lot [02:20] <^sandro^Zzz> hello everyone [02:21] <^sandro^Zzz> anyone know how to fix this error [02:21] <^sandro^Zzz> [libx264 @ 0x1781bc0] Input picture width (704) is greater than stride (0) [02:21] <^sandro^Zzz> Video encoding failed [02:39] <pasteeater> ^sandro^Zzz: is there a reason you're not using the standard x264 encoding presets? [02:52] <^sandro^Zzz> what do you mean [02:52] <^sandro^Zzz> -preset low [02:52] <^sandro^Zzz> or whatever ? [02:52] <^sandro^Zzz> doesn't change anything i have tried [02:58] <pasteeater> ^sandro^Zzz: as in ffmpeg -y -i udp://@:5000 -an -c:v libx264 -preset medium -crf 24 -map p:"167" output.ts [03:22] <^sandro^Zzz> sorry pasteeater [03:22] <^sandro^Zzz> ress [q] to stop, [?] for help [03:22] <^sandro^Zzz> [libx264 @ 0x14d8ea0] Input picture width (704) is greater than stride (0) [03:22] <^sandro^Zzz> Video encoding failed [03:22] <^sandro^Zzz> Segmentation fault [03:22] <^sandro^Zzz> same error.. using version 120 of x264 now [03:22] <^sandro^Zzz> updating ffmpeg too [03:26] <^sandro^Zzz> hey anyonek know how to fix this error on compile ffmpeg [03:26] <^sandro^Zzz> collect2: ld returned 1 exit status [03:26] <^sandro^Zzz> make: *** [libavutil/libavutil.so.51] Error 1 [03:26] <^sandro^Zzz> oops.. [03:27] <^sandro^Zzz> collect2: ld returned 1 exit status [03:27] <^sandro^Zzz> make: *** [libavutil/libavutil.so.51] Error 1 [03:27] <^sandro^Zzz> damn.. wont let me paste. [03:27] <^sandro^Zzz> http://pastebin.com/sWsJQCne [03:27] <^sandro^Zzz> that is easier :P [03:29] <VooDooNOFX_> ^sandro^Zzz, you're missing some libs on this system I think [03:34] <^sandro^Zzz> which ones is the trick [03:34] <^sandro^Zzz> :( [03:40] <pasteeater> ^sandro^Zzz: what system/os/distro/whatever [03:40] <pasteeater> and show your ./configure [03:49] <^sandro^Zzz> ubuntu [03:50] <^sandro^Zzz> ./configure --enable-libmp3lame --enable-libtheora --enable-libvorbis --enable-libfaac --disable-mmx --enable-shared --enable-libxvid --enable-gpl --enable-nonfree [03:52] <pasteeater> why disable-mmx? [03:52] <pasteeater> what ubuntu version? [03:54] <^sandro^Zzz> 11.04 [03:55] <pasteeater> http://ubuntuforums.org/showthread.php?t=786095 [04:16] <cosmint> This works: ffmpeg -i input.avi -vcodec libx264 -an -x264opts crf=8 output.flv, but when I change flv to f4v it complains "Unable to find a suitable output format" [04:20] <cosmint> -f mp4 seems to fix the problem, not that i know exactly why [04:39] <relaxed> cosmint: ffmpeg doesn't know about the f4v format. [04:40] <relaxed> if you don't force a format with -f $format, ffmpeg will use the extension to set it. [04:40] <cosmint> i see, this wiki page made me think f4v would work, http://en.wikibooks.org/wiki/FFMPEG_An_Intermediate_Guide/h.264 [04:40] <cosmint> so i made an mp4 or m4v and flash plays it [04:41] <cosmint> good enough i guess, now to find the difference between f4v and m4v [04:41] <relaxed> that is old, outdated and wrong [04:42] <cosmint> ya [04:43] <relaxed> how does f4v and m4v differ from flv and mp4? silly apply naming convention? [04:43] <relaxed> I would assume "-f flv output.f4v" would do it [04:44] <relaxed> s/apply/apple/ [04:50] <cosmint> I think I've figured it out. f4v is its own format, distinct from flv, but very close to mp4, or maybe just adobe's implementation of mp4 [04:51] <cosmint> I lose some features I don't think I need by using mp4 instead of f4v, cue points or something. [04:51] <cosmint> So I'm happy for now. [04:54] <funkster_> looking to create a photo slideshow with some fading between the photos - anyone point me in the right direction on how to do this? [04:56] <relaxed> funkster_: Shouldn't be hard- ffmpeg has a fad filter [04:56] <funkster_> yeah? hmm ok lemme try and find that [04:56] <relaxed> loop each a certain number of times and use a similar fade for each image [04:58] <funkster_> happen to have an example somewhere, the man pages just confuse me more then i am already. [04:58] <relaxed> not really, I'm just thinking out loud [04:59] <relaxed> did you find fade in the man page? [05:01] <funkster_> the main ffmpeg site, i found fade filter, yeah. [05:17] <cosmint> i found out why flashsv2 is experimental :p [05:19] <cosmint> thanks relaxed [06:42] <lahwran> how does one loop an input file into an output file? I have a file I would like to loop so it repeats, say, 10 times in a single stream [06:45] <relaxed> what kind of file? [07:04] <relaxed> lahwran: look at concat in the man page [07:06] <lahwran> relaxed: thanks [07:06] <lahwran> and, mp4 [07:50] <trupheenix> hi. i want to know how i can find out a formula to cut my video based on keyframes? i want to do this for streaming. [10:52] <Xcellence> Hi there.. is there any way atm to decode a vc-1 interlaced stream with ffmpeg? On google i just found some mentions of w32codecs but i dont know if ffmpeg supports this or how i can use them.. :/ [11:37] <Xcellence> nobody knows? :( [11:47] <zap0> nobody knows is a statement. adding a question mark doesn't make it a question. [11:50] <Xcellence> ;) [12:59] <stonie> can ffmpeg do audio streams as flac? [14:02] <bgmarete> Hello. How can I make MPEG-2 encoding more scalable? That is, make it use more threads + do more fps? At the moment, the maximum scalability I can achieve is 3 threads. [14:03] <bgmarete> That is, I am using only 3 cores, no matter my argument to the `-threads' option. [14:05] <sacarasc> Are you putting -threads in the output options? [14:19] <bgmarete> sacarasc: Yes. [14:20] <bgmarete> sacarasc: `ffmpeg -i input.mkv -vf yadif -map 0 -threads 4 -vcodec mpeg2video -r 25 -g 12 -bf 3 -s 720x576 -aspect 4:3 -pix_fmt:v yuv420p -mbd rd -cmp 2 -subcmp 2 -b:v 6000k -maxrate 6000k -minrate 6000k -bufsize 6000k -acodec mp2 -ac 2 -ar 48000 -b:a 256k -f vob -y out.mpg' [14:58] <SvenL-> is there something special to do to be able to use -pix_fmt with ffplay ? [14:58] <SvenL-> tells me it's deprecated, -pixel_format option is "not found" [15:20] <Xcellence> Hi there.. is there any way atm to decode a vc-1 interlaced stream with ffmpeg? On google i just found some mentions of w32codecs but i dont know if ffmpeg supports this or how i can use them.. :/ [15:42] <MrKeuner> hello, how can I rip bunch of vob files from out of a dvd into a single mpeg2 file without any conversion? [16:06] <relaxed> MrKeuner: you can concat a bunch of vobs without any conversion using ffmpeg [16:07] <MrKeuner> how do I do it [16:08] <relaxed> ffmpeg can't "rip" the dvd though [16:08] <MrKeuner> by rip I meant getting the mpeg2 file out of the dvd [16:09] <MrKeuner> this is a dvd created from a video8 cassette [16:09] <relaxed> this would be easier with mplayer: mplayer -dumpstream -dumpfile output.vob dvd:// [16:10] <MrKeuner> relaxed, absolutely no conversion? [16:10] <relaxed> zero [16:10] <MrKeuner> and vob to mpeg2? [16:11] <MrKeuner> file dvd/file.vob returns mpeg2 [16:11] <MrKeuner> but I don't know if simply changing the extension is ok [16:11] <relaxed> .vob = MPEG-2 PS format [16:12] <relaxed> it's a container that holds mpeg2 video and audio streams [16:14] <relaxed> if it will give you the warm fuzzies you can use `output.mpg` [16:32] <MrKeuner> if I encode some.dv to mpeg2 versus an mpeg2 in a vob, what's the difference? [16:40] <relaxed> you want to go from dv -> dvd ? [16:40] <MrKeuner> relaxed, no trying to understand what container format vob is for [16:40] <relaxed> DVDs [16:40] <MrKeuner> mpeg2 already has sound in it, no [16:43] <relaxed> mpeg2 is a video codec; mpeg ts and mpeg ps are containers that hold mpeg2 video [16:45] <MrKeuner> relaxed, what is a container for? [16:46] <relaxed> http://en.wikipedia.org/wiki/Digital_container_format [16:52] <MrKeuner> relaxed, so vob file is like a Java class inheriting mpg format so that changing the file extension from vob to mpg does not break any mpg players [16:52] <kriegerod> is there some avfilter to transform variable-frame-rate video to fixed-frame-rate one? If not, then a) what could be adviced way to do that? b) is it good idea to implement such filter? [16:56] <relaxed> kriegerod: I believe ffmpeg is incapable of outputting vfr, so cfr is all you get. [16:57] <kriegerod> relaxed, sorry i forgot to mention it, i work with API, not utility [16:58] <relaxed> still, if ffmpeg can do it you can too. [17:00] <kriegerod> sure, i just asked in hope to save efforts on researching [17:00] <relaxed> maybe ask in #libav too, they're pretty good about answering such questions. [17:02] <aClam> He did minutes before [17:02] <relaxed> oh, you're right :) [17:23] <renderTom> Hello everybody [17:23] <renderTom> I was wondering, is it possible in ffmpeg to use some other scaling method? As '-s" seems to produce "soft" scaling [17:28] <microchip_> renderTom: -sws_flags spline (for example) [17:31] <renderTom> oh, thanks, will look at that [17:43] <renderTom> I am hacing no luck finding understandable info about scaling :) [17:44] <renderTom> what I am after is pretty simple - "-s" scaling seems to get pretty soft output, so I wonder if there's a option to make it better or to use some other scaling method? [17:47] <aClam> You didn't try sws_flags ? [17:49] <renderTom> well, not yet, as I am not sure how to use it :) [17:49] <microchip_> renderTom: 'ffmpeg -h' and look for -sws_flags [17:49] <renderTom> oh yeahhhh [17:49] <renderTom> solly me :) [17:49] <microchip_> renderTom: to use it, specify the scaler you want. ie, -sws_flags lanczos or -sws_flags spline [17:49] <renderTom> silly [17:50] <renderTom> yeap, got it [17:50] <renderTom> what method could you reccomend? [17:50] <microchip_> I always use spline, but lanczos is pretty good too [17:50] <renderTom> mhm, I see [17:51] <microchip_> i recall that -s default to linear, that's why you get softness [17:51] <renderTom> will it take longer to transcode using this scaling method? [17:51] <microchip_> yes, a bit [17:51] <microchip_> spline is more demanding than lanczos, though [17:51] <renderTom> but you'd recommend spline to try at first. Thanks will give it a go [17:52] <microchip_> i recommend both spline and lanczos :) [17:53] <renderTom> :) [17:53] <microchip_> use either one of them [17:53] <renderTom> will try spline at first [17:53] <renderTom> by the way, I am on iMac i7 with 4 cores (+ 4 virtual), [17:53] <renderTom> and was wondering, how to distribute or use all cores during compression? [17:53] <renderTom> is it even possible? [17:54] <microchip_> -threads 4 [17:54] <renderTom> was trying -thred 4 with no luck [17:54] <microchip_> hmm [17:54] <microchip_> don't know, i'm not very much into ffmpeg [17:54] <renderTom> at least I didn't see any change in activity monitor [17:54] <aClam> -threads 0 [17:55] <aClam> I used that to auto detect, but it will depends on what compression library you are using [17:55] <renderTom> and also, is it important where to put this -thread 0 command? [17:55] <aClam> if you are using libx264 it will work [17:55] <renderTom> I am going from h264 to prorez [17:55] <renderTom> ffmpeg -i /_in.mov -vcodec prores -profile 2 -aspect 16:9 -s 1024x576 -an /_out.mov [17:55] <aClam> Then it will depend on the prores encoder whether it supports multithreaded encoding [17:55] <renderTom> so does it matter where to put -threads and -sws_flags command? [17:56] <renderTom> aClam, I see, that makes sence [17:56] <aClam> I haven't used those flags since the command line syntax reshuffle so I wouldn't know exactly [18:00] <renderTom> ok, I'll put -treads and sws_flags at the end of the code [18:00] <relaxed> anywhere after the input should be fine [18:01] <renderTom> yep, so it seems h264 to prores has no effect with -threads 0 [18:02] <renderTom> as I see in Activity Monitor (mac) that process jumps from one core to another [18:02] <relaxed> use -vf scale=1024:576,setdar=16:9 [18:02] <renderTom> and not 4 of them are working simultaniously [18:03] <relaxed> 0 only works with libx264. try -threads 4 [18:03] <renderTom> -vf? Is it different ftom -swf_flags? [18:04] <aClam> It's not "-treads" by the way [18:05] <relaxed> using a recent version of ffmpeg, your command should be: ffmpeg -i input -an -threads 4 -swf_flags spline -vf scale=1024:576,setdar=16:9 -vcodec prores -profile 2 output.mov [18:05] <aClam> -vf is if you wanted to say -s more verbose [18:06] <aClam> For example: "-vf scale=854:480" [18:06] <renderTom> sorry, I dont know what is verbose [18:06] <renderTom> oh, ok [18:06] <aClam> instead of "-s 854x480" [18:10] <renderTom> with that syntax <relaxed> just provided I don't see any change :) [18:11] <renderTom> any ways, I am just testing, so I have time even if one core is working [18:11] <renderTom> thanks for your imput guys [18:12] <relaxed> renderTom: prores may not be multithreaded. The syntax i used is the recommended way now. [18:15] <renderTom> I see [18:15] <renderTom> thank you [21:13] <JDuke128> hello , i can read file by writing this function ( avformat_open_input ) but how can i get from a network stream ? [21:36] <JDuke128> hey [21:36] <JDuke128> i can read file by writing this function ( avformat_open_input ) but how can i get from a network stream ? [21:46] <Zilly> Can FFmpeg encode rgb24 to Theora or only yuv? [21:52] <pasteeater> Zilly: ffmpeg will probably autoslect yuv420p for the output [21:53] <pasteeater> if your input is rgb [21:53] <Zilly> pasteeater, Yeah. I meant that I'm trying to encode a set of frames to a video. The images are in rgb24 and, when encoded with Theora, look awful. However, when encoded with x264, it looks perfect. [21:55] <teratorn> Zilly: bad quality, or corrupted? [21:56] <teratorn> JDuke128: use a correct URL and it will be opened [21:56] <Zilly> teratorn, it appears that it may be corrupted. It's not playing back properly. The part that does is pretty bad, though. [21:57] <pasteeater> Zilly: use a pastebin site to show your ffmpeg command and the complete console output [21:57] <Mista_D> Anayway to have FFMPEG store mp4 atoms in the beginning of the file? [21:57] <Zilly> pasteeater, I don't have the code with me at the moment. Do you know of anything that I could look for that would be suspicious? It complains that the images are in rgb24 instead of YUV420p. [21:58] <tweek__> AtomicParsley can do that [21:58] <teratorn> Zilly: so.. don't do that? put them in yuv420p first [21:58] <JEEB> Mista_D, the qt-faststart thingy in the tools or whatever folder [21:58] <JEEB> you have to build it separately but it's there [21:59] <Zilly> teratorn, my question was if there is anyway to have Theora encode rgb24 images instead of yuv420p. [21:59] <JEEB> Zilly, probably not unless the encoder has it as a feature :P [21:59] <JEEB> I'd be surprised if Theora supports RGB tbqh [21:59] <teratorn> Zilly: sws_scale() [21:59] <Mista_D> JEEB: thnaks. [22:00] <Mista_D> I'll stick to mp4box for container fixes. [22:03] <Zilly> teratron, thanks. It seems like FFmpeg does that by default, though, right? I'd probably be better off saving the frames in yuv420p. Is there any format that does this by default? JPEG is YcRcB, right? [22:05] <vadim_> hi 2 all [22:07] <tweek__> usually I find it's easier to just rely on mp4box exclusively for remuxing [22:09] <JDuke128> teratorn , i want to decode from char* custom binary data instead of file [22:15] <teratorn> Zilly: Y`CbCr, yea [22:16] <Zilly> teratorn, won't that format cause the same problem that rbg24 causes? [22:17] <teratorn> you said it complains that the images are not in yuv420p [22:18] <teratorn> Zilly: so just try giving it frames in yuv420p and see what happens [22:18] <teratorn> you can do the pixel format conversion yourself with sws_scale() before encoding the frames [22:18] <Zilly> teratorn, Yeah. That's what I want to do. I'm under the impression that Y`CbCr and yuv420p are not the same. [22:18] <teratorn> I don't know if it's already doing that for you or not [22:18] <teratorn> well, you have rbg24 already so just start with that [22:19] <cbreak> YCbCr is a color standard [22:19] <cbreak> it doesn't specify color subsampling [22:19] <cbreak> I use YCbCr with a 422 subsampling for example [22:19] <Zilly> teratorn, ok. I'll do that. [22:20] <Tjoppen> cbreak: don't forget color siting [22:21] <Zilly> cbreak, I see. I don't really know how to convert a JPEG to different subsamples. I'm just using PIL to do some really light editing on frames. [22:23] <teratorn> Zilly: erm, well, sws_scale() does pixel format conversions including re-subsampling I believe [22:24] <teratorn> Zilly: so load (decompress) your jpeg frame, and run it through sws_scale() specifying the correct source and destination pixel formats [22:25] <Zilly> teratorn, ok. That's what I'm assuming I'll have to do. I was just hoping that I could continue using PIL. [22:25] <teratorn> JPEG can be saved in various subsampling formats, so you'll have to get that from the header I guess [22:27] <teratorn> Zilly: you're already using ffmpeg to encode a video right? [22:27] <Zilly> teratorn, Yeah. [22:27] <teratorn> Zilly: from static jpeg files? [22:27] <JDuke128> whats the fastest way to print avcodec_decode_video2 output to put framebuffer any extra lib ? [22:27] <Zilly> teratorn, Well, PNGs at the moment--they were working better to encode with x264. [22:27] <Zilly> teratorn, but, it's no difference to use JPEG. [22:27] <teratorn> Zilly: and you're doing this from the command-line (not in C code) ? [22:28] <Zilly> teratorn, more or less. I'm just running a command from Python's commands lib. [22:28] <teratorn> Zilly: hehe. don't use the 'commands' module. use subprocess pl0x [22:29] <Zilly> teratorn, Err... WAS using commands--now using subprocess >.< [22:29] <teratorn> Zilly: and uh, I think that /should/ just work and do the correct pixel format conversion for you [22:29] <teratorn> Zilly: so sounds bug-report/mailing-list worthy [22:31] <Zilly> teratorn, wait. Should that work with PNG rgb24? Or are you saying if I have this problem with JPEG? [22:31] <teratorn> I'm saying it should work with any set of valid image files that ffmpeg knows how to load [22:32] <Zilly> teratorn, what if I gave it a pretty weird set where some have an alpha channel and some don't? Would that affect it? [22:32] <teratorn> uhhh [22:32] <teratorn> quite possibly [22:33] <Zilly> teratorn, that was my next thing to try. I just figured that I'd have to give it yuv420p since it complained about it. [22:33] <teratorn> I would tend to think the alpha channel would be ignored [22:33] <Zilly> teratorn, thanks! [22:33] <teratorn> but hard to say [22:34] <teratorn> Zilly: please do follow this up, it sounds off [22:37] <Zilly> teratorn, ok. I'm actually at work >.< This is a side project. I'll check back into IRC later tonight / tomorrow with results. [22:38] <teratorn> very good [22:39] <Zilly> teratorn, thanks! [22:56] <__vincent> does ffmpeg know how to write an mp4 header? [22:57] <sacarasc> Yes. [22:59] <jensverwiebe> hi folks [22:59] <jensverwiebe> i'am the blender OSX maintainer and ran into a problem with ffv1 both on 0.10 release and trunk [23:00] <jensverwiebe> ./ffmpeg -i /Volumes/SystemHD/Users/jensverwiebe/Downloads/Hausbrennt8d0001_0250a.mov -pix_fmt rgb32 -vcodec ffv1 -an output.mov is scrambled due pix_fmt rgb32 [23:00] <jensverwiebe> any clou ? [23:02] <jensverwiebe> made a crosscheck with 0.87 and all is k in this version, same configured and compiled [23:40] <jensverwiebe> hmm, no one ? [23:51] <relaxed> jensverwiebe: can you stick a small sample of your input somewhere? [23:51] <jensverwiebe> relaxed: sure, moment [23:52] <relaxed> sorry, I have to run right now. go ahead and paste the url and I'll look at it when I return. [23:53] <jensverwiebe> relaxed: http://www.jensverwiebe.de/temp/ffmpeg_input_rgb32.mov [23:53] <jensverwiebe> relaxed: btw, tested now all rgb24, rgb444 etc, only rgb32 seems to be broken with ffv1, but nit with huffyuf for example [23:54] <jensverwiebe> always same input movie used [23:54] <jensverwiebe> relaxed: + afaik works on Linux [00:00] --- Wed Feb 15 2012

1 0