Ffmpeg-devel-irc

Download

ffmpeg-devel-irc@ffmpeg.org

August 2014

1 participants
60 discussions

ffmpeg-devel.log.20140830
by burek 31 Aug '14

31 Aug '14

[00:06] <iive> wm4: ffmpeg is trademarked, so the debian ffmpeg "deprecated" program could be considered counterfeit. I'm sure Debian people would love us if we send them C&D letter :) [00:10] <wm4> actually that's a great idea, considering the lena crap [00:13] <jamrial> has anyone pointed them that libav also ships the same image? [00:13] <kepstin-laptop> jamrial: yes, that has been pointed out. [00:14] <jamrial> I see [00:14] <kepstin-laptop> or at least, has been mentioned in the stuff on the ffmpeg-devel list, I don't actually know if it made it to the debian folks. [00:33] <Compn> wm4 / iive : its actually terrible idea, debian will balk at distributing ffmpeg much like the iceweasel thing [00:34] <Compn> so its kind of screwed if you do, screwed if you dont type of situation [00:34] <iive> i have been sarcastic... i guess the smiley at the end wasn't enough of a hint. [00:34] <Compn> yes but wm4 wasnt [00:35] <Compn> who was working on that message ? [00:35] <Compn> for months and months [00:35] <beastd> kierank: I am disappointed because of you trying to urge michaelni into attending VDD for resolving FFmpeg vs Libav. If michaelni should go to VDD than that is only to enjoy the event, interesting talks and sessions. And IIRC j-b said it at least once before that VDD is not a forum for FFmpeg nor Libav. [00:36] <j-b> because you're a bunch of kids [00:36] <Compn> iirc kierank suggested michael come to a meeting for years [00:36] <Compn> so this is nothing new beastd [00:36] <j-b> else, FFmpeg would be the center of the multimedia community like it should [00:36] <Compn> it is the center, and theres nothing you can do about it j-b [00:36] <Compn> otherwise everyone would have switched to libav long ago :P [00:37] <j-b> Compn: right :) [00:37] <Compn> j-b : remi is difficult to talk to :\ [00:37] <Compn> ehe [00:43] <Daemon404> eh [00:43] <Daemon404> i found remi one the easiest to talk to irl [00:43] <Daemon404> then again, i am not a a vlc person. [01:41] <j-b> why the hell do people push for MSVC? [01:41] <j-b> it's slower, less maintained, less standard [01:42] <Compn> whats the alternative ? [01:43] <Compn> i'm seriously asking because i forgot [01:43] <Compn> gcc + mingw/cygwin is one of them [01:43] <Compn> i forgot what the 'standard' was now [01:43] <jamrial> gcc? mozilla uses it for all their builds for example. even the windows ones [01:43] <Compn> oh , they like msvc debugging [01:43] <Compn> theres some debugging that gdb doesnt have or something [01:44] <j-b> gcc works fine [01:44] <jamrial> or maybe not. these don't look like gcc flags... [01:47] <pross-au_> gcc can output microsoft pdb debug info now? [01:50] <Compn> theres always something the msvc people complain about gcc ;P [01:51] <J_Darnley> All those casts remind me of c++ [01:55] <j-b> pross-au_: it can. [01:58] Action: pross-au_ TIL [01:58] <j-b> pross-au_: it is not very simple yet. [01:58] <j-b> pross-au_: we're working on that. [02:05] <kierank> Compn: http://hwcdn.net/j9t9v3v5/cds/ffmpeg.tar.gz [02:05] <kierank> new dump from elemental [02:06] <kierank> ubitux: ^ [02:06] <kierank> some interesting stuff in there [02:24] <cone-440> ffmpeg.git 03Michael Niedermayer 07master:596636a474ab: avcodec/snow: check coeffs for validity [02:54] <Compn> michaelni : want to try skype or other video chat to vdd ? we can put you on a tablet or laptop and carry it around... [02:55] <Compn> you wont be there in person, but in spirit (or bytes) . can even do richard stallman, open video format only ? maybe livestream snow? :P [02:56] <Compn> j-b : will google lend us some laptop or tablet ? [02:56] <j-b> no [02:57] <kierank> you can use mine [02:57] <Compn> good battery life is a plus. [02:57] <kierank> ok then mine is out of the question [02:57] <kierank> my battery is buggered [02:57] <kierank> lasts about 5 mins [02:57] <Compn> all my laptop batteries are toast [02:58] <Compn> buy cheap chinese one off amazon? nahh [02:58] <Compn> well michaelni , if you make a livestream, we'll put it on a bunch of puters then. whatever puter we have lying around [02:58] <kierank> nervous about doing that [02:58] <Compn> use ffserver :) [02:59] <Compn> i think it would be fun to have michael (or any other devel who wont be there in person) to video chat in [02:59] <Compn> wheres the list of devs who are attending ? [02:59] <Compn> or registered [03:01] <Compn> google circle chat hahaa [03:04] <pross-au_> oh the irony [03:16] <BBB> hey guys can we drop one of the prores decoders now? [03:16] <BBB> or is there a list of which is better in terms of speed, features? can someone measure that? [03:16] <BBB> Compn: how about you [03:20] <jamrial> huh, they are both lgpl yet one is explicitly called "prores_lgpl" [03:20] <BBB> I think that historic [03:20] <BBB> they started as gpl lgpl, but then the gpl got relicensed, probably for competitiveness reasons [03:25] <pross-au_> wait, we have two encoders and two decoders! [03:44] <Compn> BBB : since kostya has quit, probably we can just merge them. but i dont know which is faster / features etc [03:44] <Compn> j-b : I complained about that several times, and was ignored. [03:44] <Compn> j-b : ok, you have a problem with anonymous code ? [03:45] <Compn> j-b : what do you suggest we do with anony code ? delete it? ask author to fess up? [03:45] <Compn> ask author to transfer copyright ? [03:45] <Compn> i'm genuinely asking , devels can vote on whatever proposal... [03:45] <j-b> either fess up, or give the right to relicensiate to the governing body [03:45] <Compn> sounds good [03:46] <j-b> or relicense to the simplest license [03:46] <j-b> aka, WTFPL or BSD 2-clause [03:47] <Compn> i'm curious about your re-licensing complaint , could you be more specific? do you mean adding copyrights from known authors when they were not present in original libav commit ? [03:48] <Compn> or the copyright multiple(a)multiple.x ? [03:48] <Compn> we could send patches to improve these things [03:48] <Compn> even i could do it :) [03:49] <Compn> it would be nice if , instead of complaining on mailing list, either patches or report to bug trac was used... [03:49] <j-b> sorry, but no. [03:49] <Compn> well i can make new bugs [03:49] <j-b> your project is hopeless wrt copyright [03:49] <Compn> so could you be specific [03:49] <Compn> argh :) [03:53] <Compn> so dont try to fix? :P [03:55] <Compn> its mplayer all over again :P [03:55] <j-b> example [03:55] <j-b> http://git.videolan.org/?p=ffmpeg.git;a=history;f=libavfilter/vf_hue.c;hb=7… [03:56] <j-b> "all contributors" [03:56] <j-b> not listed [03:56] <j-b> and when you look closely [03:56] <j-b> http://git.videolan.org/?p=ffmpeg.git;a=commit;h=ae60d2c877e452c623fbe8e312… [03:56] <j-b> This is a port of the MPlayer hue filter (libmpcodecs/vf_hue.c) [03:56] <j-b> right [03:57] <j-b> and all the MPlayer contributors agreed, right? [03:57] <j-b> they were asked, right? every single one of them? [03:57] <Compn> durandial went over the file and i think contacted every author who had copyrightable code ? [03:57] <Compn> have to ask him [03:57] <Compn> the files mplayer history [03:57] <Compn> or just the ports history [03:58] <j-b> all the code that finished in this file. [03:58] <j-b> all of it. [03:58] <j-b> it's not documented who agreed [03:58] <Compn> svn blame mplayer/libmpcodecs/vf_hue.c [03:58] <j-b> therefore, it's useless and very likely wrong [03:58] <j-b> Compn: but also svn log mplayer/libmpcodecs/vf_hue.c [03:59] <Compn> michael , carl, diego [03:59] <Compn> ok [03:59] <j-b> etc.. [03:59] <j-b> so, those are for the licensing [03:59] <j-b> because FFMpeg is Mplayer all-over-again [03:59] <j-b> mostly libavfilter and the limpcodecs (sic) [03:59] <Compn> reimar, michale , clement, diego, syrjala (7 lines) [04:00] <Compn> +carl [04:00] <Compn> almost the same list [04:00] <j-b> http://git.videolan.org/?p=ffmpeg.git;a=commit;h=9e8bbe7d4d1dcd5fec491dbfbb… [04:00] <j-b> How the FUCK this happens? [04:01] <Compn> this is only for license file , not the license in vf_interlaace.c [04:01] <Compn> as you can see from commitdiff [04:01] <j-b> sure, it just proves that you don't care enough about copyrights [04:02] <Compn> it just wasnt added ot LICENSE file , probably committer forgot we maintain such files still [04:02] <Compn> or just plain forgot [04:02] <kierank> http://git.videolan.org/?p=ffmpeg.git;a=commitdiff;h=3fce13679837ce9b43b50a… [04:02] <kierank> was committed as gpl [04:02] <kierank> at least [04:02] <Compn> right [04:02] <jamrial> configure lists it as gpl [04:03] <jamrial> so builds without --enable-gpl never compiled it [04:03] <jamrial> it was simply missing from the license file [04:03] <Compn> dont bother [04:03] <jamrial> that file is a bit unmaintained. it still mentions some idctmmx file that doesn't exist. and all the idct code we have is lgpl afaik [04:04] <Compn> we finally got rid of that test idct file ? ehe [04:04] <j-b> Compn: finally, you add the squashed code without attributions, and then the anonymous devs... [04:05] <Compn> the authors of the squashed code were added, and BBB and (other dev i forgot now) fixed all that before commit iirc [04:05] <Compn> for vp9 [04:05] <Compn> but you dont seem willing to listen [04:06] <j-b> seriously, I do [04:06] <j-b> but your attitude with copyright is a bit light [04:07] <j-b> and all the relicensing of the filters from mplayer are very dubious [04:07] <Compn> you dont think michael did the relicense honestly ? [04:07] <Compn> why not just say that ? [04:07] <j-b> Yes, I don't think so. [04:07] <j-b> I say that, I said that and I will resay that [04:08] <Compn> do you want reimar 's word of relicensing ? [04:08] <Compn> or i mean, what can prove it to you ? [04:09] <j-b> I don't know, proper documentation of the process, for a start... [04:09] <BBB> Compn: I doubt hes talking about vp9; the orig vp9 commit clearly lists all (2) authors that contributed to the whole, and we have a historic tree available that lists the individual commits leading to that whole [04:09] <Compn> i mean, what does proper documentation look like ? [04:09] <BBB> Compn: I think hes talking about other crap where that is not clearly documented [04:09] <j-b> Compn: read my blog about relicensing [04:10] <j-b> read my commits in VLC on relicensing [04:10] <Compn> i think i read it long time ago, let me dig it up [04:10] <Compn> http://www.jbkempf.com/blog/post/2012/How-to-properly-relicense-a-large-ope… [04:10] <j-b> and not the usual "this change is trivial, so not copyrighted" [04:11] <j-b> why when it's relicensed, there is not the list of "all the authors that agreed"? [04:12] <j-b> stuff like this are very bad too: http://git.videolan.org/?p=ffmpeg.git;a=search;h=f7f1f4c7ce9ce689823e13a53b… [04:12] <j-b> asking people to correctly set their git name is too hard? [04:13] <Compn> http://git.videolan.org/?p=ffmpeg.git&a=search&st=author&s=compn [04:13] <Compn> :P [04:13] <j-b> Exactly the same [04:14] <Compn> of course libav has committed me sometimes > http://git.videolan.org/?p=ffmpeg.git&a=search&st=author&s=littler [04:14] <j-b> http://git.videolan.org/?p=vlc.git&a=search&st=author&s=compn [04:14] <j-b> http://git.videolan.org/?p=vlc.git&a=search&st=author&s=littler [04:15] <j-b> Compn: do you have the mplayer git somewhere browsable? [04:15] <Compn> i was pointing people to an mplayer git for a while until i realized it was mplayer2 repo :D [04:16] <j-b> same [04:17] <Compn> so 'videolabs' is the official vlc on android play store right ? [04:17] <Compn> vlc for android beta [04:17] <Compn> but no, let me look for mplayer git somewhere. there used to be a mirror. maybe its gone [04:18] <j-b> Compn: yes it is. [04:18] <j-b> it will be renamed soon [04:18] <Compn> ok, because that confused me :) [04:19] <Compn> https://github.com/pigoz/mplayer-svn , hasnt been updated in 1 year [04:20] <j-b> https://github.com/pigoz/mplayer-svn/commit/a8065c5b042be7042c399d12eb981a7… [04:20] <j-b> example in point [04:20] <j-b> did Paul contact James Crowson before doing the relicensing? [04:20] <j-b> sure, the svn author is michael... [04:20] <j-b> BUT, it's not the copyright holder... [04:21] <j-b> and I took an easy one [04:21] <Compn> that code isnt in the port [04:21] <Compn> :) [04:21] <Compn> yes, you mad. [04:22] Action: Compn trolls j-b into reviewing mplayer code [04:22] <j-b> anyway, I lost too much time on this discussion [04:22] <j-b> but I hope you see my point. [04:23] <j-b> and read my blog about why the "code not being present anymore" is irrelevant [04:23] <Compn> of course, code was checked for copyrights, ported as gpl, relicensed after doing blame , and yet still complaints. [04:23] <j-b> doing blame? I doubt it. [04:23] <Compn> wheres your libreoffice spreadsheet ? [04:23] <Compn> i didnt see it posted... [04:24] <j-b> I can show you at VDD [04:24] <j-b> the fact that you don't agree there is a problem in the way things are held just show that we will just have to disagree. [04:24] <Compn> i agreed with you about copyrights earlier [04:24] <j-b> You are too careless about copyrights [04:24] <Compn> you are just mad to realize i'm not arguing with you [04:25] <Compn> [21:45] <Compn> sounds good [04:25] <j-b> I'm mad? [04:25] <Compn> oh no, i've insulted . i take it back [04:27] <BBB> Compn: any progress on prores? [04:27] <Compn> BBB : who answered prores questions before [04:27] <Compn> someone analyzed them [04:27] <BBB> Compn: just measure it, take a sample of 1000 frames, run time -c:v first_decoder sample -f null - [04:27] <Compn> :) [04:27] <BBB> etc. [04:27] <Compn> ok [04:27] <BBB> then make sure they all support all sample types [04:27] <BBB> (like, alpha, 444 subsampling, etc.) [04:28] <Compn> and then test all known and unknown samples eh [04:28] <BBB> no [04:28] <BBB> all of these are covered in fate [04:28] <BBB> fate has like 5 or 6 samples [04:28] <BBB> so just ffplay all of them with either decoder [04:28] <BBB> thats 10-12 ffplays of a few seconds [04:28] <BBB> cant be that bad [04:28] <BBB> for alpha maybe confirm its indeed there by looking specifically at the alpha data in some way [04:29] <BBB> its a few minutes of work [04:30] <Compn> can you force decoders with ffplay now ? i remember making feature request ... [04:30] <BBB> if not, output to png and browse through the first 20 frames [04:30] <Compn> j-b : heres celements , http://lists.ffmpeg.org/pipermail/ffmpeg-devel/2013-August/147676.html and jeremy https://lists.ffmpeg.org/pipermail/ffmpeg-devel/2013-September/147835.html [04:31] <Compn> alpha is a pain :) [04:32] <j-b> Compn: and the James and the jrsawa ? [04:34] <Compn> syrjala code isnt in the port, because it has to do with colorspaces that are handled in swscale or elsehwere [04:34] <Compn> and i told you the james patch isnt in the port code either [04:35] <Compn> but [04:35] <Compn> like you say, maybe we havent dug deep enough in all of the filters [04:35] <Compn> its possible something slipped [04:36] <Compn> vf_hue looks pretty solid though [04:37] <Compn> so thank you for reviewing it [04:37] <Compn> BBB : i'm looking for prores info [04:37] <Compn> and updating my tablet , getting it ready for trip [04:37] <Compn> android update makes camera work again, hooray [04:38] <Compn> running android 4.4 on a 2011 hp touchpad :) [04:38] <j-b> We will keep disagreeing then... [04:38] <Compn> vlc might kill my kitten! [04:38] <Compn> you say vf_hue still has copyright problem ? [04:39] <Compn> because the code that wasnt in the port maybe influenced the port ? [04:40] <Compn> or because clement , carl, diego and reimar arent listed in copyright [04:41] <Compn> sorry, its late, maybe i have skipped over some of your advice for copyright . i apologize [04:41] <Compn> too much multitasking for me [04:50] <Compn> anyways, thanks for answering my question j-b [04:50] <Compn> will relay to durandial [04:50] <Compn> maybe ubitux knows more about filter ports as well [04:58] <Timothy_Gu> wm4: what was your favorite pastebin with commandline api again? [05:08] <Compn> j-b : reviewing the anonymous code, i dont want to get any more dev arguments. there was a huge fight over prores stuff and thats where elvis presley holds copyright :\ [05:09] <Compn> j-b : so yea i agree ffmpeg has copyright problems. [05:12] <Compn> j-b : here is ticket for vf_hue and libavfilter copyrights http://trac.ffmpeg.org/ticket/3908#ticket [05:26] <Compn> ugh, i'm tired of the crap [05:27] <pross-au_> of which crap do you speak [05:31] <Compn> BBB : btw i cant do the prores stuff right now, too busy. [08:56] <jamrial> cehoyos turning a trac ticket into a flamewar with a single comment [09:09] <kurosu> at the moment I think kostya's encoder produces decodable yet invalid files (eg using a profile not having alpha but still encoding it) [09:09] <vedos> i'm using youtube-dl to download stuff from Youtube... youtube-dl defaults to avconv, but can also use ffmpeg if i install it. which should i use to dump wav files out of the videos? which is more reliable and gives better sound quality? [09:09] <kurosu> I think that's the only one having alpha, but one can probably port it from one to another (license fun) [09:10] <kurosu> I haven't looked a lot to either, but kostya's looked nice with rate control and multithreading [09:10] <kurosu> why not ask the author of the other encoder ? [09:10] <ubitux> kierank: i just made a diff with the latest you send, and i see a system of lock (avpriv_{un,}lock_decodefr) which seems to be used in vc1dec to avoid a race, i see some changes in the mpeg ps parser but it might just be to resync a fix with upstream, and i see the addition of ARIB subs but only the codec id [09:11] <vedos> kurosu: you mean avconv? [09:11] <vedos> authors? [09:12] <kurosu> vedos, sorry no time for anything, gotta go [09:22] <ubitux> vedos: this is a question for #ffmpeg; also, we support only ffmpeg, here. Last, wav is lossless so there is no quality factor going on [10:41] <wm4> what does it mean when the h264 decoder prints "mmco: unref short failure"? [10:42] <vedos> ubitux: i read somewhere that avconv corrupted audio files for someone. that's why i'm leaning toward ffmpeg. it feels to be more better. [10:44] <ubitux> you definitely won't hear that the tools from the libav project are more reliable here [10:45] <wm4> I suspect if vedos uses "ffmpeg" on his installation, he'll just use the "ffmpeg" program provided by Libav, instead of ffmpeg from the ffmpeg project [10:59] <vedos> mirsal: ffmpeg version N-63893-gc69defd Copyright (c) 2000-2014 the FFmpeg developers [10:59] <vedos> is this the "true" version? [11:00] <wm4> vedos: why are you highlighting apparently random people [11:01] <wm4> but yes it's the real ffmpeg [11:01] <wm4> (apparently an outdated git snapshot) [11:02] <nevcairiel> just 2000 revisions out [11:02] <nevcairiel> thats nothing! [11:19] <ubitux> oh ffs [11:20] <ubitux> CE creating a drama all by himself [11:21] <nevcairiel> really wish it was possible to silence him, its not helping the situation at all [11:22] <wm4> calling him names should do the trick (hurr) [11:22] <nevcairiel> you can claim its only his opinion and noone elses, but that usually never works [11:22] <ubitux> and meanwhile suggesting to --disable-opencl in x264 in the libx264 detection failure ticket [11:24] <wm4> "There is apparently some misunderstanding regarding the word thieve (and liar): Some people seem to believe that this is meant as an insult which is of course not true. " [11:24] <wm4> ahahaha [11:24] <wm4> that's hilarious [11:29] <ubitux> wm4: oh no 2011 again, please noo :( [11:30] <ubitux> you're giving him the opportunity to rediscuss all over again the same story [11:31] <wm4> oops [11:31] <ubitux> http://pastebin.com/6qZDfjv0 wat [11:32] <nevcairiel> windows batch files are fun [11:34] <nevcairiel> can someone just make the ticket go away now [11:37] <nevcairiel> on more constructive things, is the curl api just terrible or am I going crazy trying to understand it [11:47] <cbsrobot> ubitux: just tell him to tell you a number before you answer with a yes or no [11:50] <wm4> nevcairiel: AFAIK it's very terrible, yes [12:11] <wm4> you know what would be nice? if there was something occasional contributors could run FATE with their patches applied [12:12] <wm4> because keeping around about 1 GB of FATE samples and running the suite locally is a bit inconvenient [12:15] <J_Darnley> Make them available via http? [12:16] <wm4> sounds potentially helpful [12:17] <nevcairiel> what does that change [12:17] <J_Darnley> They don't have to keep the samples. They just have to d/l 1G everytime they run it! [12:18] <nevcairiel> and thats different to now, how? [12:18] <wm4> they could get the samples for the tests they want to run, or so [12:19] <J_Darnley> you can do that with rsync too [12:23] <wm4> is it right that AVInputFormat.read_packet frees the packet passed as parameter with av_free on error? [12:24] <wm4> this doesn't seem right at all [12:38] <cone-391> ffmpeg.git 03Michael Niedermayer 07master:0daff3ce7632: avfilter/vf_mp: remove incorrect usage of AVFrame.type [12:38] <cone-391> ffmpeg.git 03Michael Niedermayer 07master:19bf1ed1f426: avformat/id3v2: Fix "warning: unused variable uncompressed_buffer_size" if zlib is unavailable [13:04] <Daemon404> pretty amusing the chromium guy asks if ffmpeg supports pre-c99 compilers [13:05] <Daemon404> while using a pre-c99 msvc [13:05] <Daemon404> (unless they upgarded finally) [13:06] <wm4> gah [13:07] <wm4> j-b: not sure if you're interested in that at all, but I know someone who wants to write a POSIX layer on top of native windows (and that doesn't suck to hell and back as cygwin does) [13:07] <Daemon404> sounds doomed to fail [13:07] <Daemon404> (yeah ill just write it... wait...) [13:07] <wm4> he actually got fork() to work, so I'm not sure about that [13:08] <nevcairiel> you can emulate fork, but you cant get fork "to work" :P [13:08] <wm4> because apparently the windows kernel supports fork nativelyx [13:08] <nevcairiel> even if thats true, going that low is a recipe for disaster [13:08] <Daemon404> inb4 insanely unportable code [13:08] <wm4> nevcairiel: why? [13:08] <BBB> omg diego [13:09] <BBB> why is he so lastwordstands [13:09] <Daemon404> you used to work with GNOME [13:09] <Daemon404> you tell me. [13:09] <BBB> I left, that was probably the biggest reason [13:09] <BBB> it was impossible to have any discussion [13:09] <BBB> impossible [13:10] <BBB> all these 15-year olds that think that just because you repeated yourself once more than the other, youve won the discussion [13:10] <BBB> diego really is a 15-year old [13:14] <nevcairiel> wm4: mostly because nothing any levels higher expects you to do it, so if you use any code that you don't control yourself, who knows what weird issues it'll cause [13:15] <wm4> nevcairiel: that's surely true for fork, but unix has the same issue with fork specifically [13:16] <nevcairiel> but fork on unix is a standard thing, isnt it [13:16] <wm4> if something uses threads in the same process, using fork is already sketchy (and unless you exec right after fork, definitely buggy) [13:17] <nevcairiel> anyhow, trying to use kernel functions that noone else otherwise uses is a good way to find problems :D [13:17] <wm4> many win32 functions are just small wrappers around the kernel functions, and drivers use the same functions directly [13:18] <wm4> but it gives you the chance to work around entirely broken crap like the winsock API [13:18] <Daemon404> is stability even guaranteed? [13:18] <Daemon404> i doubt it. [13:19] <wm4> it probably is [13:19] <wm4> compatibility goes back to win2k and earlier [13:19] <nevcairiel> windows doesnt usually break API or ABI at all [13:19] <wm4> so it'd be strange if they changed that [13:22] <kierank> I give up. Both sides have their heads so firmly up their arses [13:23] <wm4> sad [13:30] <nevcairiel> i should stop trying to post on trac, its just impossible with this person around [13:35] <wm4> what is happening to the timestamps I set on a packet in read_packet? [13:35] <wm4> ffprobe -show_packets is reporting "pts=N/A" [13:37] <cone-391> ffmpeg.git 03Michael Niedermayer 07master:97cebf3139cc: avformat/udp: Move variables used only with HAVE_PTHREAD_CANCEL, under the #if [13:37] <cone-391> ffmpeg.git 03Michael Niedermayer 07master:e16b7338d877: avcodec/aarch64/h264qpel_init_aarch64: mark src as const [13:40] <BBB> kierank: I actually feel for you [13:41] <nevcairiel> wm4: did you set a timebase [13:41] <wm4> nevcairiel: yes [13:44] <durandal_1707> what is error code 139? [13:45] <wm4> durandal_1707: use the function that converts error codes to strings? [13:45] <Daemon404> [12:22] <+kierank> I give up. Both sides have their heads so firmly up their arses <-- great, you can join me in the 0 fucks corner [13:45] <Daemon404> there's lots of room. [13:46] <JEEB> aye [13:53] <durandal_1707> what is wrong with hue relicensing? [13:56] <cone-391> ffmpeg.git 03Michael Niedermayer 07master:140287360110: avcodec/vp3data: use more compact data types [14:13] <cbsrobot> is there a prores sample with yuv444p10 (without alpha) ? [14:13] <cbsrobot> i thought it is either yuva444p10 or yuv422p10. [14:15] <cbsrobot> durandal11707: j-b rised some questions tonight, but I think vf_hue was just an arbitrary choice. [14:18] <kurosu_> cbsrobot, not sure but the format and some software seem to handle it [14:18] <kurosu_> http://trac.ffmpeg.org/ticket/3846 [14:21] <kurosu_> the file is converted to yuva444p10 but because the profile was left as is, the alpha wasn't encoded [14:21] <kurosu_> the 2 tested software didn't seem to complain or decode it incorrectly, of course besides the alpha [14:21] <kurosu_> as for an actual sample, I don't know - I've just discovered prores with that ticket [14:21] <kurosu_> (or rather another one about a segfault) [14:22] <wm4> so how the fuck do you compile ffmpeg in a way to make it debuggable? since ffmpeg doesn't compile with -O0 [14:22] <kurosu_> cbsrobot, fate seems to contain some tests from actual files, eg prores/Sequence_1-Apple_ProRes_with_Alpha.mov [14:23] <nevcairiel> wm4: if you have a recent gcc, does Og work? [14:24] <wm4> CLAGS=-Og ./configure ? [14:24] <wm4> *CFLAGS [14:24] <nevcairiel> --extra-cflags maybe? no idea [14:27] <cone-391> ffmpeg.git 03Christophe Gisquet 07master:a808733675cc: proresenc_ks: allow auto-selecting profile [14:27] <kurosu__> not sure if any of my messages went through, so there's (at least) one fate sample with alpha: prores/Sequence_1-Apple_ProRes_with_Alpha.mov [14:28] <kurosu__> software don't seem to mind 444 content encoded with an incorrect profile [14:28] <kurosu__> cf ticket: http://trac.ffmpeg.org/ticket/3846 [14:32] <wm4> god I'm dumb [14:32] <wm4> of course av_new_packet() overwrites all packet fields [14:33] <cbsrobot> kurosu__: but Sequence_1-Apple_ProRes_with_Alpha.mov seems to be yuva444p10le [14:34] <kurosu_> sorry [14:34] <cbsrobot> I was asking for yuv444p10 (without alpha). [14:34] <kurosu_> I somehow managed to warp your question meaning meanwhile [14:34] <kurosu_> I don't know of samples [14:35] <kurosu_> the ticket seems to show software to handle it [14:35] <kurosu_> *that handle it [14:35] <kurosu_> I'm not sure that spec-compliant, but... [14:40] <cbsrobot> kurosu_: it seems apple released a new prores whitepaper: https://www.apple.com/final-cut-pro/docs/Apple_ProRes_White_Paper.pdf [14:43] <kurosu_> probably each plane sampling could be checked to force 4444 profile (which is suggested to support lack of alpha) [14:43] <kurosu_> "In addition to supporting YC B C R or RGB 4:4:4 pixel data, the Apple ProRes 4444 XQ [14:43] <kurosu_> and Apple ProRes 4444 codec types [...]" [14:47] <cone-391> ffmpeg.git 03Carl Eugen Hoyos 07master:494cbc4238db: ffmpeg: Clean up if filter initialisation failed to avoid a memleak. [14:47] <cbsrobot> and: Like Apple ProRes 4444 XQ and Apple ProRes 4444, all Apple ProRes 422 codecs can in fact accept image samples even greater than 10 bits [14:47] <cone-391> ffmpeg.git 03Carl Eugen Hoyos 07master:f22c24bd7a57: lavf/rtpdec_hevc: Fix compilation with -DDEBUG. [14:47] <cone-391> ffmpeg.git 03Michael Niedermayer 07master:ce36d8088113: Merge remote-tracking branch 'cehoyos/master' [14:53] <kurosu_> cbsrobot, yeah, that's about bitdepth, but "chroma" sampling seems to imply one or the other [14:54] <cbsrobot> it's more like a sidenote to me (and anyone interested) [15:09] <ubitux> < wm4> so how the fuck do you compile ffmpeg in a way to make it debuggable? since ffmpeg doesn't compile with -O0 // ffmpeg doesn't compile with --disable-optimizations? [15:09] <wm4> ubitux: ok [15:39] <nevcairiel> ubitux: it does, but thats still O1, which makes debugging not as easy as it could be [15:39] <nevcairiel> which is why i'm wondering if Og works, which is only available in GCC 4.8+, but supposedly still optimizes a bit, but specifically designed for debugging [16:33] <kurosu_> regarding the x264asm patch by bugmaster, yeah, he seemed interested in adding this when I mentioned that in #x264dev [16:33] <kurosu_> I didn't get to test his WIP at the time, though [16:33] <cone-391> ffmpeg.git 03Reimar Döffinger 07master:e2cd28c9265e: fft: add missing const. [17:00] <cone-391> ffmpeg.git 03wm4 07master:b173f5c15572: oggdec: fix invalid free on error [17:51] <cone-391> ffmpeg.git 03Paul B Mahol 07master:bd6f14582077: avformat/dfa: use avio_feof() [18:03] <cone-391> ffmpeg.git 03Reimar Döffinger 07master:8c63a0d17114: xv.c: Add missing const to lookup table. [18:52] <wm4> illegal file ready [18:52] <wm4> sounds like a great marketing slogan [18:53] <nevcairiel> What's the point of that ticket [20:36] <cone-391> ffmpeg.git 03Luca Barbato 07release/1.1:694b7cd873f8: mpegts: Define the section length with a constant [20:36] <cone-391> ffmpeg.git 03Luca Barbato 07release/1.1:addbaf134836: mpegts: Do not try to write a PMT larger than SECTION_SIZE [20:36] <cone-391> ffmpeg.git 03Michael Niedermayer 07release/1.1:a507fea70769: Merge commit '694b7cd873f8b06af109036eff1ccd741afdd28e' into release/1.1 [20:36] <cone-391> ffmpeg.git 03Michael Niedermayer 07release/1.1:807b7388402e: Merge commit 'addbaf134836aea4e14f73add8c6d753a1373257' into release/1.1 [20:45] <cone-391> ffmpeg.git 03Michael Niedermayer 07release/1.1:437848e37ae7: vp3: Copy all 3 frames for thread updates [20:45] <cone-391> ffmpeg.git 03Reinhard Tartler 07release/1.1:8da037af3327: Update Changelog for v9.15 [20:45] <cone-391> ffmpeg.git 03Reinhard Tartler 07release/1.1:e86074e6ef23: Prepare for 9.15 Release [20:45] <cone-391> ffmpeg.git 03Reinhard Tartler 07release/1.1:bd41211395fd: Re-release 9.15 as 9.16 [20:45] <cone-391> ffmpeg.git 03Michael Niedermayer 07release/1.1:dcce698fd887: Merge commit '437848e37ae7ef73cd8101031dc570d1f009ffd5' into release/1.1 [20:45] <cone-391> ffmpeg.git 03Michael Niedermayer 07release/1.1:f953c3c2343d: Merge commit 'bd41211395fd1f968e9f3a4746daffebea60f41e' into release/1.1 [20:57] <cone-391> ffmpeg.git 03Michael Niedermayer 07release/1.1:e4fb53c73abe: ffv1dec: check that global parameters do not change in version 0/1 [20:57] <cone-391> ffmpeg.git 03Anton Khirnov 07release/1.1:bbd632082b18: mpegenc: limit the maximum muxrate [20:57] <cone-391> ffmpeg.git 03Michael Niedermayer 07release/1.1:081f4f5f56e0: Merge commit 'e4fb53c73abece15a7c5df0019df9a0371db2297' into release/1.1 [20:57] <cone-391> ffmpeg.git 03Michael Niedermayer 07release/1.1:0a6d39791165: Merge commit 'bbd632082b18e6c5ce9c2d6be8bc260c05ae9417' into release/1.1 [21:08] <cone-391> ffmpeg.git 03Anton Khirnov 07release/1.1:8d7839fc7c52: avconv: fix the muxrate values for -target [21:08] <cone-391> ffmpeg.git 03Anton Khirnov 07release/1.1:e1f0c41e1aa3: avconv: fix parsing the AVOptions for -target [21:08] <cone-391> ffmpeg.git 03Michael Niedermayer 07release/1.1:fed28fe054de: Merge commit '8d7839fc7c52574dfc22db0181b1cef9cb929910' into release/1.1 [21:08] <cone-391> ffmpeg.git 03Michael Niedermayer 07release/1.1:946a106995bc: Merge commit 'e1f0c41e1aa37a9c166c43abf1b526c796ed7649' into release/1.1 [21:13] <cone-391> ffmpeg.git 03Luca Barbato 07release/1.1:124ec8b1303d: pulse: Add a wallclock option to be compatible with other other captures [21:13] <cone-391> ffmpeg.git 03Michael Niedermayer 07release/1.1:d03fd2c8052e: Merge commit '124ec8b1303d4f29b833099ce9008e31ac6d7c86' into release/1.1 [21:17] <Timothy_Gu> michaelni: are you going to release another 1.1 soon? [21:34] <michaelni> Timothy_Gu, i planed to, yes [21:43] <Timothy_Gu> I have a list of stuff to back port: https://gist.github.com/TimothyGu/f69b74bb24eb15e2d005 Probably not all of them apply to 1.1 but you can pick and choose. [21:44] <cone-391> ffmpeg.git 03Michael Niedermayer 07release/1.1:5865d599c388: avcodec/iff: check pixfmt for rgb8 / rgbn [21:44] <cone-391> ffmpeg.git 03Christophe Gisquet 07release/1.1:11a61dd0e2b5: proresenc_kostya: report buffer overflow [21:44] <cone-391> ffmpeg.git 03Michael Niedermayer 07release/1.1:c58d7f9eb56a: avcodec: fix aac/ac3 parser bitstream buffer size [21:44] <cone-391> ffmpeg.git 03Michael Niedermayer 07release/1.1:0bf0de718524: avcodec/utils: add GBRP16 to avcodec_align_dimensions2() [21:44] <cone-391> ffmpeg.git 03Michael Niedermayer 07release/1.1:96d1a8f0145b: avcodec/snow: check coeffs for validity [21:51] <cone-391> ffmpeg.git 03Michael Niedermayer 07release/1.1:195fcbff2b64: Update for 1.1.14 [21:51] <cone-391> ffmpeg.git 03Michael Niedermayer 07release/1.1:4ede955d864d: remove VERSION file [21:51] <cone-391> ffmpeg.git 03Michael Niedermayer 07release/1.1:3ed4dc92284c: version.sh: Print versions based on the last git tag for release branches [22:09] <ubitux> we should disable completely the markup in trac [22:10] <ubitux> users are completely clueless about it [22:10] <ubitux> having every message in a <pre> with some basic regex to match url & hashes would be more than enough [22:11] <ubitux> or should we patch ffmpeg to print {{{ and the beginning and }}} at the end? :P [22:11] <cone-391> ffmpeg.git 03Jon Morley 07release/1.1:552fe9b07f6b: avcodec/adpcm: Fix incorrect AVSampleFormat for sample_fmts_s16p [22:11] <cone-391> ffmpeg.git 03Christophe Gisquet 07release/1.1:3231e7ab64ef: wavpack: report if there is no bits left [22:12] <cone-391> ffmpeg.git 03Christophe Gisquet 07release/1.1:7f7cf051edd3: alacenc: increase predictor buffer [22:12] <cone-391> ffmpeg.git 03Christophe Gisquet 07release/1.1:60f94f708414: alacenc: fix extra bits extraction [22:12] <cone-391> ffmpeg.git 03Piotr Bandurski 07release/1.1:f3c8a8b087d0: avcodec/lcldec: fix decoding of YUV444 sample [22:17] <michaelni> Timothy_Gu, backported what applied cleanly and didnt break compilation. In the future i suggest you backport the commits to release/x.y and ask me to merge/pull, thats safer than backporting what applies cleanly as ommited commits could have side effects though thats probably not common [22:17] <michaelni> also if someone wants to maintain old release branches or a specific old release branch [22:17] <michaelni> would be welcome ... [22:18] <michaelni> my maintaince of them is kind of only "make release if there are some new commits somewhere like locally or from libav or from elsewhere" [22:19] <michaelni> and backporting critical security issues of course [23:24] <cone-391> ffmpeg.git 03Reimar Döffinger 07fatal: ambiguous argument 'refs/tags/n1.1.14': unknown revision or path not in the working tree. [23:24] <cone-391> Use '--' to separate paths from revisions [23:24] <cone-391> refs/tags/n1.1.14:HEAD: xv.c: Add missing const to lookup table. [00:00] --- Sun Aug 31 2014

1 0

ffmpeg.log.20140830
by burek 31 Aug '14

31 Aug '14

[00:00] <ac_slater_> klaxa: huh, I guess specifying two outputs is ok [00:10] <ac_slater_> ls [00:10] <ac_slater_> ;) oops [01:48] <jorb> klaxa: thx for you help .. I did figure out that the only way some of the files would work is via the older Subsonic compiled version of ffmpeg.. where it just threw out the "video" stream for the album cover [01:54] <jorb> pasted the other output .. not sure what release of ffmpeg the "no drop" transcode uses ... http://pastebin.com/Tf875r8E [02:40] <wumzi> Hi. I've been trying to compile ffplay.c only from the latest development version of ffmpeg. I've been trying things like that http://pastebin.com/AbJhnYZx without success. I'm running Ubuntu 14.04 64bits. I'm a bit lost ... Building the whole project with the Ubuntu Compilation Guide works perfectly. Do you have any idea ? [02:54] <editor24> Hey [02:54] <editor24> Anyone here played around with x265 (HEVC)? [02:57] <c_14> editor24: presumably several, what's your question? [02:59] <c_14> wumzi: why not just ./configure [options]; make ffplay ? [03:00] <editor24> c_14: a few questions, actually. I'll start with mcdeint [03:01] <editor24> I'm trying yadif and mcdeint to get from 30i DV to 30p HEVC MP4 [03:01] <editor24> is there a setting you can suggest? I have yadif=1:1,mcdeint=0:1:10 [03:07] <c_14> Hmm, don't do much deinterlacing, but that should work. Anything wrong with it? [03:19] <editor24> c_14: No, it worked well, afaik. I got it from a forum, but without much explanation [03:19] <editor24> Guess I'll keep it [03:20] <editor24> Do you know what framestep=2 will do? [03:21] <editor24> My full command: [03:21] <editor24> ffmpeg.exe -v verbose -i test.avi -threads auto -sws_flags lanczos -vf "yadif=1:1,mcdeint=0:1:10,framestep=2,scale=iw*sar:ih,unsharp=3:3:2.03" -codec:v libx265 -b:v 1500k -minrate 1k -maxrate 128k -flags +loop -me_method umh -g 250 -qcomp 0.6 -qmin 01 -qmax 81 -qdiff 4 -bf 16 -b_strategy 4 -i_qfactor 0.81 -af "aresample=44100" -codec:alibvo_aacenc -b:a 128k -ac 2 -y test.mp4 [03:21] <c_14> afaik it will drop every other frame [03:22] <c_14> or duplicate every frame [03:22] <c_14> ehhh [03:22] <c_14> https://ffmpeg.org/ffmpeg-filters.html#framestep [03:22] <c_14> If whatever it says there is of any help to you. [03:23] <c_14> You might not want to use libvo_aacenc, the native encoder is probably better [03:23] <editor24> okay [03:23] <c_14> See: https://trac.ffmpeg.org/wiki/Encode/AAC [03:24] <editor24> what is it..? [03:24] <c_14> An aac encoder. [03:24] <editor24> i mean, how do i put it in the command line? [03:24] <c_14> use -c:a aac -strict -2 instead of -c:a libvo_aacenc [03:25] <editor24> thanks. what's -strict -2? [03:26] <c_14> The built in aac encoder is marked as experimental (although it's pretty stable) so you need to enable experimental encoders. [03:27] <editor24> oh ok [03:27] <editor24> any other thoughts about the command i pasted? i just copied it ... [03:30] <c_14> Nothing major, but a lot of those options are very specific. What you might want to do is adjust the video and audio bitrates. [03:31] <c_14> The verbose loglevel might just spam you, and not sure if -threads auto does anything for x265 [03:31] <c_14> You'll also only need the audio filter if the sampling rate isn't already 44.1KHz (if you even want a sampling rate of 44.1KHz) [03:32] <c_14> And if you want more than 2 audio channels (or less) you'll need to adjust the -ac option [03:32] <c_14> You'll probably also want to get rid of the -y (so you don't accidentally overwrite existing files) [03:34] <editor24> How do I add a mode to mcdeint? The FFmpeg documentation says I can set it to fast, medium, slow, extra_slow. [03:35] <c_14> mcdeint=mode=[mode]:parity=[parity]:qp=[quantization] [03:36] <c_14> If you don't name the arguments, they are assumed to be in that order. [03:36] <c_14> The {fast,medium,slow,extra_slow} names are probably mapped to numbers, but I don't know how. [03:36] <c_14> In any way, if you want to force a setting just use the name. [04:26] <editor24> hello [04:27] <editor24> how do i change the mcdeint speed to extra_slow? [06:08] <rule_2> I am trying to test capture functionality with a webcam before I use ffmpeg to record from an HDMI input card I plan on getting [06:09] <rule_2> I tried recording from a webcam but I get unknown input format: 'vl42' [06:09] <rule_2> if I run ffmpeg -list_formats all -f v4l2 -i /dev/video0 I get Unrecognized option 'list_formats [06:10] <rule_2> am I looking at outdated documentation to try and use ffmpeg? [06:22] <chriys> hi all [06:23] <chriys> I'm looking to setup HLS live streaming from my camera. I'll send a signal using FMLE or Wirecast. The server will be on a vps (Ubuntu or Debian). Do you have a link for the steps to advise me ? [07:29] <techtopia> hello [07:29] <techtopia> i run this line [07:29] <chriys> sup techtopia [07:29] <techtopia> ffmpeg.exe -y -i %input01% -sws_flags spline -sn -vf yadif=1:0,crop=%width%:%height%:%left%:%top% -s 704x394 -c:v libx264 -preset slow -crf %crf% -r 50 01.mp4 [07:30] <techtopia> with no -acodec specified because a demux the audio and encode it seperatly [07:30] <techtopia> but the encoded 01.mp4 still has the source audio file present [07:31] <techtopia> is there soway to specify that i just want to encode the video and forget about the audio [07:31] <techtopia> someway* [07:32] <ghospich> -an [07:32] <ghospich> maybe, if i understood correctly [07:32] <techtopia> thanks, i will do a quick test :) [07:35] <techtopia> it still has the audio track with -an [07:36] <techtopia> ffmpeg.exe -y -i test.ts -sws_flags spline -sn -vf yadif=1:0,crop=720:576:0:0 -s 704x394 -c:v libx264 -preset slow -crf 21 -r 50 -an 01.mp4 [07:43] <ghospich> try something like "-map -0:a:0" [07:44] <ghospich> also, position of argument matter, so try paste -an after -sn, not to the end. [09:07] <vedos> howdy [09:09] <vedos> i'm using youtube-dl to download stuff from Youtube... youtube-dl defaults to avconv, but can also use ffmpeg if i install it. which should i use to dump wav files out of the videos? which is more reliable and gives better sound quality? [10:40] <techtopia> is there a way to use -concat so it will process all files in a directory [10:40] <techtopia> instead of having to create a text file or an elaborate batch file [11:22] <techtopia> http://pastebin.com/6qZDfjv0 [11:22] <techtopia> how can i get this to work [11:22] <techtopia> atm it just asks for a yes or no response instead of asking for a number [12:49] <wodim> please recommend me one of those modern shiny codecs that produce a good quality video in a small file. [12:55] <techtopia> x264 and x265 [12:56] <techtopia> i use libx264 with ffmpeg [12:56] <techtopia> quick encodes, decent filesize, good quality [12:57] <techtopia> x265 will get you a smaller filesize for a comparable quality but takes longer to encode [12:57] <Chaz6> vp8? [12:59] <JEEB> vp8 is spec-wise close to baseline profile H.264, and the encoder is slow [13:00] <JEEB> vp9 is better than high profile H.264 spec-wise but the encoder is even more slow [13:00] <JEEB> (since it doesn't even have multithreading) [13:09] <K4T> btw: better buycpu with more cores or higher frequency for ffmpeg? [13:09] <K4T> I need to stream multiple streams with ffmpeg in real time [13:10] <K4T> some of them I have to reencode [13:56] <techtopia> im enocding from a transport stream and the resulting encodes have this menu text in them http://i.imgur.com/uYwt552.jpg [13:56] <techtopia> this is my line "ffmpeg.exe -y -i 01.ts -sws_flags spline -sn -vf yadif=1:0,crop=720:576:0:0 -s 704x394 -c:v libx264 -preset slow -crf 21 -r 50 aaa.mp4" [13:56] <techtopia> how can i omit the text from the encode [13:57] <techtopia> without remuxing at the end [14:18] <relaxed_> techtopia: add -map 0:v -map 0:a after the input [15:17] <chriys> hi all [15:20] <chriys> how can I make ffmpeg to grab a file sent from a rtmp server? [15:21] <chriys> I'm going to send the feed from FMLE then it will go to red5 (rtmp server) then to ffmpeg, then to ffserver (this one will will HTTP serve the feed)? [15:22] <chriys> I'm also interessed if there another way to grab the feed without passing by the rtmp server. [15:30] <chriys_> how can I make ffmpeg to grab a file sent from a rtmp server? [15:30] <chriys_> I'm going to send the feed from FMLE then it will go to red5 (rtmp server) then to ffmpeg, then to ffserver (this one will will HTTP serve the feed)? [16:15] <c_14> ffmpeg -i rtmp://[server] [foobar] [16:41] <Andy2> Hi guys, are asked this question what seems to be a bug a couple of days ago. http://ffmpeg.org/pipermail/ffmpeg-user/2014-August/023110.html [16:41] <Andy2> I got some initial responses buit now nothing happens anymore. [16:41] <Andy2> Can someone tell me how to proceed? [16:41] <Andy2> Thanks [17:54] <active8> Hi. I'm converting webm to mp4 using -q:v 0 (maybe 1 is the limit - I don't always get a warnig on that) and without any options I get pixelation on the output file. using -q:v 0 -c:v mpeg4 for an output specifier, it's ok. if ffmpeg guesses the codec, why would I need -c:v mpeg4 to get a good conversion. Also wondering if c:v copy is better than -q:v 0 . [17:54] <Fjorgynn> I see [17:55] <Fjorgynn> -c:v means that you don't reencode the file just copy the encodec [17:55] <Fjorgynn> and you should probably use libx264 [17:56] <Fjorgynn> and -crf instead of -q:v 0 maybe [17:56] <Fjorgynn> active8: but what do I know [17:58] <active8> your first stmnt: -c:v means... you mean that -c:v copy means to just copy? does ffmpeg then pick a codec to make the mp4? [17:59] <Fjorgynn> no I think it just copies it to another container [17:59] <Fjorgynn> good if you want to remove video from audio or vice versa [18:00] <active8> so vlc is trying to play an mp4 which is actually a webm stuck in an mp4 container and thus the pixelation? [18:01] <Fjorgynn> ffmpeg -i test.webm -c:a copy -c:v libx264 -crf 22 test.mp4 [18:04] <active8> what is -crf [18:04] <active8> ? [18:05] <active8> looks like frame rate,but i thought that was just -r [18:09] <JEEB> active8, x264 setting, constant rate factor. it is not (since we have no thousands of little brains in our PCs) constant quality level, but it's the closest we have to it [18:09] <JEEB> 23 is the default, and higher values use compress more, lower values compress less. the idea is to find the highest value that still looks good [18:10] <active8> oh. so codecs have options, too. man page didn't turn up anything on that. maybe there's a man (X) ffmpeg that does or a man libx264 [18:12] <JEEB> there was a way to output encoder-specific help with -help but don't ask me how :P [18:12] <JEEB> never used that feature [18:12] <JEEB> http://ffmpeg.org/ffmpeg-all.html contains some of the stuff [18:15] <Fjorgynn> 0~0~ [18:15] <active8> grep didn't work for some reason, but ffmpeg --help | less and then /-crf <enter> took me there. Thanks JEEB and Fjorgynn [18:18] <durandal_1707> ffmpeg -h encoder=libx264 [18:28] <active8> durandal_1707, what system are you using? encoder=libx264 doesn't do anything more here. I just get the same output as ffmpeg -h [18:32] <active8> you want a pastie for what? [18:33] <durandal_1707> output of ffmpeg [18:33] <durandal_1707> anyway you are either not using ffmpeg or using very old version [18:38] <active8> i compiled it maybe 6 mos to a year ago - oh... i was calling help on the one installed from a safe repository. I see the diff now when I run it with encoder=libx264 on the one I compiled myself - ffmpeg version git-2013-08-04-61af627 [18:39] <active8> that's a nice feature. specific help. thanks [19:47] <anshul_mahe> please help me with some alternative of -list_devices true -f dshow to be used when directly using libav library [20:16] <TwisteR> Greetings! I need to embed an ass softsub to mkv file and set subtitle's 'default' flag to 1 [20:17] <TwisteR> here is my command line: ffmpeg -i in.mkv -c:v copy -c:a copy -metadata:s:s:0 language=eng -metadata:s:s:0 default-flag=1 out.mkv [20:17] <TwisteR> tried also this: -metadata:s:s:0 disposition:default=1 [20:18] <klaxa> afaik ffmpeg does not support that yet [20:18] <TwisteR> but ffprobe output tells me that those flags are in some incorrect place: [20:18] <TwisteR> DISPOSITION:default=0 [20:19] <TwisteR> TAG:DISPOSITION:DEFAULT=1 [20:19] <klaxa> https://trac.ffmpeg.org/ticket/2402 [20:19] <TwisteR> klaxa: yep, that is my case, thanks [20:20] <klaxa> you can try mkvmerge or mkvmerge-gui from mkvtoolnix instead [20:20] <TwisteR> klaxa: thanks, I'll try them [21:00] <tfirg_> i have a trivial and probably common question: [21:01] <tfirg_> i have a ffmpeg x11grab line that cooks my laptop, and i am looking for ways to make it less intensive on my system while preserving reasonable quality [21:01] <tfirg_> ffmpeg -f alsa -ac 2 -i pulse -f x11grab -r 30 -s 1366x768 -i :0.0 -acodec pcm_s16le -vcodec libx264 -preset ultrafast -crf 0 -threads 0 ${1}.mkv [21:02] <tfirg_> that is what i have [21:02] <tfirg_> so i tried less fps (15) and preset slow [21:02] <tfirg_> but that still boils my cpu [21:02] <c_14> That's cooking your laptop? [21:02] <c_14> preset slow will make it use more cpu, btw [21:03] <tfirg_> o ok [21:03] <tfirg_> thats my i came here [21:03] <tfirg_> for suggestions [21:03] <c_14> If you have hard drive space to spare, you could just use rawvideo. [21:03] <tfirg_> i want something that is light but still reasonable [21:04] <tfirg_> but that needs to be processed sooner or later anyway [21:04] <tfirg_> i have a 180 gig ssd in there but i was hoping for other suggestions [21:04] <tfirg_> you know that example in man ffmpeg is nice an light [21:04] <tfirg_> but its limited [21:05] <tfirg_> no sound and i cant get it to do my resolution [21:05] <tfirg_> so i was hoping for a middle ground [21:07] <c_14> Maybe try one of the other lossless video codecs? [21:07] <c_14> ffmpeg -codecs | grep ' .EV..S' [21:10] <tfirg_> thanks, ill try, in the meantime if there are other suggestions, please let me know [21:35] <klaxa> <c_14> If you have hard drive space to spare, you could just use rawvideo. [21:35] <klaxa> that will pretty much fuck up I/O though [22:11] <pixaal> hey folks - i'm converting an image sequence to a video (with audio as well) but it seems to ignore the '-r 24' framerate [22:12] <pixaal> the audio is 5m 11s long, and there are 7485 frames which would make 5m 11s at 24fps, but the video finishes just before 5m while the audio carries on to 5:11 [22:16] <pixaal> the command is: ffmpeg -i "02\img_%%04d.png" -i "02\audio.wav" -c:v libvpx -b:v 3M -c:a libvorbis -r 24 02.webm [22:16] <pixaal> (double % to escape the % in a bat script on windows) [22:20] <c_14> https://trac.ffmpeg.org/wiki/Create%20a%20video%20slideshow%20from%20images… [22:21] <c_14> Try making your command look more like that (just with the added audio track and the codec/bitrate settings) [22:21] <pixaal> will try, thanks [23:26] <relaxed> pixaal: by default ffmpeg uses 25 fps with images as input- put -r 24 before -i "02\img_%%04d.png" [23:27] Action: pixaal tries [23:45] <pixaal> thanks relaxed, worked fine. i swear i tried that... [23:45] <pixaal> o/ [00:00] --- Sun Aug 31 2014

1 0

ffmpeg-devel.log.20140829
by burek 30 Aug '14

30 Aug '14

[00:48] <llogan> i'm surprised that there was any response at all [01:09] <cone-123> ffmpeg.git 03Vittorio Giovara 07master:4d686fb721b4: matroskaenc: convert avstream stereo3d side data during encoding [01:09] <cone-123> ffmpeg.git 03Michael Niedermayer 07master:39cd9fdbf802: Merge commit '4d686fb721b485ebbc4c7779d927d876c1e630f7' [01:09] <Compn> cant say we didnt try [01:19] <cone-123> ffmpeg.git 03Diego Biurrun 07master:9e8bbe7d4d1d: license: Mention that vf_interlace is GPL, not LGPL [01:19] <cone-123> ffmpeg.git 03Michael Niedermayer 07master:7771daaea7d5: Merge commit '9e8bbe7d4d1dcd5fec491dbfbb98ed2038a7bed5' [01:39] <cone-123> ffmpeg.git 03Michael Niedermayer 07master:37520a91ace8: avformat: drop redundant MATROSKA_VIDEO_STEREO_MODE_COUNT identifier [01:39] <cone-123> ffmpeg.git 03Michael Niedermayer 07master:22652dc2b8d6: avcodec/avcodec: fix missing doxygen comment marker [04:29] <cone-123> ffmpeg.git 03Vignesh Venkatasubramanian 07master:4c9204783ad2: fate: Add basic tests for WebM Dash Manifest [05:58] <pross-au> cehoyos: why label #3898 as 'fails to play broken file' [07:35] <ubitux> kierank: no one mentioned 2011 yet [07:57] <ubitux> Compn: the fact that concat is seemingly simpler in mencoder is not the reason people stick with it [07:57] <ubitux> Compn: the lack of some features might be [08:15] <wm4> ubitux: "seemingly"? [08:58] <ubitux> wm4: replace with "limited and" [08:58] <plepere> good morning all [09:50] <Guest61930> hi, i am using ffmpeg to get Hardware encodded H264 video. i am using this in such way " ffmpeg -f v4l2 -i /dev/video0 -map 0 -c:v copy -an -b:v 64k out.mkv. The output video format is undf and it is availble with overshooted bitrate upto 144Mbps app. can any one help me out get the Hardware encoded H264 video? [10:18] <plepere> Guest61930, try asking in #ffmpeg [10:54] <ubitux> oh ffs koda asking for meeting irl as if it will make any difference [12:40] <cone-881> ffmpeg.git 03Paul B Mahol 07master:be3d8073ed83: avcodec/wavpack: increase WV_MAX_SAMPLES [12:47] <wm4> wow pgs looks pretty simple [12:52] <ubitux> wm4: you mean "sup"? :P [12:52] <wm4> yes [12:52] <ubitux> yes, less insane than vobsub [12:52] <ubitux> at least you don't need another index [12:53] <ubitux> but you can have only one stream afaict [12:53] <ubitux> it's almost a raw muxer [12:54] <Compn> dont worry, i'm sure sony made it insane in some ways :P [13:10] <cone-881> ffmpeg.git 03Michael Niedermayer 07master:5ce98e774eea: fate/vpx: make webm dash manifest tests depend on the WEBM_DASH_MANIFEST demuxer [13:10] <cone-881> ffmpeg.git 03Michael Niedermayer 07master:72732f2dddab: avformat/webmdashenc: use av_strlcpy() and allocate enough space [13:13] <Compn> ubitux breaking out the logic arguments :D [13:13] <Compn> re mailing list discussions [13:44] <ubitux> wm4: i know diego didn't keep ffmpeg in cc, but please make sure you keep both ml when replying [13:44] <wm4> didn't pay attention to that [13:44] <ubitux> yep, i know :) [13:45] <ubitux> that mail was targetted @ libav anyway [13:51] <wm4> btw. what does Libav use instead of lena.pnm? [13:51] <ubitux> Libav debian package has the issue [13:51] <ubitux> it was just not raised [13:53] <wm4> hurr [13:57] <ubitux> i can't find the mail of the guy @ http://blog.ricbit.com/2009/11/lena-e-ilena.html [13:57] <ubitux> can anyone confirm? [13:57] <ubitux> i'll probably contact directly http://www.ilafox.com/2010/04/contato.html to get a "raw" version otherwise [14:03] <ubitux> wm4: it seems you're making some friends& [14:14] <cone-881> ffmpeg.git 03Peter Ross 07master:9b8eedd736ea: avformat/wtvdec: seek over broken chunks [14:33] <kierank> i like the way libav people stop ccing ffmpeg-devel [14:33] <kierank> well diego [14:34] <ubitux> kierank: might not be malice, really [14:35] <wm4> it's good that you try to interpret it not in that way [14:38] <wm4> what's the best way to create a lavfi filter graph with an optional additional output pad? [14:38] <wm4> i.e. I don't know whether the graph will have that output or not [14:45] <__gb__> ubitux, michaelni: have you considered using one of those images from fotolia for a lena replacement? that used to be free but I am not so sure nowadays [14:45] <ubitux> i'd personally like to keep a lena.pnm file, because we might have a few references [14:46] <ubitux> like typically in the bug tracker [14:46] <ubitux> also, it has a special symbolic [14:46] <ubitux> that's why i liked the idea of using the "free version of lena" from the blog post [14:46] <__gb__> yes, but if the fotolia conditions are ok, there could be some girly face in there, isn't it? [14:47] <ubitux> i don't know fotolia [14:48] <__gb__> or someone could find lena of today and ask for shooting her :) [14:49] <__gb__> shooting, as in "take a photo" of course, hey [14:49] <ubitux> https://www.flickr.com/search/?text=lena&sort=relevance&license=4%2C5 [14:49] <durandal_1707> why not make selfie... [14:50] <ubitux> there are tons of possibilities really [14:52] <__gb__> but the difficulty is to find something with the same characteristics, or having debian just decide on the way they prefer? [14:52] <ubitux> decide on something [14:52] <ubitux> there are various prosposal [15:04] <Daemon404> btw, asm dudes (BBB, ubitux, jamarial (not here)), bugmaster posted a useful patch fo x86inc you may want to sync once pushed [15:04] <Daemon404> it detects instructions used in the wrong cpu level [15:04] <Daemon404> e.g. mmx2 in mmx [15:04] <ubitux> ooh [15:04] <ubitux> i saw that discussion, wasn't kurosu talking about that? [15:04] <Daemon404> maybe [15:05] <Daemon404> i just saw a patch materialize yesterday [15:05] <ubitux> it sounds pretty cool [15:05] <ubitux> Daemon404: url? [15:05] <plepere> oh, great news ! [15:05] <Daemon404> [19:09] < BugMaster> fionag: Gramner: http://privatepaste.com/d30aa7a940 patches can squashed together [15:05] <Daemon404> CAUTION: unmerged and unreviewed [15:05] <Daemon404> so dont go pushing it to ffmpeg ;) [15:06] <ubitux> would be nice to make it possible to error out [15:07] <ubitux> instead of just warn [15:07] <Daemon404> should be easy enough [15:07] <Daemon404> you should tell bugmaster / fionag [15:07] <nevcairiel> Warnings help enough really, since the developer will see it [15:07] <ubitux> i'm assuming it might not be reliable enough so it's kept as warning [15:07] <Daemon404> nevcairiel, not in the wasteland of ffmpeg warnings [15:07] <Daemon404> ubitux, maybe [15:07] <ubitux> but yeah really, cool stuff [15:08] <ubitux> i wonder if it will detect issues :) [15:08] <Daemon404> my money is on yes [15:08] <nevcairiel> Probably some sse2 in SSE, not like anyone really tests sse-only hardware anymore [15:09] <ubitux> we have such box afaik but it probably doesn't cover every asm ofc [15:10] <Daemon404> nevcairiel, i used to have an atom [15:10] <Daemon404> which didnt even have sse [15:10] <Daemon404> clang couldnt even build for it [15:10] <Daemon404> :D [15:10] <Daemon404> er [15:10] <Daemon404> nto atom [15:10] <Daemon404> geode [15:21] <J_D> I tested that patch from bugmaster on ffmpeg here and it built without issue [15:22] <J_D> and I think the commit message is slightly wrong [15:23] <J_D> it uses the %error feature of the preprocessor [15:23] <Daemon404> ah [15:23] <J_D> Perhaps I should test it to make sure it is a fatal error [15:29] <J_D> Oh it isn't [15:30] <J_D> Why the eck is it called %error then? [15:30] <J_D> *heck [15:31] <Daemon404> hell if i know [15:31] <nevcairiel> It should be [15:32] <nevcairiel> Although they may have %fatal if it isn't [15:33] <nevcairiel> At least nasm has thay [15:33] <nevcairiel> that* [15:36] <J_Darnley> No it does not have %fatal [15:38] <J_Darnley> The comments about using %error in the yasm docs are then just inaccurate. [15:38] <J_Darnley> "any user who fails to understand the way your code is supposed to be assembled will be quickly warned of their mistake, rather than having to wait until the program crashes" [15:39] <J_Darnley> Not so much in this case [15:39] <J_Darnley> It will be assembled just fine until you run into an illegal instruction. [15:39] <J_Darnley> ... during execution [15:40] <J_Darnley> Anyway. The patch is definitely useful for people adding or porting code [15:53] <saste> anybody willing to review the avstring/utf8 decoding patch? [15:53] <saste> otherwise i'll probably push it in a day [15:58] <kierank> 2:05 PM <"Daemon404> CAUTION: unmerged and unreviewed [15:58] <kierank> 2:05 PM <"Daemon404> so dont go pushing it to ffmpeg ;) [15:58] <kierank> LOL [15:58] <J_Darnley> If I knew anything about text encoding I might. Sorry. [16:11] <nevcairiel> J_Darnley: the documentation does say warned, not fail [16:11] <nevcairiel> In any case, developers should read their warnings ;) [16:12] <nevcairiel> Easy enough to grep through a build log now to check too [16:15] <J_Darnley> ffmpeg doesn't produce any new warning when using the patch. [16:44] <ubitux> why Diego doesn't want to update the message with "The ffmpeg program has been deprecated *in the Libav project*"? [16:44] <ubitux> :( [16:48] <wm4> it's funny because this message actually is the biggest reason why users hate Libav and think it's malicious [16:48] <wm4> debian users, I mean [17:04] <ubitux> haha, he really doesn't want to say "deprecated in the Libav project" and is trying to move it to the next sentence [17:06] <ubitux> kierank: thx [17:11] <iive> ubitux: it kind of sounds like "FFmpeg have deprecated the Libav project" :) one does not want to cause such confusion :P [17:11] <Daemon404> g 31 [17:11] <Daemon404> woops [17:12] Action: ubitux increments the Daemon404 irc fail counter [17:12] <Daemon404> :D [17:15] <ubitux> [~]- grep -iE 'Daemon404> (g )?[0-9]+$' .irssi/irclogs/2014/freenode-\#ffmpeg-devel.log|wc -l [17:15] <ubitux> 12 [17:15] <ubitux> so there is now 12 fails this year on this channel [17:15] <Daemon404> thats inaccurate [17:15] <ubitux> ah? [17:15] <Daemon404> sometimes i type / N [17:15] <Daemon404> so it ends up with just a num [17:16] <ubitux> yes, i handle that [17:16] <ubitux> http://pastie.org/pastes/9513188/text [17:16] <wm4> ubitux stalks ffmpeg devs! [17:16] <ubitux> how evil [17:16] <Daemon404> ah [17:16] <ubitux> i'm stealing what they say [17:17] <Daemon404> maybe it's secretly a code [17:17] <ubitux> Daemon404: 22 times in #libav-devel btw [17:17] <ubitux> this year only too [17:17] <ubitux> you seem to like the 42 window [17:17] <Daemon404> darkhold too [17:17] <Daemon404> 42 is my work irc window. [17:18] <ubitux> how suitable [17:18] <Daemon404> it just happened to work out that way [17:18] <Daemon404> happy coincidence [17:18] <ubitux> 16 on the only darkhold chan i'm on [17:19] <ubitux> so on a limited scope of 3 channels, you're doing about 100 fails a year [17:20] <ubitux> (a bit less) [17:21] <iive> - is that sickle and hammer? aka USSR national emblem. [17:21] <ubitux> yes [17:21] <iive> you are evil! [17:22] <Daemon404> lol [17:22] <ubitux> it becomes (and change color) when the command fails nowadays [17:22] <Daemon404> look at my orgs on github [17:22] <Daemon404> one is CCCP [17:22] <Daemon404> with sickle/hammer [17:22] <Daemon404> ;) [17:27] <iive> ;) [17:29] Action: J_Darnley should get some interesting unicode chars in his terminal [18:10] <Compnn> panasonic using our samples to test its apps :) [18:18] <kierank> Compnn: which app? [18:18] <Compnn> "I'm in charge of testing for Mobile Apps that test multiple codec types." [18:19] <Compnn> no idea [18:59] <cone-449> ffmpeg.git 03Reimar Döffinger 07master:be4aac302b0c: patcheck: check for pointer arrays that are not const. [18:59] <cone-449> ffmpeg.git 03Reimar Döffinger 07master:d9e2aceb7f1c: Add missing "const" all over the place. [19:06] <Compnn> ubitux : and a new troll emerges on deb-dev [19:06] <Compnn> sigh [19:08] <Compnn> well yesterday [19:08] Action: Compnn catches up [19:19] <wm4> link? [19:19] <wm4> not following that flame [19:20] <Compn> https://lists.debian.org/debian-devel/2014/08/msg00914.html [19:20] <wm4> ah [19:21] <Compn> it continues over 3 or more mails [19:21] <Compn> https://lists.debian.org/debian-devel/2014/08/msg00916.html [19:22] <Compn> carl still continues to upset libav. [19:22] <wm4> the first mail is constructive and brings up important issues (although there are worse problems that need to be solved before that) [19:23] <wm4> and the second mail you link, meh. [19:23] <Compn> i'm curious what an ffmpeg tree sans merges would look like [19:23] <Compn> no soc code. no ffmpeg-mt code... [19:23] <wm4> merges are ok, but the libav merges are insane [19:23] <Compn> back to svn days i'd assume. [19:23] <wm4> git wasn'T made for that [19:24] <Compn> i dont think there are any objections to squashing merged code history like that [19:24] <Compn> although it would probably take a long long time to compress... [19:26] <nevcairiel> Carl should probably be forbidden to talk in the name of the project, that would make a bunch of people happier. :P [19:27] <Compn> there is no one that talks in the name of ffmpeg except michael really [19:27] <Compn> reimar , maybe other roots [19:27] <Compn> and mostly michael talks for himself [19:27] <Compn> any talking carl is doing is for himself as well [19:27] <nevcairiel> Then forbidden to talk at all [19:27] <Compn> lol [19:27] <Compn> ffmpeg is not a project that controls its voluntary developers :P [19:27] <Compn> imo [19:28] <Compn> you've seen michael's posts, when he wants a change (bug bounties) he asks on list if other devels agree [19:28] <wm4> the thing is that Libav folks think CE is talking in the name of ffmpeg [19:28] Action: Compn gives up explaining [19:29] <wm4> and since nobody stops him, maybe they're right... [19:29] <Compn> well, you try explain to libav that carl speaks only for himself [19:29] <Compn> good luck... [19:29] <Compn> there isnt much censorship here in ffmpeg-land [19:29] <Compn> i'm glad the project believes in free speech and not banning developers :) [19:31] <wm4> in general, I'd say if the projects are merged, the source tree should probably start out with Libav, and then all ffmpeg changes should be carefully added back [19:32] <wm4> of course that'd be a lot of work [19:32] <kierank> they won't be merged [19:32] <nevcairiel> Sounds like a fun project for a year [19:32] <Compn> whos going to test all of those changes though ? [19:32] <wm4> for some thing it'd be trivial (simple decoders/demuxers), for some almost impossible (changes to mpegvideo and stuff like this) [19:33] <nevcairiel> And every single change reviewed on a ML? :D [19:33] <nevcairiel> It'll take two years [19:33] <wm4> ok I'm giving up the thought.... it's all fucked up [19:34] <Compn> but keep thinking. always good to keep coming up with ideas [19:34] <Compn> the better an idea, the more people will agree. [19:34] <Compn> cant we go into git history and delete all empty merges ? [19:34] <nevcairiel> Not counting new stuff at all, I don't know how big a diff would be even. A lot of changes get reimplemented or ported eventually [19:34] <Compn> i know git wasnt 'built for that' but git is open source and we are open source developers... [19:35] <Compn> nevcairiel : someone posted a ffmpeg v libav diff a while back. [19:35] <Compn> or maybe someone even keeps it up to date, forgot the url [19:35] <kepstin-laptop> rewriting history in git is really annoying, since it changes all the commit hashes [19:35] <Compn> maybe ubitux has it [19:35] <wm4> nevcairiel: my idea was that most common code is actually relatively small compared to the total size of the code [19:35] <wm4> e.g. utils.c is very central, but also just a few kloc [19:35] <Compn> kepstin-laptop : so theres no way to save the hashes and delete the crap ? [19:35] <kepstin-laptop> the hashes are calculated from the contents of the git repo, so changing the git repo changes the hashes [19:36] <kepstin-laptop> i like the merges, they preserve commit authorship and commiter info, including commit dates, and allow you to look up libav git hashes in the ffmpeg repo. [19:36] <wm4> anyway, there's no way Libav folks will give up their "clean" git history (even though the start of it is full of cvs and svn crap) [19:37] <Compn> i didnt realize that was a breaker problem [19:37] <Compn> maybe we could come up with a list of problems that need to be sorted out [19:37] <kepstin-laptop> it's not like it actually causes any issues with git; even fancy stuff like bisect will work properly. [19:37] <nevcairiel> Well have to be a bit careful when bisecting though [19:38] <nevcairiel> When you go into their tree, ffmpeg binary vanishes [19:38] <Compn> wm4 : what about just starting today as ffmpeg git day one (as-in, zero history at all) , and then no one looks at the old git tree? :P [19:39] <wm4> Compn: that would be a problem [19:39] <Compn> bah [19:39] <nevcairiel> That's terrible [19:39] <kepstin-laptop> huh, forgot about that :) [19:40] <kepstin-laptop> in most cases you could just hop over to the nearest ffmpeg merge, but in some cases, the issue might be in the middle of a series of commits in libav. [19:40] <nevcairiel> We have a script that simply skips revisions that don't have the binary. So in the end you would just end on the merge and not the proper commit [19:40] <nevcairiel> But merges contain mostly only one commit these days, so that's not a problem [19:41] <iive> https://lists.debian.org/debian-devel/2014/08/msg00918.html Lovely reply by debian developer :) [19:41] <gnafu> fflibav-ng [19:42] <iive> gnafu: put 2 somewhere :) [19:42] <Compn> again , while i dont speak for ffmpeg, nor would i want to censor anyone. i suggest ffmpeg devs please do not flame on debian-devel list [19:43] <gnafu> iive: f2libav-ng (where "f2" in the logo is stylized like F²). [19:44] <J_Darnley> I was very impressed with the work done to migrate ffmpeg to git. Then I started hating all the merge commits that appeared. [19:44] <iive> gnafu: purrrrfect [19:45] <J_Darnley> "ng"? [19:45] <gnafu> Next Generation [19:45] <gnafu> It used to be a common addition to the name of a fork. [19:45] <J_Darnley> :) [19:46] <iive> startrek-ng [19:47] <nevcairiel> Technically that was tng [19:47] <kepstin-laptop> in the musicbrainz project, we rewrote our tagger program as a "next generation" version, and named the new one "picard" :) [19:48] <gnafu> OOOOOOOOOOoooooooooooohhhhhhhhhhhhhhhhh. [19:48] <gnafu> kepstin-laptop: I never put that together, but that's awesome XD. [19:48] <iive> LoL [19:54] <cbsrobot> I didn't bisect too often, but imho the atmoic merge commits are the way to go if you need to merge [19:54] <nevcairiel> It just makes the history soooo ugly ;) [19:59] <Compn> maybe we should make a new channel for talking about ffmpeg/libav [20:00] <Compn> and leave this one to technical discussions only :P [20:00] <Compn> or maybe just take all the ffmpeg/libav stuff to #libav :P [20:00] <Compn> er #libav-devel [20:00] <durandal_1707> too much weed recently? [20:01] <Compn> i did eat too much the other day and felt woozy for 24 hours durandal_1707. but no i'm good now [20:01] <Compn> just seems like we go on never ending bikesheds here [20:01] <Compn> sometimes :P [20:05] <ubitux> < wm4> in general, I'd say if the projects are merged, the source tree should probably start out with Libav, and then all ffmpeg changes should be carefully added back // or maybe you could git log from 3 years ago and try to review it and fix accordingly, it will be faster and less error prone [20:05] <ubitux> because some merges actually fixed some merged input [20:06] <durandal_1707> libav can do that already [20:06] <wm4> my idea was mostly like: take a diff between ffmpeg and libav, then split that diff into lots of patches (but of course that would lose most of authorship info etc.) [20:07] <ubitux> wm4: erm, you know there are like 100.000 more lines of code? [20:07] <wm4> most of them are in cleanly separate, like individual filters or demuxers etc. [20:08] <wm4> +files [20:08] <Compn> thats still 2+ years of patches to review [20:08] <ubitux> 3+ years [20:08] <ubitux> that won't happen, really [20:08] <wm4> anyway, it's either that, or accept the ffmpeg history [20:08] <wm4> and the latter won't happen [20:08] <Compn> or delete all history :P [20:08] <ubitux> there is no way the whole FFmpeg can be reconstructed completely from this [20:09] <ubitux> and it's an insane waste of time [20:09] <ubitux> IMO [20:09] <Compn> +1 [20:09] <ubitux> try it, i'm pretty you'll abandon after maybe 10-15 patches [20:09] <ubitux> +sure [20:09] <ubitux> out of thousands [20:09] <wm4> :) [20:10] <iive> is there anything of substance that is in libav but not in ffmpeg? [20:10] <wm4> iive: cleaner code [20:10] <durandal_1707> different xbm decoder [20:10] <ubitux> (but with more bugs on common parts?) [20:10] <wm4> ubitux: yes [20:10] <wm4> probably [20:11] <ubitux> what's the point of "cleaner code" then [20:11] <durandal_1707> K&R [20:11] <ubitux> except faping over git history :p [20:11] <iive> LoL [20:11] <durandal_1707> rotfl [20:11] <wm4> ? [20:11] <Compn> faping = masturbating [20:12] <wm4> ask Daemon404 what he thinks of mini's merge commits [20:13] <ubitux> arg again an interview about "code quality" in FFmpeg [20:13] <durandal_1707> what should i reply? [20:14] <ubitux> "go look at Libav, code is prettier" [20:14] <ubitux> i might answer him this week end [20:15] <wm4> of course both look like shit (if you look at the dark corners) [20:16] <wm4> (like with any software) [20:16] <ubitux> yes but it's the fault of FFmpeg [20:16] <ubitux> Libav is cleaning up all this mess fortunately [20:16] <Compn> i blame that gerald launtau guy :P [20:16] <Compn> ehe [20:22] <iive> well, what they claim might not be what they actually do. [20:22] Action: Compn thinks about printing up tshirts 'carl eugen has a posse' [20:23] <ubitux> iive: i hope you understand that was irony :p [20:23] <iive> well, pow law [20:26] <iive> btw, I thought of explaining why carl uses the strongest words against them, given that all involved parties are now ffmpeg developers. But imho, it is kind of disproportional response. [20:47] <ubitux> the main problem is not that he's right or wrong, the problem is that showing off such hostility really doesn't help FFmpeg at all [20:47] <ubitux> but i guess he's aware of that and just think that by principle he must say this [20:48] <ubitux> his answers are used against the project, that's the unfortunate part [21:06] <iive> ubitux: i can turn that around, but it would involve opening old wounds. And IMHO, the fact that 2 people pointed to the same email, just comes to show that there are not that many and diverse insults. [21:07] <wm4> are you still talking about carl? [21:07] <iive> yes [21:08] <wm4> he uses hate speech against libav all the time [21:08] <ubitux> on every occasion yes [21:08] <iive> i know. [21:31] <iive> actions speak more than words. [23:16] <kierank> Daemon404: http://hwcdn.net/j9t9v3v5/cds/ffmpeg.tar.gz [23:16] <kierank> another elemental dump [23:16] <Daemon404> whats different [23:16] <kierank> see your email [23:17] <kierank> it's basically an acknowledgement that the first dump they sent me was bullshit [23:17] <Daemon404> theyre still full of shit [23:17] <Daemon404> as seen in your reply [23:21] <kierank> I still don't see the mxf patches [23:21] <wm4> is this communication public somewhere? [23:21] <kierank> not yet [23:22] <kierank> jesse is the cto [23:34] <kierank> i see nothing mxf related in the dump [23:35] <kierank> +//donn [23:35] <kierank> +//patch for sar issues with avci-50 being backwards (3:4): [23:35] <kierank> hahahah [23:35] <kierank> this is the avc intra [00:00] --- Sat Aug 30 2014

1 0

ffmpeg.log.20140829
by burek 30 Aug '14

30 Aug '14

[00:13] <llogan> lkiesow: i've removed and copyedited some of his stuff. scatterbrained. [00:41] <peshalto> would ffmpeg be the right tool if I want to create a one-frame video from a still image, matching the format of this video? [00:41] <peshalto> http://pastebin.com/KkvAtYVy [00:41] <peshalto> why do you want to create a one-frame video [00:42] <peshalto> I dont, really, but Ill tell you why. Because what I really want to do is prepend one image to a video. [00:42] <peshalto> anyone know how to do that with ffmpeg? [00:42] <c_14> Just concat the image to the video? [00:43] <c_14> https://trac.ffmpeg.org/wiki/How%20to%20concatenate%20(join,%20merge)%20med… [00:43] <peshalto> yes I tried that [00:43] <peshalto> (sort of) [00:44] <peshalto> the input types dont match what I have, and I dont know how to change those command lines to match what I have [00:44] <llogan> you can attempt make a 1 frame movie and then you can use the concat demuxer if you want to attempt to avoid encoding the "main" video, or use the concat filter if you want to encode everything [00:44] <llogan> which do you prefer? [00:44] <peshalto> Id like to avoid decoding/encoding& would prefer just passing stuff through as raw as possible [00:47] <llogan> and you only want one frame? or do you mean you just want the image to appear for a certain duration? [00:48] <peshalto> llogan: the ultimate purpose is to have it be the image that is displayed as a thumbnail by the OS, when browing the file system in a GUI [00:48] <peshalto> so one frame should be enough [00:48] <peshalto> obviously this assumes the OS chooses the first frame as the thumbnail. if it doesnt, Ill have to look for another approach. [00:49] <llogan> the OS automatically chooses the first frame as the thumbnail? that would be boring for most movies. [00:49] <peshalto> remarkably, thats what they seem to do, yes [00:50] <peshalto> anyway it works for my purposes [00:50] <peshalto> can I use ffmpeg for creating a one-frame video from a still? [00:52] <peshalto> using c_14s technique I get the still image for the duration of the video [00:52] <peshalto> and it wont open with QuickTime, although VLC opens it [00:54] <c_14> peshalto: just use the image as the input file with -r 1 and then output to whatever you want [00:54] <peshalto> and the bitrate went from 17886 to 739 kbs [00:55] <peshalto> c_14: just use which command line options exactly? [00:56] <peshalto> Theres also an FAQ entry on this, sort of, but the code produced errors, and when I got past those, a corrupted file [00:56] <peshalto> but then, I probably just dont know what settings to use for that ffprobe output [00:56] <peshalto> (ffprobe output in pastebin) [00:57] <c_14> ffmpeg -r 1 -i image -c:v libx264 -pix_fmt yuv420p -vf 'colormatrix=[srccolormatrix]:bt709,scale=1280x720,fps=119.99' out.mov [00:58] <c_14> something like that should do it or at least get you close [00:58] <c_14> You might not need the scale, and the [srccolormatrix] needs to be replaced with the images colorspace. [00:58] <llogan> and you need audio too, which can be generated with "aevalsrc=0" [01:00] <c_14> Ah, I'm just reading that you can use the scale filter to set the output colormatrix. [01:00] <c_14> https://ffmpeg.org/ffmpeg-filters.html#scale-1 [01:01] <llogan> i didn't know that and i stare at scale all day [01:03] <peshalto> whoa thanks [01:03] <peshalto> where do I put that aevalsrc=0 ? tried putting it after an -i and got a no such file or directory error [01:04] <c_14> either -af or -f lavfi -i [01:06] <peshalto> [Parsed_colormatrix_0 @ 0x7fd049501f20] Unspecified destination color space, [01:06] <peshalto> then [01:06] <peshalto> [AVFilterGraph @ 0x7fd049501c00] Error initializing filter 'colormatrix' with args '' [01:06] <c_14> What's your current complete command? [01:06] <peshalto> error opening filters [01:07] <peshalto> ffmpeg -f lavfi -i "aevalsrc=0" -r 1 -i thumbnail.png -c:v libx264 -pix_fmt yuv420p -vf 'colormatrix=[srccolormatrix]:bt709,scale=1280x720,fps=119.99' out.mov [01:07] <c_14> eh, you were supposed to replace everything in the [] with the source colormatrix of the png [01:08] <peshalto> oh shit where s my input movie& [01:08] <peshalto> is this only for the first step, making the .mov file from the still? [01:08] <c_14> ye [01:09] <Neppy> hmm... any way to get more info on what happened when ffmpeg says "input/output error"? [01:09] <peshalto> where can I get that srccolormatrix? [01:10] <c_14> peshalto: maybe drop the colormatrix filter and just use the out_color_matrix option of the scale filter [01:10] <peshalto> doing man ffprobe now... [01:10] <c_14> Neppy: input/output error is usually os-side [01:10] <Neppy> hm [01:11] <llogan> peshalto: also you'll need "-profile:v high" [01:11] <peshalto> so& vfbt709,scale=out_color_matrix:1280x720 like that? [01:11] <peshalto> + space [01:11] <Neppy> can i have it keep going even when that error happens? [01:13] <c_14> Neppy: that error means that ffmpeg can no longer write output data, continuing would be extremely unproductive [01:14] <peshalto> llogan: in what position? [01:14] <c_14> peshalto: as an output option [01:14] <Neppy> hm.. [01:14] <Neppy> thats annoying ^^" [01:15] <peshalto> no such filter: out_color_matrix [01:15] <peshalto> ffmpeg -f lavfi -i "aevalsrc=0" -r 1 -i thumbnail.png -c:v libx264 -pix_fmt yuv420p -vf 'out_color_matrix,bt709,scale=1280x720,fps=119.99' -profile:v high out.mov [01:15] <Neppy> so the only thing i can do is restart the encoding then? [01:15] <peshalto> or maybe it should be colormatrix=out_color_matrix [01:16] <c_14> peshalto: scale=w=1280:h=720:out_color_matrix=bt709 [01:18] <peshalto> heh ok now running without errors [01:18] <peshalto> but [01:18] <peshalto> still running& and running&. [01:18] <c_14> Add a duration switch to the aevalsrc or add -shortest as an output option. [01:19] <peshalto> 3 hours long when I Ctrl-Ced it [01:19] <c_14> 3 hours of silent audio and 1 frame of video [01:19] <peshalto> heh [01:19] <c_14> just what you've always wanted [01:20] <peshalto> yeah it was black after one frame& call hollywood! [01:21] <llogan> ffmpeg -framerate 119.99 -i input -filter_complex "aevalsrc=0[a]" -map 0:v -map "[a]" -s 1280x720 -pix_fmt yuv420p -c:v libx264 -x264opts colormatrix=bt709 -profile:v high -c:a aac -strict -2 -shortest output.mov [01:22] <llogan> i wanted to use format and scale instead of -pix_fmt and -s but i'm too hungover to get it to only do one frame [01:23] <llogan> and -vframe won't include the audio [01:23] <peshalto> ok so then I put the output filename along with my movie-to-be-prefixed.mov filename in mylist.txt, and did: ffmpeg -f concat -i mylist.txt -c copy output.mov [01:23] <peshalto> and got corrupted output similar to what I got when I tried the faq entry [01:24] <peshalto> still it seems Im getting closer [01:24] <llogan> i don't know if using output.mov for an input and output.mov for the output will grok [01:24] <peshalto> ffmpeg -f lavfi -i "aevalsrc=0" -r 1 -i thumbnail.png -c:v libx264 -pix_fmt yuv420p -vf 'scale=w=1280:h=720:out_color_matrix=bt709,fps=119.99' -profile:v high -shortest thumbnail.mov [01:24] <peshalto> comparing mine with llogans now [01:25] <llogan> several ways to flog a horse or kick a cat i guess [01:25] <peshalto> I get an error with that one [01:25] <llogan> workedforme [01:26] <peshalto> whats input in your example? mine is a png file [01:26] <llogan> your input file name [01:26] <peshalto> it worked for me too but did you try the next step, prepending it to a video file? [01:26] <peshalto> thats the part that doesnt work for me [01:27] <peshalto> well I mean c_14s worked for me [01:27] <peshalto> [concat @ 0x7fc16b800000] DTS 28 < 279 out of order [01:27] <peshalto> [mov @ 0x7fc16b825e00] Invalid DTS: 379 PTS: 28 in output stream 0:0 [01:27] <peshalto> [mov @ 0x7fc16b825e00] Application provided invalid, non monotonically increasing dts to muxer in stream 0: 279 >= NOPTS [01:27] <llogan> no. too lazy to try. [01:27] <peshalto> av_interleaved_write_frame(): Invalid argument [01:27] <peshalto> yeah I have something not matched pu [01:27] <peshalto> up* [01:31] <llogan> peshalto: maybe it will still play fine [01:32] <llogan> oh nevermind, i missed some of your output [01:32] <llogan> anyway, it seemed to work for me [01:33] <peshalto> the input file was 30MB, the output (concatenated) file is 9.7k [01:35] <peshalto> the other command (pasted above) gave me a reasonable output file size, but the file system cant make a thumbnail out of it, and VLC thinks its 1 second long (should be 13 secs long) [01:36] <peshalto> I guess what I need to learn is how to read the ffprobe output, and translate that into the right ffmpeg input [01:40] <peshalto> a killer youtube tutorial would be someone going through a screen full of ffprobe output and explaining all the stuff bit by bit [01:41] <llogan> peshalto: here's one you can test http://avmule.com/junk/peshalto.mov [01:42] <llogan> not exactly the same as yours, but at least it will see if the OS chooses 1st frame [01:42] <llogan> or see if it actually plays correctly in your players of choice [01:43] <peshalto> nice [01:43] <peshalto> yeah, first frame, and it plays [01:44] <peshalto> assuming the first frame is the one with the 0 on it [01:45] <llogan> no. it's a screenshot from the 1986 NES game "Pro Wrestling" [01:45] <llogan> "a winner is you" [01:46] <peshalto> ah right [01:46] <peshalto> yeah it does NOT show the first frame [01:46] <llogan> http://knowyourmeme.com/memes/a-winner-is-you (although no sarcasm was meant with this image choice) [01:47] <peshalto> Ill have to play with it to see what its doing. maybe be a heuristic based on what it finds in the file. I wonder if theres any metadata you can set to control which frame is used. [01:48] <peshalto> gotta go soon, will report back if I figure this out [01:49] <llogan> peshalto: http://avmule.com/junk/clock.mov [01:50] <llogan> i guess "clock" is not an accurate name, but it may allow you to see what frame it chooses (if it always chooses a certain frame) [01:51] <peshalto> cool& thanks llogan and c_14 [01:51] <peshalto> later [01:51] <llogan> or maybe it just chooses a certain dura...........poof [02:31] <diegoviola> i'm trying to encode some mp4 video file to webm, but i want to preserve the same video quality, what option should I look for? [03:12] <diegoviola> any ideas? [03:14] <diegoviola> how do i find out a list of presets? [03:27] <radleader> i'm exporting a video from after effects that contains alpha data that i want to use ffmpeg filter graph to composit together with a series of PNG images (that show through the alpha) then produce an mpeg output - what format would be lowest computational overhead for ffmpeg for the input alpha video [03:28] <radleader> should i use h264 and have an extra video file with the alpha mask in it? [03:28] <radleader> should i use an avi containerand just keep it uncompressed [03:28] <radleader> not sure what the best plan here is [03:41] <jorb> http://superuser.com/questions/804508/some-transcoded-streams-delayed-or-no… [05:34] <myubuntu> how to mp4's indexs from tail to head? [05:47] <myubuntu> how to move mp4's indexs from tail to head? [06:06] <relaxed> myubuntu: ffmpeg -i input -movflags faststart output.mp4 [06:37] <hanshenrik> where can i buy a cat? [06:46] <ghospich> you can download absolutely free (as freedom and beer) cat which comes with GNU coreutils. [06:48] <hanshenrik> oh ok, but i already have the source and gcc [06:49] <hanshenrik> cat* [06:49] <hanshenrik> so.. i want to grab the first 4 minutes and 10 seconds off a video, losslessy, umm, ffmpeg -i foo -c:v:copy -c:a:copy bar ? [06:51] <rav0> -t duration https://www.ffmpeg.org/ffmpeg.html [06:55] <hanshenrik> ffmpeg -i foo -c:v:copy -c:a:copy -t 00:04:10 bar then, look about right? [06:55] <hanshenrik> thanks btw [06:55] <rav0> yep [06:56] <rav0> if there isn't a keyframe where you specified, ffmpeg will cut at the next keyframe. if your content has a visible cut, there is likely a keyframe nearby, you can try specifying 0:04:09 to get what you want [06:57] <hanshenrik> Unknown encoder '-c:a:copy' , and i am using a 2014 build :s [06:58] <relaxed> -c:a copy [06:58] <hanshenrik> oh, [06:59] <rav0> for video also [06:59] <rav0> or just -c copy [06:59] <hanshenrik> ok that worked great, thanks guys [07:00] <hanshenrik> though i must say, i prefer the old (deprecated? remvoed?) -acodec / -vcodec :p [07:04] <hanshenrik> so.. about 1am today, went down on my gf... she must have been very relaxed, cause she fell asleep [07:41] <myubuntu> relaxed: thanks. [07:42] <myubuntu> test [07:43] <relaxed> microphone check [07:43] <relaxed> you're welcome [09:26] <ewa> hi [09:26] <ewa> i have a problem with micro on Debian [09:26] <ewa> i need to record a video with ffmpeg but using external microphone not front mic in my notebook [09:26] <ewa> how can I mute front mic and use external mic? [09:27] <ewa> right now I record with both of them at the same time so the quality is bad [10:03] <K4T> is it possible with ffmpeg to restream multicast input? As input I would like to have multicast ip address: -i udp://239.239.1.1:4321 [10:03] <K4T> for example [10:58] <BillyZane> is it possible to setup a server on a computer that does not have X installed and broadcast a video without the audio or video playing locally? [11:06] <myubuntu_> ls [11:41] <Speed``> i just got my ducky shine 3 :) [11:41] <Speed``> oops wrong channel [11:43] <pron> gz [12:50] <pmarty> do ffmpeg/ffplay read any configuration files by default? i would like to put -hide_banner option there. [13:05] <pmarty> then i might just add shell alias ffmpeg='ffmpeg -hide_banner' [13:20] <Nopik> hi, I'm having small problem with ffserver, it does not restart ffmpeg (Launch config line) after they exit.. anybody had such problem, how to fix it? [14:06] <KAT_> which CPU is the best for ffmpeg? I should look on to amount of cores a frequency? [14:06] <KAT_> cores or frequency* [14:45] <vlatkozelka> hi , is it possible to add a new "layer" once streaming has started ? for example if i need to add some "urgent news" at the bottom in real time . I know this can be done if u know u gonna add a video from the beginning ... but is it possible in real time after the streaming started ? [14:58] <KAT_> vlatkozelka, you should use something like CasparCG to do thing like this [15:32] <jorb> anyone have a sec to try and answer this question? http://superuser.com/questions/804508/some-transcoded-streams-delayed-or-no… [15:32] <jorb> or just... comment on it here w/e [15:37] <jorb> or just tell me i am lazy for not typing the question into irc [15:55] <jorb> http://pastebin.com/2uCuFLFB [16:07] <disruptive> hi all [16:08] <disruptive> Im doing some searching around and didnt find much, but is there anything along the lines of FEC or some sort of error correction on UDP streams? [16:08] <disruptive> I see that there is some error concealment messages when using rtsp over tcp [16:09] <jnvsor> UDP has nothing at all, it's just data. The only reason it's used is because while TCP has error correction it also has throttling built in [16:10] <jnvsor> If you want error correction you'll have to DIY or have it built into whatever protocol you run over UDP [16:10] <disruptive> thanks jnvsor, I understand that UDP just spits whatever. Is there any sort of server-client already available with ffmpeg implementations that have some sort of error recovery in the mix? [16:10] <disruptive> I see [16:10] <jnvsor> Which protocol are you using to transfer? [16:10] <jnvsor> UDP by itself doesn't really do anything unless you have a protocol for it already [16:11] <disruptive> if Im doing it right, Im using multicasting [16:12] <jnvsor> So, mpegts format? [16:12] <disruptive> its h.264 in rtp? [16:13] <jnvsor> Hmm, never used rtp sorry [16:14] <disruptive> in the case of mpegts would there be some light in this question? [16:15] <jnvsor> No idea, I just noticed it was the example code for all the ffmpeg multicast info I could google :P [16:15] <disruptive> lol [16:28] <disruptive> thanks again for your replies jnvsor [16:37] <jorb> ..anyone seen an issue with video files jamming up ffmpeg streaming to stdout? [16:37] <jorb> http://pastebin.com/2uCuFLFB [16:38] <jorb> err audio files with images in the container [16:39] <jorb> http://superuser.com/questions/804508/some-transcoded-streams-delayed-or-no… [16:39] <disruptive> Ive ran by this before: http://lists.mplayerhq.hu/pipermail/mplayer-users/2011-August/083163.html [16:39] <klaxa|work> i never encountered that problem [16:39] <disruptive> but that was back in 2011 [16:40] <jorb> well mplayer runs it OK eventually [16:40] <jorb> but other players will just drop the streams because they are "unrecognized" [16:40] <klaxa|work> can you try writing it to a file and play it with a different player then? what formats do you use? is it always mp3? [16:41] <jorb> everything is transcoded to mp3 by default [16:41] <jorb> yes, writing toa file removes the issue.. [16:42] <jorb> i'm pretty sure when i was running Ubuntu 12.04 this was not an issue too [16:43] <klaxa|work> that is weird indeed, can the players play it like this: cat file.mp3 | player - ? [16:44] <jorb> yes [16:44] <jorb> it is something that ffmpeg is doing presumably [16:44] <jorb> i am guessing just pading 000000000000000 or something [16:45] <jorb> kind of a minor issue anyway, but seems like others would have run into it [16:46] <klaxa|work> can you try to write to a file using a pipe and compare that with the file produced without a pipe? [16:47] <jorb> yeah, it works fine if i pipe it to a file [16:47] <jorb> the problem is definitely rooted in the audio files themselves though [16:47] <jorb> just cause a bug in ffmpeg maybe... [16:48] <jorb> something crappy / non standard in the container maybe [16:48] <klaxa|work> does it work with other files? [16:49] <jorb> most files run fine, just particular ones have an issue [16:49] <jorb> those with album art is one category [16:53] <klaxa|work> maybe try adding "-map 0:a" to the ffmpeg command [16:53] <klaxa|work> ffmpeg tries to put in the cover art into the audio-file, do you want that even? [16:54] <jorb> yeah, kind of nice [16:54] <jorb> haha [16:54] <jorb> does that front load the audio? [16:54] <klaxa|work> it tells ffmpeg to only use the audio from the first input [16:55] <jorb> yeah, same result with the -vn option i believe [16:55] <jorb> does work, but no album art for the player [16:59] <klaxa|work> i can't find documentation on where id3 metadata is stored [17:00] <klaxa|work> what i can imagine is that the metadata is placed at the end of the file with pointers to it at the beginning or something [17:04] <jorb> yeah that would definitely be an issue.. i'll study the troublesome files some more [17:04] <jorb> seeing some folks with the issue also from their forum [17:04] <jorb> http://forum.subsonic.org/forum/viewtopic.php?f=2&t=14644 [17:05] <klaxa|work> turns out id3v1 does it like that, id3v2 does not [17:21] <kaotiko> hi [17:26] <jorb> fuck can't get by the spam detection on forum.subsonic.org lol [17:31] <jorb> the question was what year was the Battle of Hastings [17:34] <mifritscher> hi [17:35] <mifritscher> how can I set the mtu for udp streaming? [18:29] <mifritscher> In case somebody answered about the mtu-question: sorry, had connection problems [18:37] <ac_slater_> hey all. I'm doing something like `ffmpeg -i hd.mpeg -f rtp rtp://127.0.0.1:6666 > test.sdp` and `ffplay test.sdp` .... this works but the video quality is super low on the receiver side. I tried setting bitrate to ensure it was at least what ffprobe detected. Any clues? [19:28] <ac_slater_> Any hints how to take an mp4 (h264) with video and audio and reencode it to mpegts? [19:28] <ac_slater_> oh, wait, I got it, really simple [19:29] <Mavrik> just remux it with -codec copy [19:30] <Mavrik> no need to reencode [19:31] <ac_slater_> Mavrik: right! [19:32] <ac_slater_> Mavrik: apparently, 'h.264 bitstream malformed... ' [19:33] <ac_slater_> Mavrik: ah it even gave me a hint, `-bsf h264_mp4toannexb` [19:33] <Mavrik> mhm :) [19:38] <mifritscher> hi [19:45] <ac_slater_> Mavrik: it's also very frustrating that ffmpeg's rtp layer only supports one stream per session. Kind of defeats the purpose of it [19:48] <Mavrik> ac_slater_, I avoid RTP like the plague [19:48] <Mavrik> it's a shitty format that rarely works on any half complicated environment [19:49] <ac_slater_> Mavrik: it is really crappy. But sadly, I need it. [19:50] <ac_slater_> Mavrik: I've been using ffmpeg for mux/demux and live555 for delivery... live555 is a mess, and nothing from it plays on anything that doesnt use live555 for recieving - ie, not standard compliant [19:51] <Mavrik> mhm [19:51] <Mavrik> we just gave money to wowza at the end [19:52] <ac_slater_> Mavrik: does wowza work on ARM? Linux? [19:52] <ac_slater_> Mavrik: ah hold on, looking at the site... at first glance it's confusing [19:52] <Mavrik> ac_slater_, no idea [19:53] <Mavrik> I mean, we did use it on Linux :) [19:53] <Mavrik> but we had a x86 cluster [19:57] <ac_slater_> Mavrik: yea :( I'll be streaming this stuff from a 733mhz ARM [19:57] <Mavrik> ah, good luck :) [19:57] <ac_slater_> (not the ras-pi ... but an SoC) [19:57] <ac_slater_> Mavrik: yea ... ffmpeg's rtp stuff is just lacking, I'll ditch it and start over [19:57] <ac_slater_> thanks!@ [19:58] <Mavrik> why not use something sane like MPEG2-TS over UDP, RTSP which is slightly better or something like that? [20:09] <ac_slater_> Mavrik: right, I've moved to RTSP and it seems to work ... I just dont need to VCR-like features or the TCP requirement. [20:09] <ac_slater_> MPEG2-TS is ok only if I can have h.264 video, which it looks possible, maybe not via ffmpeg though [20:12] <Mavrik> ac_slater_, of course its possible [20:12] <Mavrik> MPEG2-TS is what we use for IP-TV distribution, DVB-T and DVB-S [20:13] <Mavrik> and also in US for broadcasting :) [20:13] <Mavrik> it's just that there was no need for update in MPEG4 standard so there was no MPEG4-TS [20:13] <ac_slater_> Mavrik: that makes sense. I just know not to make assumptions with container formats ;) Though , TS == transport stream, so it's prob different [20:13] <ac_slater_> I see. [20:13] <ac_slater_> interesting [20:14] <Mavrik> anyway, it's a format made for streaming and usually works rather well over UDP [20:16] <ac_slater_> Mavrik: what about jitter compensation, redelivery, etc... sure, it's UDP and therefore simple, but RTP is does that ... at least the standard says it does, I cant find a solid implementation [20:16] <Mavrik> you practically never want redelivery in a streaming video [20:16] <Mavrik> (and you can also always do tcp streaming) [20:16] <ac_slater_> Mavrik: unless it's a simple 1500byte slice [20:17] <ac_slater_> Mavrik: RTP seems convincing, but yes, it does produce latency with redeliveries. Thankfully, most implementations dont implement redelivery [20:17] <Mavrik> having redelivery on any transmission is just asking for trouble [20:17] <Mavrik> if the packet is gone, it's gone [20:17] <ac_slater_> I agree. [20:18] <Mavrik> resending a packet which should be already played will just clog your buffers and network [20:18] <ac_slater_> resulting in latent video, which is lame [20:18] <ac_slater_> I just need RTP for it's standardization. [20:18] <Mavrik> unless you have a huge buffer... but most of applications have shown that using TCP is just a source of trouble on network layer [20:18] <ac_slater_> ie - most things play it [20:18] <ac_slater_> right, udp is the way to go [20:19] <ac_slater_> I avoided mpeg2-ts for a long time, since RTP is everywhere these days ... at least in my target audience [20:21] <ac_slater_> I still havent messed with oRTP, etc... mainly because most implementations dont handle packetization of video, they are intended for audio really. [20:22] <ac_slater_> I need h.264 and a specialized data format which I've written an encoder/decoder for ... MPEG2TS is my container format already since it allows an arbitrary data stream. I guess it wouldnt be terrible to test MPEG2-TS delivery [23:39] <ac_slater_> I wonder how much it'll take to add support for multiple streams to ffmpeg [23:39] <ac_slater_> (/ libavformat) [23:41] <klaxa> depending on the definition of "multiple streams" it already supports multiple streams [23:41] <klaxa> what do you define as "one stream" exactly? [23:50] <ac_slater_> klaxa: I see. The RFC has all of that in there... that is, RTCP will do some synching [00:00] --- Sat Aug 30 2014

1 0

ffmpeg-devel.log.20140828
by burek 29 Aug '14

29 Aug '14

[00:04] <J_Darnley> Oh wait... there is no --enable-gplv3 [00:08] <cone-843> ffmpeg.git 03Michael Niedermayer 07master:35debfc366cd: avformat/rdt: fix compiler warning about const qualifier being discarded [00:08] <cone-843> ffmpeg.git 03Michael Niedermayer 07master:e6516944a3d5: avformat/rtpdec_asf: fix compiler warning about const qualifier being discarded [00:43] <J_Darnley> WTF?! What happened to the -preset option? [00:45] <J_Darnley> Oh wait.. is that because I didn't enable x264? [02:02] <cone-843> ffmpeg.git 03James Darnley 07master:8c1b942ad7b6: Replace incorrect use of "multiply" with "multiple" [02:03] <Timothy_Gu> Daemon404 & Compn: from my experiences tonal languages are easier at first, when you are learning the basics like pronunciation [02:04] <Timothy_Gu> But when you get to higher level stuff like conjugation and tenses, Chinese is way easier because it doesn't have tenses. [02:05] <llogan> What about Cantonese? [02:06] <llogan> Do nine men interpret? Nine men, I nod. [02:06] <Timothy_Gu> It doesn't even have plural forms except when you are talking about people, like I vs we ( vs ì), but even then it's crazily easy as you just add a character, no weird I vs we, he vs they stuff [02:06] <Timothy_Gu> Cantonese is Chinese [02:06] <Timothy_Gu> With a special pronunciation [02:07] <llogan> And Inupiaq is American. My Mandarin speaking friend can't understand a damned thing in Cantonese. [02:07] <Timothy_Gu> I _speak_ Mandarin and a little bit of Shanghainese, but everything written is the same [02:10] <Timothy_Gu> And most Cantonese-speaking population in China speak Mandarin too (sometimes with an accent so rich you can't even understand it, which happened to me when my family went to a "Chinese " restaurant operated by people from Guangdong (another name for Canton)) [02:16] <Daemon404> i think you may be a bit biased [02:16] <Daemon404> i must say ive made efforts to learn cantonese in my mid-20s... nope.jpg [02:16] <Daemon404> tones just dont click [02:19] <llogan> Daemon404: but Swahili does. ha ha ha....eeehhhhh. sorry. [02:19] <Daemon404> funny [02:25] <Timothy_Gu> Daemon404: how am I biased? I know quite a few Chinese people who just learned Cantonese by singing songs made by Hong Kong singers. Of course it's easy to learn it! (If you know standard Chinese/Mandarin, that is...) [02:27] <Daemon404> i am assuming you learned a tonal language from a vey young age [02:27] <Daemon404> very* [02:27] <Daemon404> "I know quite a few Chinese people who just learned Cantonese by singing songs" -> that completely ignores the tonal aspect [02:27] <Timothy_Gu> Also IIRC UN officially recognized Cantonese as a language some time ago. It's really unfair because Cantonese isn't even the most spoken Chinese dialect else than Mandarin (Wu, the higher classification of Shanghainese, is) [02:27] <Timothy_Gu> What do you mean by tonal languages? [02:28] <Timothy_Gu> Oh Chinese [02:28] <Daemon404> http://simple.wikipedia.org/wiki/Tone_language [02:28] <Timothy_Gu> Never mind [02:29] <Timothy_Gu> I meant English, Spanish etc. are easier in the beginning, harder later [02:30] <Daemon404> maybe [02:30] <Daemon404> it say it relies heavily on your linguistic background [02:30] <Daemon404> a to which is easier in the long run [02:30] <Timothy_Gu> There are no subjunctives or fancy participles in Chinese. There are also no difference between past tense versus present perfect [02:31] <Daemon404> chinese has two things going against it that are a (massive) speedbump for me [02:31] <Daemon404> 1) tones (cantonese) [02:31] <Daemon404> 2) you cant practice it usign writing [02:31] <Daemon404> e.g. french i learned via writing and speaking. [02:31] <Timothy_Gu> Well I know one tonal languages and one Indo-European languages at least semi-natively so I guess I'm pretty good with languages :D [02:32] <Timothy_Gu> Cantonese tones are weird, I have heard. [02:32] <Daemon404> i really dislike the sound of cantonese [02:32] <Daemon404> it is had on my ears [02:32] <Daemon404> mandarin is ok. [02:32] <Daemon404> hard* [02:33] <Daemon404> (my SO speaks cantonese to her parents) [02:33] <Timothy_Gu> (Its kinda like Vietnamese) [02:33] <Timothy_Gu> SO? [02:33] <Daemon404> signigicant other [02:33] <Timothy_Gu> hmm [02:35] <Timothy_Gu> She can't read or write Chinese can she? [02:35] <Daemon404> no [02:35] <Timothy_Gu> Sad [02:36] <Daemon404> it has no practical value [02:36] <Timothy_Gu> You can use it to decipher Chinese DTS specs ;) [02:36] <kierank> chinese dca [02:37] <Timothy_Gu> Whatever [02:37] <kierank> dra i mean [02:38] <Timothy_Gu> Anyways I gotta start on doing AP World homework now (if you know what AP World is and how AP world homework is like) [02:40] <Timothy_Gu> BTW CE flamed me for writing the patch that changed Gerard/Fabrice's name and Michael for applying it [02:41] <Timothy_Gu> Legal stuff, blah blah blah [02:41] <llogan> i was one of the dumb/smart kids who didn't want to do AP because i was too busy playing video games and bullshitting with friends. [02:42] <kierank> Timothy_Gu: he might be right [02:42] <wm4> Timothy_Gu: where? [02:43] <Timothy_Gu> I'm currently in a Catholic school where 80% of all the people are white not counting Hispanic and 15% are Hispanic. And in my AP World class 50% are Asian. [02:43] <Timothy_Gu> Not to be racist, but it's the truth [02:43] <wm4> I mean, where did CE flame you? [02:43] <Timothy_Gu> Privately [02:44] <Timothy_Gu> Well he mainly blamed Michael as he flamed him for merging a Libav change that changed copyright some time ago [02:44] <Timothy_Gu> And Michael still aplied my patch [02:45] <llogan> don't let it bug you. anyway you have to go study or you'll end up like me. [02:45] <Timothy_Gu> kierank: I'm perfectly fine with taking 3 AP classes at one time, but stupid rules of my new school don't let me [02:45] <Timothy_Gu> llogan: thanks [02:45] <Timothy_Gu> Bye then [02:50] <Compn> interesting [02:51] <Compn> i thought poeple hated it when describing mandarin and or cantonese as 'chinese' [02:53] <Daemon404> oh so it's not just kierank and I who CE flames privately (not me anymore, since i dont think he wants to talk to me) [02:53] <Daemon404> \o/ [02:53] Action: Daemon404 sleeps [02:53] <Compn> i watch films from china, vietnam, fillipenes, korea, thailand... anywhere i can get my hands on. i've picked up some chinese. ni hao :) [02:53] <Compn> CE flames me too :P [02:54] Action: Compn waves to carl [02:54] Action: Daemon404 throws some wuxia at Compn and goes back to sleep [02:55] <Compn> i dont really mind mandarin vs cantonese, the only thing that bugs me is that the film has to be in the original language, many times the dubbing into the other language is terrrrrible and high pitch. [02:55] <Compn> sometimes its the original cast doing both ? [02:55] <Daemon404> they have to do that because many cant read enough symbols or whatever [02:56] <Daemon404> but i agree it looks retarded. [02:56] <Compn> china film industry is really bad now. :( i cant remember anything big in the last 10 years [02:56] <Compn> only big news was destroying actors careers because of sex tapes :P [02:56] <Daemon404> Compn, http://en.wikipedia.org/wiki/What_Women_Want_(2011_film) [02:56] <Daemon404> ill just leave that here [02:56] <Daemon404> yes HK remade that mel gibson movie. enjoy. [02:56] Action: Daemon404 runs [02:57] <Compn> lol [02:57] <Compn> no more gwanji films :( [02:57] <llogan> what am i missing? In [1] can't the suggestion (or joke?) from Nicolas [2] be used instead of wget? [02:57] <llogan> [1] https://lists.ffmpeg.org/pipermail/ffmpeg-devel/2014-August/161881.html [02:57] <llogan> [2] https://lists.ffmpeg.org/pipermail/ffmpeg-devel/2014-August/161872.html [02:58] <Compn> Daemon404 : http://en.wikipedia.org/wiki/Jiangshi [02:58] <Compn> cant even remember the name of the films :\ [02:58] <Compn> er genre [02:59] <Compn> llogan : ffmpeg can be used to download the image, yes [02:59] <Compn> llogan : did you md5 verify its the same though ? [02:59] <Compn> iirc ffmpeg might remux it [03:00] <Compn> whereas mplayer -dumpstream would not :D [03:01] <llogan> md5 is same. and anyway i'm a curltard and don't have wget on my system [03:04] <Compn> so ping michael or send patch [03:05] <Compn> or just commit directly ;D [03:05] <Compn> but really no wget? :P [03:18] <BBB> ubitux: you mean reproduce? it is; only first 3 frames was sufficient form what I remember [04:13] <cone-843> ffmpeg.git 03Deb Mukherjee 07master:b6a0b8bcf497: Adds support for setting aq_mode in libvpx encoder [05:04] <Timothy_Gu> I have no idea when Chinese people's English began to be this terrible. Maybe it has never been good? [05:06] <Timothy_Gu> Oh, ubitux, FYI I've been to Japan once and the English spoken by people there is incomprehensible even to my parents, who both have graduate degrees in English [07:32] <ubitux> Timothy_Gu: it's fine in most big cities IMO; but as soon as you go sidetrack/countryside, please just don't speak at all [07:33] <ubitux> it's more like they prefer not to talk than even try :p [07:33] <ubitux> at least from my experience [07:34] <ubitux> (why did i wrote "please", i meant "they") [09:58] <nevcairiel> when i tell swscale to downscale HD to SD, can i also tell it to convert bt709 to bt601? I don't think it has such an option, has it? [09:58] <wm4> yes [09:58] <wm4> via avoptions [09:58] <nevcairiel> but does it actually process this stuff when just scaling yuv->yuv? [09:58] <wm4> whether it works, dunno [09:58] <wm4> you'll probably have to try [09:59] <wm4> michaelni: can it? [10:04] <wm4> hm I couldn't get it to convert them [10:05] <nevcairiel> there is a lavfi filter that does it, but dont really feel like putting lavfi into this conversion code, its not used there yet at all [10:07] <nevcairiel> ... and i highly doubt android players would read the color matrix tag from a h264 file, if i try setting that instead [10:14] <wm4> also looking at the API again, what I said about avoptions was incorrect too - you have to use sws_setColorspaceDetails, which fails for yuv->yuv [10:18] <nevcairiel> its not like anyone complained about this issue yet [10:18] <nevcairiel> maybe i'll just ignore it for a few months more [12:20] <cone-226> ffmpeg.git 03Stefano Sabatini 07master:fa39f8c0e2c9: doc/ffprobe: fix -show_entries example [12:48] <Daemon404> "I think for there to be a motvation the payment should be above the [12:48] <Daemon404> minimum wage of the average western country [12:48] <Daemon404> :D [12:48] Action: Daemon404 loled [12:58] <BBB> the problem with bug payment is, it could take a few minutes to a few months [12:58] <BBB> $50/bug is thus highly inefficient [12:58] <BBB> it promotes working on silly one-liner bugs [12:58] <BBB> (not these that take 2 days of debugging) [13:02] <J_Darnley> If it gets "silly one-liner bugs" fixed quickly, good. [13:03] <J_Darnley> They must have been present long enough for someone to report them. [13:03] <beastd> BBB: silly one liner fixes can take weeks to debug in sometimes. anyway it is somthing else that is important. if those silly bugs are away they are out of the list and people start to focus on what is remaining. so IMHO 50$/bug seems not too bad. [13:10] <beastd> I'd say for the more involved bugs, the 50$ bounties won't help in any way. It is eithter the person wants to work on it and therefore heavily invest in it or not. That person would get paid 50$ in the end, but it would just be a tiny partial refund of the overall investment she did already. [13:14] <nevcairiel> ubitux: so you got luca to implement your pkg-config problem somehow? :P [13:14] <ubitux> not really [13:15] <wm4> heh [13:15] <ubitux> i'm preventing him from doing changes [13:15] <wm4> if some jerk on ffmpeg rejects something, just send a patch to Libav [13:15] <ubitux> the pkg-config problem is solved [13:15] <wm4> that's the nice side of the split [13:15] <ubitux> --pkg-config=true is what CE is looking for [13:15] <ubitux> there is no need to add an ambigous --pkg-config=false [13:16] <nevcairiel> especially because false is also an executable [13:16] <ubitux> (because false can be a program and it's used as a special key here) [13:16] <ubitux> yes, right :) [13:16] <ubitux> anyway, i found the solution for CE [13:16] <ubitux> he just has to use --pkg-config=true --extra-ldflags=-lx264 [13:16] <ubitux> he didn't reply in the thread though [13:17] <nevcairiel> he probably wants to not have to do the last part [13:18] <ubitux> he's specifying the other ldflags anyway [13:18] <ubitux> and since this -lx264 is wrong in the case of static typically there is no reason to hardcode it [13:19] <ubitux> what would be the point of using pkg-config if we had to also hardcode an incomplete version of the flags.. [13:20] <nevcairiel> i'm not saying he is being logical [13:21] <ubitux> and anyway, he will have to add --pkg-config=true in his cli anyway as a retro-compat option [13:21] <ubitux> so he can add -lx264 as well [14:02] <ubitux> we don't use "local" anywhere in the configure [14:02] <ubitux> it's not standard? [14:04] <ubitux> (there is some evil in the configure around that) [14:16] <ubitux> ok i just reworked the whole pkg/configure thing [14:16] <ubitux> it looks more usable [14:20] <ubitux> so here is a PoC http://pastie.org/9509979 [14:21] <ubitux> and it looks like this: http://pastie.org/pastes/9509980/text (version check hacked) [14:30] <Daemon404> ubitux, dont worry, carl will reject it [14:30] <Daemon404> i am certain [14:32] <ubitux> we'll see [14:39] <durandal_1707> is anyone writting new filter(s)? [14:40] <kierank> after ibc i will make vf_scale properly support interlaced chroma [14:42] <saste> durandal_1707, no, are you? [14:42] <wm4> durandal_1707: someone still needs to port pp7 [14:42] <saste> who is willing to help me with ASM? [14:42] <ubitux> durandal_1707: mmh i'd like to, but my TODO list is filled with fixes & improvements in filters [14:42] <saste> in that case i might consider to port eq/eq2 [14:42] <ubitux> saste: o/ [14:43] <saste> i should probably start with a C-only version [14:43] <ubitux> yes [14:43] <cone-226> ffmpeg.git 03Michael Niedermayer 07master:d6bd29492ca2: avformat/icoenc: Remove deprecated use of codec_name [14:44] <durandal_1707> saste: i'm working on silence removal, filter from sox [14:44] <ubitux> saste: there isn't much asm actually [14:45] <ubitux> and it's just poor mmx [14:50] <J_Darnley> I can probably help with asm [14:50] <J_Darnley> x86 simd assembly that is. [14:50] <Daemon404> i dont get why port sox filters [14:51] <Daemon404> just vf_sox.c [14:51] <Daemon404> because NIH? [14:51] <wm4> because having them in a single big project with unstable ABI is better than a light weight plugin interface [14:53] <J_Darnley> To be honest, one advantage to porting is not having to make some interface between av_opt and, in this case, sox's options (whatever they might be like) [14:54] <Daemon404> thats not really an advantage [14:54] <Daemon404> thats laziness [14:55] <durandal_1707> sox does not support floats and doubles [14:55] <J_Darnley> True, you still have to dea with the escaping hell anyway. [14:56] <J_Darnley> *deal [14:56] <Daemon404> it most certain does support floats [14:56] <Daemon404> certainly* [14:56] <durandal_1707> nope 32bit ints [14:56] <Daemon404> all of its documents indicate it sure does support f32 and f64 [14:57] <durandal_1707> that is something else [14:57] <Daemon404> i love your vague responses [14:57] <Daemon404> completely devoid of any actual information [14:58] <ubitux> (last commit from sox is from nov 2013?) [15:00] <durandal_1707> define SOX_SAMPLE_MAX (sox_sample_t)SOX_INT_MAX(32) [15:00] <ubitux> durandal_1707: if you're porting a silence cut filter; could you instead port the code to silence detect, and make it possible for aselect to drop using the meta? [15:01] <ubitux> the silence detect code in af_silence is really weak, so you can really do what you want with it if you don't break the interface [15:02] <durandal_1707> that would be nightmare to work with [15:02] <ubitux> -af "silencedetect,aselect='not(silence)'" ? [15:03] <durandal_1707> silenceremove can remove actual nonsilence [15:04] <wm4> a terrible DSL is growing here... [15:04] <durandal_1707> and metadata is not designed for such thing, as filter must operate with samples and not frames [15:05] <ubitux> silencedetect can control the frame size [15:05] <ubitux> but yeah i get your point [15:05] <ubitux> if you don't want to use the meta & aselect for that, then at least share the code between silencedetect and silencecut [15:06] <ubitux> so both would benefit from the same code & efficiency [15:07] <durandal_1707> there is not much to share except A < B [15:08] <wm4> hm, is accessing AVIndexEntry a sane idea? I suspect not, but it seems attractive for certain uses [15:10] <durandal_1707> from AVStream? [15:11] <wm4> yes [15:12] <wm4> mapping byte positions to timestamps for visualizing what of a network stream has been cached [15:12] <durandal_1707> if you want to be mplayer you can use it [15:14] <wm4> that means it's a "don't go there" API? [15:15] <durandal_1707> i see big warning message several lines above [15:17] <wm4> oh it's in the private section [15:17] <wm4> that took me several seconds to get [15:17] <wm4> great API [15:19] <wm4> ubitux: with the audio visualizers, what's a good recommendation to keep them in sync? [15:19] <ubitux> keep them in sync with the audio? [15:20] <wm4> you have to do correct "scheduling" of audio frames [15:21] <wm4> well, hard to explain, but you need to ensure sane dataflow, without ending up with excessive buffering somewhere [15:22] <wm4> can one output be used as input to multiple inputs? [15:23] <cone-226> ffmpeg.git 03Deb Mukherjee 07master:b543a2902023: avcodec/libvpxenc: Adds support for constant quality mode in VP9. [15:24] <ubitux> wm4: parse error [15:24] <ubitux> you can use split [15:24] <ubitux> if you want to dup one output to feed it to several other inputs [15:25] <wm4> so maybe it would be best to funnel the audio data always through lavfi, and produce video frames as side-effect of the visualization filter [15:31] Action: ubitux doesn't get it [15:31] <wm4> what I said or what cehoyos said? [15:31] <ubitux> what you said [15:32] <ubitux> but i admit i don't understand much what's going on everywhere these days, so i don't get exactly what CE is talking about either :p [15:33] <wm4> a video player basically decodes audio and video independently from each other, and then times them to show up synchronized using the timestamps [15:33] <wm4> but with visualization, video depends on audio decoding [15:33] <wm4> so what do you do, add a second FIFO for audio? or a FIFO for the produced video frames? [15:37] <ubitux> ok get it, but i don't know [15:37] <ubitux> i believe the request_frame() thing is used for that [15:37] <ubitux> to force to get a frame (audio or video) at any point [15:37] <ubitux> i'm not the best person to ask [15:45] <J_Darnley> wm4: I don't know about rendering but I have a working visualiser in lavfi [15:47] <J_Darnley> I just ask the filter chain to provide me with audio frame of constant size. [15:49] <J_Darnley> If you want to look: https://gitorious.org/jdarnley/advanced-visualization-studio/source/avf_avs… [15:49] <ubitux> wm4: ffplay is able to playback and show visualization at the same time [15:49] <ubitux> so i would guess you could do it as well [15:51] <cone-226> ffmpeg.git 03Michael Niedermayer 07master:87961a8a504a: avcodec/flashsv2enc: fix "unused variable" warnings [16:09] <cone-226> ffmpeg.git 03Michael Niedermayer 07master:ef768ab9763d: avcodec/h264: Move h264_vdpau_class under ifdef to avoid unused variable warning [16:39] <wm4> ubitux: how do you call ffplay to visualize audio? there are options for some manual ffplay-specific rendering, but that can't be it [16:39] <ubitux> you're not gonna like it ;) [16:39] <ubitux> ffplay -f lavfi "amovie=$HOME/samples/music/heart.flac,asplit[out0][x];[x]showwaves=mode=cline:split_channels=1[out1]" [16:41] <wm4> oh, that's cheating [16:41] <wm4> mpv can do the same... [16:41] <ubitux> ffplay doesn't have -filter_complex unfortunately [16:42] <wm4> why does ffplay call the decoders at all [16:42] <wm4> it could just use lavfi :D:D:D [16:42] <ubitux> yes [16:43] <ubitux> and we need a sink encode filter ;) [16:43] <wm4> buffersink? [16:43] <ubitux> no [16:43] <ubitux> just like [a]movie [16:43] <ubitux> but for outputs [16:43] <wm4> I don't get it [16:44] <ubitux> -vf "testsrc, encode=bla.mp4" [16:44] <wm4> ah [16:45] <ubitux> haters gonna hate [16:45] <wm4> if you actually had scripting language bindings, that would be even useful [16:47] <ubitux> oh [16:47] <ubitux> i did it with ffmpeg [16:47] <ubitux> wm4: ./ffmpeg -i ~/samples/music/heart.flac -filter_complex "asplit[a][x];[x]showwaves=mode=cline,format=yuv420p[v]" -map '[v]' -f sdl - -map '[a]' -f alsa default [16:47] <ubitux> it's not fluid though [16:48] <wm4> I wonder how that synchronizes [16:48] <ubitux> doesn't work with showspectrum :( [16:48] <ubitux> wm4: badly it seems :p [17:22] <kierank> 3:43 PM <"ubitux> and we need a sink encode filter ;) --> waaa? [17:22] <kierank> wtf [17:22] <ubitux> 16:45:13 <@ubitux> haters gonna hate [17:23] <durandal_1707> gui for lavfi? [17:23] <ubitux> kierank: the rational was that we could have a fffilter being able to do most of what ffmpeg does but just being a filtergraph string [17:24] <kierank> my head [17:24] <kierank> boom [17:24] <ubitux> and so you could indeed design a GUI to create a filtergraph to pipe to it, or whatever workflow you fancy [17:25] <ubitux> kierank: filtergraph already have various sources, either being synth (aevalsrc, sine, testsrc, smptebars, mandelbrot...) or actually using real file sources (movie, amovie) [17:26] <ubitux> so having a special sink doing the encode would be consistent ;) [17:26] Action: kierank invokes godwin's law and says it's like ffsystemd [17:26] <ubitux> :D [17:26] <ubitux> speaking of this, we need ffserver integration in systemd ;) [17:27] <wm4> shouldn't systemd be part of libavdevice? [17:28] <wm4> kierank: the worst about this is that conceptually it's not even a bad idea, but in practice it's worse than hitler [17:29] <ubitux> ./fffitler "amovie=bla.mp3, showwaves, encode=waves.mp4:vcodec=libx264" [17:29] <kierank> wm4: it'll end up being a purely academic exercise [17:30] <ubitux> ff > ss [17:30] <ubitux> i need a blazon [17:31] <Daemon404> "fitler" [17:31] <Daemon404> lol [17:31] <Daemon404> how apt. [17:31] <ubitux> :) [17:31] <ubitux> we probably can't name it as such unfortunately [17:31] <ubitux> because debian will reject it [17:32] <durandal_1707> like lena [17:32] <ubitux> let's replace lena.pnm with hitler.pnm [17:33] <Daemon404> i really would like to just tell the debian folks to go forth and multiply. [17:34] <kierank> ubitux: lol [17:36] <ubitux> actually, we can [17:36] <ubitux> https://commons.wikimedia.org/wiki/File:Adolf_Hitler_cropped_restored.jpg is under CC [17:36] <ubitux> that could be a nice troll [17:36] <wm4> but it's B&W [17:36] <wm4> such a problem [17:37] <ubitux> https://commons.wikimedia.org/wiki/File:Hitler,_recoloured.jpg [17:37] <ubitux> what about this one? [17:37] <wm4> also needs the debian logo somewhere [17:37] Action: ubitux gonna send a patch [17:37] <wm4> or is it copyrighted? [17:38] <kierank> wm4: LOL [17:47] <ubitux> ffplay http://lucy.pkh.me/lena.pnm [17:52] <wm4> I have to say lena looks more attractive [17:52] <cone-226> ffmpeg.git 03Paul B Mahol 07master:0ddb0516f113: avformat/soxenc: return more meaningful error code [17:56] <ubitux> wm4: maybe we can make the picture a bit more pinky [17:56] <ubitux> would that help? [17:58] <ubitux> ffplay http://lucy.pkh.me/lena.pnm -vf hue=h=-30 [17:59] <Daemon404> ubitux, use eval to create a picture of hitler from fractals [17:59] <saste> ubitux, wm4 find an hitler pic, and replace the swastika symbol with a debian logo [17:59] <wm4> why is there hue but not eq? seems strange [17:59] <ubitux> saste :D [18:00] <wm4> saste: that was the idea [18:00] <wm4> they're playing copyright nazis [18:00] <saste> wm4, i'm all for it [18:01] <Daemon404> wm4, i prefer to think of them as teh freedom enforcement squad [18:01] <Daemon404> KGB?SS [18:07] <ubitux> ffplay http://lucy.pkh.me/lena-debian.pnm [18:07] <ubitux> alternative [18:09] <ubitux> anyway, gtg [18:10] <durandal11707> push it without review [18:10] <saste> shouldn't be controversial [18:12] <wm4> too crappy [18:51] <llogan> hangover........ [19:40] <durandal_1707> who is stalking libav devs? [19:43] <kierank> carl [19:44] <kierank> (allegedly) [19:51] <BtbN> Is it even supposed to produce identical results with all optimizations? [19:51] <BtbN> @ that ticket [19:52] <nevcairiel> Not necessarily, mpeg2 is not strictly specified like that [19:52] <nevcairiel> Although he seems to test copying, not decoding [19:58] <BtbN> Aparently it's not the video that causes it to be diffrent [19:58] <BtbN> it's the audio [19:58] <BtbN> But as long as the result is still valid and plays fine, i wouldn't see that as a problem or even bug [19:59] <BtbN> My guess would be that the atom is lacking some cpu extensions, and the optimized asm code on the i7 produces slightly diffrent results than the non optimized code. [20:00] <nevcairiel> None of the streams are being decoded, and if he used audio independently its also the same md5 [20:38] <TheFluff> some decoders produce different outputs even on the same input on the same machine [20:39] <TheFluff> like the mp3 one for example produces different outputs every time [20:39] <TheFluff> or at least it used to [21:37] <cone-123> ffmpeg.git 03Luca Barbato 07master:f07a4290a0e8: x11grab: K&R formatting cosmetics [21:37] <cone-123> ffmpeg.git 03Michael Niedermayer 07master:5393c9dadd23: Merge commit 'f07a4290a0e8f31796e348edd3ed06b8d15132d8' [21:45] <cone-123> ffmpeg.git 03Luca Barbato 07master:20e82b41fcad: x11grab: Small near-cosmetic refactorings [21:45] <cone-123> ffmpeg.git 03Michael Niedermayer 07master:65c9e2d0030d: Merge commit '20e82b41fcad23ebbb18d89948aebfcc53921198' [22:01] <ubitux> beastd: let's not play the game of who is willing or not in this thread [22:02] <ubitux> beastd: this is merely an attempt to summarize who would be ok with that [22:02] <ubitux> so we can see who seems to be willing to cooperate [22:03] <beastd> ubitux: Only wanted to point out that the person complaining about hostility of FFmpeg is completely hostile towards FFmpeg in the very same email. [22:03] <beastd> Occurred to me kind of ironic... [22:04] <ubitux> yes i know but it doesn't matter :) [22:04] <Compn> [03:52] <courmisch> If FFmpeg was actually compatible, then how come the soname changes (just like libav's) [22:04] <Compn> [03:53] <courmisch> and you call that a long-haul compatible binary interface? [22:05] <Compn> if someone wants to explain to remi about soname changes... i do not. [22:06] <cone-123> ffmpeg.git 03Luca Barbato 07master:7bb505a33ca1: x11grab: Drop a spurious space in the extension reporting message [22:06] <cone-123> ffmpeg.git 03Michael Niedermayer 07master:1fc396318833: Merge commit '7bb505a33ca131906b2ceb2f298e104c862740ea' [22:06] <Compn> ubitux , beastd : make kierank explain things. it was his idea [22:06] <Compn> :P [22:06] <ubitux> i believe my post was clear enough [22:12] <cone-123> ffmpeg.git 03Luca Barbato 07master:58396e806c65: x11grab: Use a typedef for the context, as most other code does [22:12] <cone-123> ffmpeg.git 03Michael Niedermayer 07master:c6a3b0092498: Merge commit '58396e806c65fe0eb00e6ccf1980f810cdceed05' [22:20] <cone-123> ffmpeg.git 03Vittorio Giovara 07master:f2583bc86eb5: matroska: list supported extensions [22:20] <cone-123> ffmpeg.git 03Michael Niedermayer 07master:56cc754c76d3: Merge commit 'f2583bc86eb5d7decc639fc9a36edc93e6003eef' [22:24] <Compn> ubitux : after dvd support in ffmpeg ... then binary codec loader and we can kill mencoder :) [22:24] <Compn> well and multiple file input ... [22:25] <ubitux> vobsub muxer [22:25] <kepstin-laptop> what about multiple file input is missing? [22:26] <Compn> kepstin-laptop : mencoder file.rm file.avi file.mpeg -oac lavc -ovc lavc -vf scale=640:480 -o file.avi [22:26] <Compn> might be possible with advanced filtering and concat in ffmpeg. [22:26] <kepstin-laptop> what does that do, concat them? or output as multiple video tracks? [22:26] <Compn> concat them [22:27] <kepstin-laptop> yeah, the concat filter can do that. [22:27] <Compn> assuming i put the -vf scale at the start of the command (oops) [22:27] <Compn> then maybe make concat by default [22:27] <Compn> multiple video tracks bah :P [22:28] <ubitux> you know we can't do that [22:29] <kepstin-laptop> but yeah, that example would be equivalent to ffmpeg -i file.rm -i file.avi -i file.mpeg -filter_complex [22:29] <kepstin-laptop> er, silly enter key [22:29] <ubitux> yes, we have the concat filter that does exactly that indeed [22:30] <ubitux> it's slightly less intuitive but it works [22:30] <kepstin-laptop> ffmpeg -i file.rm -i file.avi -i file.mpeg -filter_complex [0]scale=640:480[i0];[1]scale=640:480[i2];[2]scale=640:480[i2];[i1][i2][i3]concat=n=3 file.avi [22:31] <kepstin-laptop> (concat defaults to having 1 video stream and 0 audio streams in the output, otherwise it would need more parameters) [22:32] <kepstin-laptop> with some fixing of my silly off-by-one errors, of course. [22:33] <cone-123> ffmpeg.git 03Vittorio Giovara 07master:9301486408a4: avcodec: add stream-level stereo3d side data [22:33] <cone-123> ffmpeg.git 03Michael Niedermayer 07master:f30a8154abf2: Merge commit '9301486408a480629336af4d7fd873c0f28fb2d5' [22:34] <Compn> fair enough :) [22:35] <jamrial> now if we could get that working with a command line as simple and intuitive as the one from mencoder :P [22:37] <kepstin-laptop> pretty much the last thing I was holding out for from mencoder was a working detelecine, and it looks like ffmpeg has that now. [22:46] <Compn> jamrial : yes, it would be nice to make it automagically work... [22:52] <J_Darnley> I believe that ffmpeg should do less "automagically" [22:53] <Compn> then mencoder continues to live :P [22:54] <Compn> which i dont have a problem with, i cant speak for ffmpeg project on what ffmpeg should do [23:36] <kierank> ubitux: look at all the good natured conversations on the mailing list [23:36] <kierank> such harmoney [23:36] <kierank> ... [23:54] <cone-123> ffmpeg.git 03Vittorio Giovara 07master:d4ae8ac92f61: matroskadec: parse stereo mode on decoding [23:54] <cone-123> ffmpeg.git 03Michael Niedermayer 07master:31f9c228c042: ffmpeg: Fix missing {} in Stream to Packet side data code [23:54] <cone-123> ffmpeg.git 03Michael Niedermayer 07master:fcc39099087e: Merge commit 'd4ae8ac92f619507aadd021bb67b517d39d3a36f' [00:00] --- Fri Aug 29 2014

1 0

ffmpeg.log.20140828
by burek 29 Aug '14

29 Aug '14

[02:26] <voip_> Hello guys. I need to trans code 150 streams to h264 flash. How many CPUs approximately i will need ? [02:28] <llogan> voip_: not enough information to provide answer, because you could do it with 1 cpu. [02:29] <voip_> llogan, somting like -acodec aac -strict experimental -ar 44100 -ac 2 -b:a 96k -vcodec libx264 -b:v 900k -threads 4 -f flv [02:29] <klaxa> you still can do what with one cpu [02:30] <voip_> with this parameters u have 7% cpu usage on 8 core xeon e5 [02:31] <voip_> when runing only 1 stream [02:32] <voip_> how to calculate how many servers we need for transcoding 150 streams ? [02:33] <klaxa> oh you mean in real time [02:33] <klaxa> i don't know, try to do as many as you can with that setup, when it crashes: that's too much for one single machine [02:34] <klaxa> it also depends on the source material i think [02:34] <llogan> why -b:v 900k? [02:36] <Hello71> f [02:36] <Hello71> and parameters. [02:36] <Hello71> obviously you will be fine on a p4 if you use -preset ultrafast [02:37] <voip_> guys i need raltime for IPTV [02:38] <klaxa> if you expect a mathematical way to determine how many machines you need to do what you want to do, you will be disappointed [02:39] <llogan> voip_: ...because if you're limited by bandwidth in any way you should be using VBV instead of just -b:v [02:39] <llogan> with -maxrate and -bufsize [02:43] <voip_> llogan, i am not limited by bandwith in datacenter side, my some custumers with 1mb DSL may be limited. So we planing take 150 streams from privider trnascode and pass to wowza. [02:47] <llogan> progressive download or live streaming? (sounds like live) [02:48] <voip_> live streaming [02:54] <llogan> use VBV [02:55] <voip_> its will decrease cpu ussage ? [02:57] <llogan> [1] https://lists.ffmpeg.org/pipermail/ffmpeg-devel/2014-August/161881.html [02:57] <llogan> obviously that is the wrong channel, llogan [03:05] <voip_> llogan, i can increase with vbv cpu ussage? [04:30] <surf2b1> Here is my question: http://pastebin.com/w0DbyzPD Any help is greatly appreciated. [07:35] <tm512> why is ffmpeg so finicky when it comes to not having the audio and video desync terribly? If I use a alsa and a pulse source, it's fine, but two pulse sources, and it desyncs [07:49] <tm512> I get a lot of "Non-monotonous DTS in output stream 0:0; previous: 26603, current: 26588; changing to 26603. This may result in incorrect timestamps in the output file." [08:09] <tm512> I used to have this working just fine. I don't know what went wrong [09:47] <termos> I have an issue with a multithreaded program that crashes on av_interleaved_write_frame, it seems to happen when I start another thread with another encoding job. Can I not write frames while doing something like avformat_open_input? [10:13] <K4T> http://ffmpeg.gusari.org/viewtopic.php?f=11&t=1648&p=4654#p4654 can someone look on this? [11:31] <lkiesow> Hi, yesterday a college of mine brought to my attention the following sentence from the FFmpeg wiki: [11:31] <lkiesow> If you use -ss with -c:v copy, the resulting bitstream might end up being choppy, not playable, or out of sync with the audio stream, since ffmpeg is forced to only use/split on i-frames. [11:31] <lkiesow> I'm regularly using bitstream copy together with seeking and can't remember having any problems. Is this statement still true? [12:37] <groundnuty> hey, I've recorded a tutorial video for the a purpose of presentation, si there some magic way to make the video autoamticly pause at specyfied times? [13:25] <lkiesow> groundnuty: Not really, I guess that is something your player would have to do [13:29] <Diaz2211> hi all experts, [13:29] <Diaz2211> I am new here [13:29] <Diaz2211> I am searching the fixed point decoder for WMA [13:30] <Diaz2211> is there any information about that? [13:30] <Diaz2211> in libav, wma is not fixed point, when I try to use that to decode wma audio file, [13:31] <Diaz2211> it takes 100% cup usage, there is not float instructions support on my embedded cpu [13:31] <Diaz2211> I would like to search fixed point decoder [13:31] <Diaz2211> any ideas? [14:08] <Diaz2211> I can see the code wmafixed from vlc source code [14:08] <Diaz2211> but ffmpeg does not have that code, [14:09] <Diaz2211> is it possible to support wmafixed in ffmpeg source code? [16:06] <badcompiler> ?w [16:06] <badcompiler> ?? [16:06] <badcompiler> :help [16:13] <badcompiler> I have a really basic question about using ffmpeg. I'm not sure if this is codec/conatiner specific or just all files in general but. If I have a source file (QT Animation from Adobe Package I think) and I do a straight copy with ffmpeg -i input.mov -vcodec copy -acodec copy output.mov why do I get different duration always by a frame or more (40-160 ms) ? [16:13] <klaxa|work> what version are you using [16:14] <klaxa|work> of ffmpeg that is [16:14] <badcompiler> i think its 2.3.3 [16:14] <badcompiler> ffmpeg version N-62249-g845414b [16:14] <badcompiler> mandelbrot ? [16:15] <klaxa|work> that's weird, if it really bugs you, maybe upload a sample and file a bugreport? [16:16] <badcompiler> ffprobe -i input.mov Duration: 00:00:35.00, start: 0.000000. ffprobe -i output.mov Duration: 00:00:35.04, start: 0.000000, [16:17] <badcompiler> now I do get Unsupported codec with id 0 for input stream 1 when I run the ffmpeg copy [16:29] <OutBound> Hi... perhaps someone can help: When I try to capture the screen with x11grab and also 2 audio streams into three files, the audio drifts apart. After recording for two minutes, the audio files show being almost a second longer, increasing with longer recording session. Infos here: http://pastebin.com/N6hmV22Y [16:32] <Hello71> why don't you put everything into one file [16:32] <OutBound> because I need them seperate for different editing purposes [16:47] <klaxa|work> OutBound: using -r 30 as an input option might be the cause [16:48] <klaxa|work> i think there is a parameter specifically for x11grab to always produce the same amount of frames regardless of whether or not your hardware is fast enough to get them through x11 [16:49] <badcompiler> can somene else verify for me...if you down load a refernce file (eg http://a1408.g.akamai.net/5/1408/1388/2005110403/1a1a1ad948be278cff2d96046a… ) and run ffmpeg -i sample_iTunes.mov -vcode copy -acodec copy output.mov and compare ffprobe with both files ? [16:49] <klaxa|work> see: http://ffmpeg.org/ffmpeg.html#X11-grabbing for examples on how to use -framerate instead of -r [16:49] <OutBound> thanks klaxa|work [16:50] <badcompiler> every time the file output from ffmpeg adds 40 milliseconds, [16:51] <badcompiler> I cant be a bug it would be too obvious, there must be a good reason for it that I don't understand [16:52] <klaxa|work> how can this not be a bug? [16:52] <badcompiler> because I figure someone else would have definitely noticed it before me if it is a bug. [16:53] <klaxa|work> what happens if you remux the file ffmpeg produced again? [16:56] <badcompiler> hmm if i run the command ffmpeg -i output.mov -vcodec copy -acodec copy output2.mov, I get no change in the duration from ffprobe. In other words the inital 40 milliseconds added is still there but doesn't increase on second copy [16:57] <badcompiler> is there a built in padding time or something.... [16:58] <klaxa|work> i would assume it has to do with ffmpeg's mov muxer which apparently is slightly different from apple's one [17:00] <badcompiler> if I remux again that file to a mkv , it adds 200 milliseconds ! ffmpeg -i output.mov -c:v libx264 -preset slow -crf 22 -c:a copy output.mkv [17:00] <badcompiler> so thats 240 milliseconds from the original mov [17:01] <badcompiler> intresting you say that klaxa|work, because, the issue I'm trying to get to the bottom of, is incorrect audio sync playback in apple quicktime player !! [17:01] <badcompiler> and the first place I started looking was the difference in duration ffmpeg is adding when transcoding from mov files [17:02] <badcompiler> I dearly want to replace a production pipeline using ffmpeg and open source, but so far issues like these mean I have to slip audio and video streams by a few frames to compensate [17:12] <badcompiler> I found this which seems like the same thing I'm seeing http://permalink.gmane.org/gmane.comp.video.ffmpeg.devel/175181 [17:32] <NucWin> is it posssible to tell the encoder to use half the bitrate of the source without knowing the sources bitrate before hand? [17:34] <klaxa|work> can you tell what the number i'm thinking of is divided by 2? [17:34] <NucWin> yes n/2 [17:36] <NucWin> i want to batch convert a lot of mkv's from h264 to h265 at half the bitrate [17:36] <NucWin> but they all have slightly different bitrates [17:36] <klaxa|work> well you can use ffprobe to get the bitrate [17:49] <NucWin> no other suggestions before i use perl to do a nasty scrape on the output of ffprobe? [18:57] <diegoviola> hi [18:57] <diegoviola> can i make a gif from a video file? [18:57] <diegoviola> animated gif [19:00] <sacarasc> ffmpeg -i blah.rv -pix_fmt rgb output.gif [19:00] <sacarasc> Maybe? [19:01] <diegoviola> will try that, thanks [19:17] <vlatkozelka> hi , i want to stream a video froma .ts file , the video is h264 encoded , i wanna transcode it to h265 cuz lower bitrate ... i tried -vcodec libx265 but the quality is really poor that the video isnt even showing at the client [19:18] <vlatkozelka> also tried this : ffmpeg -i INPUT -c:a copy -x265-params crf=25 OUT.mov [19:18] <vlatkozelka> but the quality is still really poor [19:23] <sacarasc> What is the client? [19:24] <vlat> oops [19:24] <vlat> <sacarasc> What is the client? [19:24] <vlat> <vlatkozelka> oh [19:24] <vlat> <vlatkozelka> i mean when i played it on vlc or with ffplay [19:24] <vlat> <vlatkozelka> wich is basically simulating a client [19:25] <BtbN> keep in mind that there is no way you will ever be able to encode h265 in real time [19:25] <vlat> im just pasting the conversatio before i DC ed [19:25] <BtbN> sacarasc, he just posted what he already send but got disconnected... pastbining that seems overkill [19:26] <sacarasc> No, pastebinning the whole of ffmpeg stuff. [19:26] <vlat> alright [19:26] <vlat> one sec [19:26] <sacarasc> The output is the important bit, BtbN. [19:26] <diegoviola> ok so that `ffmpeg -i output.mp4 -pix_fmt rgb24 output.gif` works, but how do i make it smaller? [19:27] <sacarasc> diegoviola: Video dimensions or HD space? [19:27] <diegoviola> both [19:27] <BtbN> vlat, if you want to live stream h265, you better have 32 physical cores or more [19:27] <diegoviola> sacarasc: not too small, just a compression of both [19:27] <diegoviola> small compression [19:28] <sacarasc> I think gifs can do -q, also, -vf scale should work as normal. [19:28] <diegoviola> thanks [19:28] <vlat> sacarasc : http://pastebin.com/PitGVyrc [19:28] <vlat> i do have that available [19:28] <vlat> well im just testing one video on my pc but for the real thing i have that processing power available [19:29] <BtbN> you never told it which encoder to use [19:29] <BtbN> so it defaults to mpeg2 [19:29] <sacarasc> Yeah, that. ^ [19:29] <vlat> me ? [19:30] <sacarasc> Yeah, you need -c:v libx265 [19:30] <vlat> i also tried that [19:30] <vlat> the bitrate is really low with that [19:31] <BtbN> That [19:31] <BtbN> 's the whole point [19:31] <vlat> and the video doesnt show on the client [19:31] <vlat> one sec ill try again [19:32] <vlat> oh lol [19:33] <vlat> i see now my cpu jumped to 100% [19:33] <vlat> i guess thats the problem [19:34] <vlat> well ... i will try tomorrow on the server we have at work [19:34] <vlat> thanks for the help guys :) [19:36] <BtbN> i wasn't joking about the 32 cores [19:36] <vlat> yeah i see [19:36] <vlat> that compression must've killed my cpu [19:36] <vlat> but it should work with that command on a better server right ? [19:36] <BtbN> it should allways work [19:36] <BtbN> just be extremely slow [19:37] <vlat> hmm [19:37] <vlat> lowering res and framerate (idc if slow just for a test ) might make it work here [19:37] <vlat> ? [19:37] <vlat> just for the sake of testing it [19:38] <BtbN> transcoding an entire 30 minute video would propably take hours [19:39] <vlat> ahh [19:39] <vlat> hmm [19:39] <vlat> i could instead of live streaming it , transcode it first , save , then stream ? [19:39] <vlat> or would i lose the low bitrate ? [19:40] <BtbN> that's your only option if you don't actualy have 32 fast CPU cores [19:40] <vlat> alrithy then [19:40] <BtbN> Or just go with h264, nothing wrong about it [19:40] <vlat> its wrong when ur boss nags :/ [19:40] <vlat> "PEOPLE ARE USING H265 ... WE SHOULD TOO :@ " [19:41] <vlat> :P [19:44] <vlat> well since u wannna know "whats wrong about it " , in my country internet is real slow so using lower bitrate would give me an advatage ;) [19:44] <Baked_Cake> i can encode a 2 hour movie in x265 in about 30 hourss [19:44] <Baked_Cake> thats with 64 bit ffmpeg [19:45] <Baked_Cake> veryslow setting [19:45] <Baked_Cake> but the result is a work of art [19:45] <Baked_Cake> my cpu hates decoding it tho [19:45] <vlat> haha [19:46] <Baked_Cake> hogs up 20% and spikes at 30% [19:46] <vlat> the last app i made took 60 % [19:46] <vlat> but it was decodeing 30 videos at once and displaying them [19:46] <vlat> and recording [19:46] <vlat> :P [19:46] <Baked_Cake> damn [19:46] <vlat> yeah was for millitary purpose [19:47] <vlat> i love programming for people who have super computers lol [19:47] <vlat> there's just no limits :P [19:47] <Baked_Cake> ya [19:48] <Baked_Cake> gotta get those confesions compressed and sent out to the magistrate [19:52] <BtbN> If you want to stream to normal users, h265 is a poor choice anyway. It easily overloads client machines. [19:52] <BtbN> I'd not use it untill h265 hw decoders have become a common thing [19:58] <vlat> ok :) [19:58] <vlat> thx [20:45] <voip_> hello guys i have same eroors [20:45] <voip_> invalid band type [20:45] <voip_> Error while decoding stream #0:0: Invalid data found when processing input [20:46] <voip_> SSR is not implemented. Update your FFmpeg version to the newest one from Git. If the problem still occurs, it means that your file has a feature which has not been implemented. [20:46] <voip_> [aac @ 0x3d02a20] If you want to help, upload a sample of this file to ftp://upload.ffmpeg.org/incoming/ and contact the ffmpeg-devel mailing list. (ffmpeg-devel(a)ffmpeg.org) [20:46] <voip_> Error while decoding stream #0:0: Not yet implemented in FFmpeg, patches welcome [20:46] <voip_> log: http://pastebin.com/kjvGw71J [21:23] <diegoviola> how do i shrink file size when generating a gif? [21:23] <diegoviola> ffmpeg -i output.mp4 -vf scale=640:-1 -r 10 output2.gif [21:24] <diegoviola> i have this [21:32] <xqdr> hello, has anyone tested https://github.com/drocon11/ffmpeg-qsv ? [21:33] <xqdr> I compiled it on ubuntu x64, kernel 3.8.0-29generic, with IntelMediaSDK2014R2 [21:33] <xqdr> without any problems, but when I try using h264_qsv codec I get a segfault [21:41] <lkiesow> Hi, I already asked this on the ffmpeg-users list but as I didn't get an answer yet, I might as well try it here again. [21:41] <lkiesow> Yesterday a college of mine brought to my attention the following sentence from the FFmpeg wiki: [21:41] <lkiesow> If you use -ss with -c:v copy, the resulting bitstream might end up being choppy, not playable, or out of sync with the audio stream, since ffmpeg is forced to only use/split on i-frames. [21:41] <lkiesow> I'm regularly using bitstream copy (-ss as input option) together with seeking and can't remember having any problems. Is this statement still true? [22:27] <diegoviola> I can't find a way to decrease the image quality in ffmpeg when generating a gif file [22:27] <diegoviola> any ideas [22:27] <diegoviola> I'm looking to get a small file size [22:33] <Nitori> gif is not that flexible. [22:36] <diegoviola> what do you suggest i use instead? [22:36] <diegoviola> webm? [22:37] <thebombzen> lkiesow: bitstream copy means it won't decode the file when it copies (you know this). This means it can't seek to anywhere, but only to the edge of one coding Block. If the video is encoded for streaming, this shouldn't be much of an issue because they have really small blocks. If the video is encoded for storage, you could have very large blocks and that makes the compression ratio better. [22:38] <thebombzen> But it makes seeking way harder. Those are the ones it might mess up on [22:42] <llogan> diegoviola: you'll probably get better results outputting to png then using convert from imagemagick or graphicsmagick to create the gif [22:42] <llogan> AFAIK [22:44] <diegoviola> ok i'll try that, ty [22:47] <lkiesow> thebombzen: But wouldn't that just mean that the seek point is off a bit? So you end up a second before the actual seek point [22:47] <lkiesow> The wiki says it might be not playable [22:53] <voip_> llogan, hello yestorday i asced about 150 livestream transcode processing you said i must use VBV. You tinking i can decrease CPU's usage ? [22:54] <llogan> lkiesow: refer to the docs when in doubt. note that the "Community Contributed Documentation" (which includes the article you referenced) section of the FFmpeg Wiki is often written by users who may not substatiate claims [22:55] <llogan> voip_: using VBV is not about reducing CPU usage, but for providing a reliable stream (as in no buffer waiting except during beginning) [22:57] <llogan> instead you're just declaring an apparently arbitrary bitrate [22:58] <voip_> i know, int not about CPU usage , beacuse a im asking again:) I have no problem with bitrates. I need to buld system for trnscoding 150 streams [22:58] <voip_> i realy dont know how many servers and CPUs i need [22:59] <llogan> you'll just have to run some tests [23:00] <lkiesow> You are right, but there seemed to be a reason in the past why it wouldn't work ao I am a bit worried. And I couldn't find something that made things clear in the official docs. Though from what I read I would *guess* that using -ss as input option together with -c copy is safe [23:01] <lkiesow> &while -ss as output option together with -c copy miht not work [23:01] <voip_> llogan, did you ever benchmarked aproximatly how many streams can processed intel xeon with 6 cores with fast preset ? [23:01] <llogan> lkiesow: i've never really had any trouble with it, IIRC, using it either way [23:02] <llogan> voip_: no. and it depends on too many factors for me to be able to give you an actual number. [23:02] <voip_> ok [23:03] <llogan> just try it [23:03] <lkiesow> llogan: Me neither and I do a lot of encoding :) [23:03] <lkiesow> anyway, thanks for the reply [23:05] <llogan> lkiesow: you can ask the author for more details or why he added that. looks like it was added on version twelve by slhck. his site is slhck.info [23:17] <lkiesow> ok, thanks [23:46] <lkiesow> Seems like the sections was originally introduced on version six by rogerdpack [23:47] <lkiesow> *section [00:00] --- Fri Aug 29 2014

1 0

ffmpeg-devel.log.20140827
by burek 28 Aug '14

28 Aug '14

[02:19] <cone-186> ffmpeg.git 03Michael Niedermayer 07master:9e8ab36f5419: avformat/mxfenc: dont warn about d10_channelcount being ignored if its not set [02:29] <Timothy_Gu> I just noticed that DonDiego never wrote a single en/decoder/(de)muxer. I wonder why he is the person with the second highest number of commits in FFmpeg. [02:30] <Timothy_Gu> Even koda wow one [02:30] <Timothy_Gu> *wrote [02:30] <Timothy_Gu> Damn Android keyboard [02:34] <kierank> Timothy_Gu: cosmetics of course [02:34] <Timothy_Gu> Well I know that [02:34] <Timothy_Gu> And splitting things [02:35] <Timothy_Gu> And fixing the build system although it was a while back [02:35] <Timothy_Gu> But nothing so significant for the features of FFmpeg [02:53] <wm4> splitting crap was probably pretty significant [02:55] <wm4> for what's it worth, I think he did mostly cosmetic commits in mplayer too [02:55] <wm4> and build system things [02:55] <Compn> he split up a lot of dec/enc so that makes ffmpeg faster, sometimes [02:55] <Compn> also writes docs . webpage [02:59] <Timothy_Gu> Compn: I don't recall seeing him writing documentation on libav-devel [03:00] <Timothy_Gu> Compn: how does splitting things make it faster? [03:00] <Compn> well if you load ffmpeg binary that was compiled without encoders... [03:00] <Compn> :P [03:00] <Compn> technically loading a 40mb binary vs 100mb binary :P [03:00] <Timothy_Gu> lol [03:00] <kierank> Compn: you could do that already [03:01] <kierank> load an enc without a dec [03:10] <Timothy_Gu> michaelni: WARNING: Big Xvid C IDCT merge conflict incoming: https://lists.libav.org/pipermail/libav-devel/2014-August/062691.html [06:42] <jamrial> config.h is not an installed header, right? [06:45] <nevcairiel> Right [06:46] <jamrial> great, because lavu/opencl.h, an installed header, includes it [06:46] <jamrial> unless i'm missing something, that shouldn't happen [06:46] <nevcairiel> The whole opencl stuff is majorly screwed up [06:48] <jamrial> yeah [06:50] <jamrial> this basically means that nothing/nobody outside of lavfi really got to use the api at all [07:22] <ubitux> Timothy_Gu: well, he just picked 298b3b6c1f8f66b9bc6de53a7b51d3de745d946b, reindented, and dropped signed-off [07:23] <ubitux> ah, and merged it in another file [07:24] <Timothy_Gu> That's the catch [09:19] <spky_> hi ,guys, i set a timeout for socket, and if av_read_frame return timeout, i want to reconnect it. Q1. when timeout happens, av_read_frame return eof, no flag specify the timeout happens, this will be fine, whatever, i reconnect it Q2. then i call avio_close(format_ctx->pb), avio_open(&format_ctx->pb, filename ,AVIO_FLAG_READ), if open success, then i call av_read_frame, but it just dump the metadata of rtmp, then the function s [09:21] <spky_> any other cleanup or resetup functions should i call for the reconnect? [09:23] <spky_> it just behaves like : myffplay "rtmp://xxxx timeout=3" , and add the reconnect function [10:27] <__gb__> hi nevcairiel, why is this an issue wrt. linking against libva? that's a totally independent library from x11 or wayland backends [10:27] <__gb__> actually, I would like to just move more work into libavcodec, instead of having applications or libraries handle all the same stuff [10:28] <__gb__> i.e. we should normally just need to provide a VADisplay to libavcodec, and the rest (surfaces management, etc.) handled implicitly [11:05] <wm4> __gb__: libva breaks so many abstractions, you can't have libavcocec handle surfaces :D [11:06] <wm4> anyway, that kind of stuff is what lu_zero on Libav wants to do [11:07] <__gb__> please elaborate, I didn't catch everything :) [11:07] <__gb__> e.g. what breaks, what does luca wants to do? [11:08] <wm4> luca wants to move more parts of hwaccel to lavc (for all hwaccel backends) [11:08] <wm4> and the issue with libva at least I see is that all surfaces need to be preallocated [11:08] <__gb__> yes, that's the correct thing to do, along others (like handling fallbacks to SW decently) [11:08] <wm4> if the application keeps a surface longer than expected, things can go wrong [11:10] <__gb__> unfortunately, some pretty ancient HW/driver relies on that behaviour (pre-allocate surfaces within a context) [11:10] <__gb__> though, I personally don [11:10] <wm4> luca's idea has some other practical issues, like the fact that you don't know from the start which hwaccel API you want to use [11:10] <__gb__> 't care :) [11:11] <wm4> hm, so you can just skip the preallocation? [11:12] <__gb__> on intel hd graphics driver, we don't mind. though, I'd say that nowadays with MVC support, that comes to be necessary. I think I can arrange for that [11:13] <__gb__> wm4, for the selection of the hwacell API to use, I had proposed in the v2 an hwaccel_id, which could be quite straightforward [11:13] <__gb__> + a hook similar to get_pix_fmts() -> get_hwaccel_ids() [11:14] <wm4> the thing is, you somehow need to ask the application to create a hw context (like VADisplay) [11:14] <__gb__> though, did not have the time to update that since the past 3 years /o\ [11:14] <__gb__> yes [11:16] <__gb__> for vaapi, at least for any driver depending on libdrm, I foresee ways to not care about that. i.e. have VA display internally allocated + adequate buf sharing interfaces for rendering [11:16] <__gb__> it depends on how much of legacy we want to maintain, for sure, we can do better things [11:17] <wm4> that would be nice (and VDA kind of works this way), but of course the other APIs like vdpau or dxva should also be supported [11:17] <wm4> and the user also must have control over what's created [11:18] <__gb__> yes, that doesn't solve issues for other apis for sure, I was just pointing out that this is technically possible, if needed :) [11:18] <wm4> what if the system has 2 GPUs from different vendors, running on different X servers [11:18] <wm4> would buffer sharing be possible between them? [11:19] <__gb__> if there is kernel support for it, yes [11:19] <__gb__> however, performance wise, this would all depends on other requirements [11:20] <__gb__> (e.g. implicit conversions/copies needed if different tiling, cache invalidation for sync'ing things?) [11:21] <__gb__> I believe, there has been experiments with nvidia gpus (+nouveau), though I don't know the outcome of it [11:21] <wm4> interesting stuff [11:22] <wm4> it would be nice if there were some way of sharing buffers, independent of specific graphics APIs (vaapi, opengl, etc.) [11:22] <__gb__> on linux, that's dma_buf (PRIME fd) [11:22] <wm4> speaking of, is there a way to do opengl interop with vaapi yet? other than going over X11 [11:23] <__gb__> yes, I have to write/clean a sample app for it, support code for libva, driver and mesa (yuv) is available already though [11:23] <wm4> just some documentation would be great already [11:24] <wm4> we've also tried to make vaapi work on wayland, but nothing really worked well [11:25] <wm4> well, I'm off again [11:25] <__gb__> for any new VA-API, I wanted proper Doxygen documentation nowadays. please let me know if the following is not good enough: [11:25] <__gb__> http://cgit.freedesktop.org/vaapi/libva/commit/?h=staging&id=544d31d2fc096a… [11:27] <__gb__> thanks wm4 for your comments, I am finalizing a few deinterlacing fixes & tests, and I should be able to get back to the EGL/vaapi interop stuff + MT issue along the way [11:50] <nevcairiel> __gb__: my problem with linking is that its an extra dependency, since i distribute binaries. If it were like dlopen'ed at runtime, then I could simply test if its available, instead of forcing everyone to have it [12:00] <__gb__> nevcairiel, I don't think it was in lavc (ffmpeg/libav) policy to dlopen shared libs, but your suggestion makes sense too -- better ask/propose a patch to the -dev lists then? [12:00] <__gb__> I personally don't mind either way, as long as the vaapi code still works :) [12:01] <__gb__> (and that we still make efforts in keeping a unique SONAME, which is the case) [12:37] <cone-711> ffmpeg.git 03Anton Khirnov 07master:75bbaf2493a7: mpegenc: limit the maximum muxrate [12:37] <cone-711> ffmpeg.git 03Michael Niedermayer 07master:127aee4ccb5e: Merge commit '75bbaf2493a71ee66eaabe3c21fadd84d07888de' [12:53] <cone-711> ffmpeg.git 03Anton Khirnov 07master:1688eef25385: avconv: fix the muxrate values for -target [12:53] <cone-711> ffmpeg.git 03Michael Niedermayer 07master:ad207f1119d9: Merge commit '1688eef25385089026aba55da1885f70a57815ab' [13:11] <cone-711> ffmpeg.git 03Anton Khirnov 07master:f5245a9c6206: avconv: fix parsing the AVOptions for -target [13:11] <cone-711> ffmpeg.git 03Michael Niedermayer 07master:2082c00ea2ed: Merge commit 'f5245a9c6206878b892adf3ccbccc9311c202af5' [13:14] <ubitux> BBB: can you try http://ffmpeg.org/pipermail/ffmpeg-devel/2014-August/161917.html and see if it's enough to reproduce & fix the regression? [13:39] <ubitux> so what's the plan for lena.pnm? [13:55] <Compn> ubitux : draw on it and call it a parody (fair use copyright) ? :P [13:55] <ubitux> it's already fair use [13:55] <ubitux> the issue is that it's not free and so debian can't distribute it [13:56] <ubitux> s/can't/don't want to/ [13:56] <Compn> ah [13:57] <ubitux> we should just add a lena source filter [13:57] <ubitux> and hide that shit here [13:59] <Compn> -vf lenalives [13:59] <Compn> was that the official debian rejection reason? [14:01] <ubitux> no [14:01] <ubitux> but our packager worries about it [14:01] <ubitux> and i believed that's because he's tired of waiting, so he's looking for issues where there isn't :p [14:29] <cone-711> ffmpeg.git 03Michael Niedermayer 07master:7a67ab5cba8e: ffmpeg_opt: reduce diff to libav in opt_target() by 16 lines [14:36] <pross-au> i'm tired of waiting too [14:54] <nevcairiel> They are intentionally delaying anyway, and at this point its extremely unlikely to get into Jessie, so they have another 2 years to talk about it [15:02] <cone-711> ffmpeg.git 03Jean-Francois Thibert 07master:12df9b9a1510: Improved AC3 decoder level support (heavy drc, dialnorm) [15:28] <Daemon404> kierank, i cannot believe how bad yodel are... [15:41] <kierank> forgot if i posted tihs here: http://www.obe.tv/about-us/obe-blog/item/14-using-opus-audio-in-broadcasting [15:44] <cbsrobot> kierank: a picture showing a opus live stream - seems legit [15:46] <ubitux> kierank: nice [15:46] <kierank> http://www.scei.co.jp/ps4-license/sf_ffmpeg.ZIP [15:46] <kierank> ffmpeg from ps4 [15:46] <ubitux> is that smptehdbars with drawtext? [15:47] <kierank> yes [15:47] <ubitux> \o/ [15:47] <kierank> (my main reason for moving to ffmpeg) [15:48] <ubitux> durandal_1707 will be glad to hear that [15:50] <thardin> nice [15:51] <kierank> in an alternative universe that blog post would be a BS filled white paper [15:51] <kierank> but i couldn't bring myself to do that [15:52] <Timothy_Gu> kierank: nice house [15:52] <kierank> it's a nice place but the owner said she is going to sell it next year [15:52] <ubitux> is that a zotac? [15:53] <kierank> no [15:53] <Timothy_Gu> Hmm, the house is not your own? [15:53] <kierank> i rent it [15:53] <kierank> it's a flat [15:54] <Daemon404> kierank, my landlady recently fell ill and theyre selling the house [15:54] <Daemon404> i have 2-3 months to find a place [15:54] <Daemon404> sucks [15:56] <Timothy_Gu> Our landlady is a ~60-year-old Indian woman whose husband teaches C++ and whose son's startup was bought be Amazon [15:57] <Timothy_Gu> And of course our rent rises each year [15:59] <J_Darnley> That's how money works! [16:00] <J_Darnley> You charge more and more for the same stuff [16:00] <Timothy_Gu> Damn [16:01] <Timothy_Gu> *Damn 'šÀ [16:05] <Timothy_Gu> (I still don't know how to say that word in English) [16:06] <ubitux> https://translate.google.com/#auto/en/%E9%80%9A%E8%B4%A7%E8%86%A8%E8%83%80 ? [16:06] <kierank> inflation [16:11] <Timothy_Gu> Thx [16:12] <Timothy_Gu> There you go [16:12] <Timothy_Gu> It literally means "everything is inflated" [16:12] <Timothy_Gu> ubitux: are you still learning Japanese? [16:13] <ubitux> i stopped a while ago [16:13] <ubitux> will restart soon"®© [16:14] <Timothy_Gu> Kanji too hard? [16:14] <Timothy_Gu> Or too many? [16:15] <ubitux> i don't think that's particularly hard, but it needs a lot of engagement in it (time wise) [16:15] <ubitux> i'm a slow learner, especially when there is no direct usage [16:16] <ubitux> tools like anki are really cool because it makes me realize how stupid i can be [16:16] <ubitux> like, asking 10 times what kanji it is in 5 times and failing everytime i'm asked which it is :p [16:17] <ubitux> in 5 minutes* [16:18] <ubitux> anyway, i need to re-allocate between 2 and 4 hours a day to it [16:19] <ubitux> it might be possible, but i need to schedule it properly [16:19] <ubitux> </mylife> [16:21] Action: Timothy_Gu fails to visualize how ubitux pronounces Kapanese [16:21] <Timothy_Gu> *Japanese [16:22] <ubitux> what? :) [16:22] <ubitux> Japanese is easy to pronounce in French [16:22] <ubitux> (every vocal is available in french) [16:22] <ubitux> there is no tones like in Chinese [16:23] <Timothy_Gu> Tones are easy [16:23] <ubitux> the pronunciation & oral understanding aren't really an issue [16:23] <Timothy_Gu> http://people.wku.edu/shizhen.gao/Chinese101/gift/tone/tones.gif [16:24] <Timothy_Gu> Anyways gotta go to school now [16:24] <ubitux> cya [16:24] <Timothy_Gu> BTW I'm taking Spanish but seems like nobody speaks Spanish here [16:25] <Timothy_Gu> I know reynaldo does and probably DonDiego although he lives in Germany or something [17:01] <Daemon404> kierank, i found a sort-of close hash that it could have been forked from [17:02] <Daemon404> but it's almost like they updated everything independently... [17:05] <kierank> no surprise [17:05] <Diogo> hi i have a mpegts file and i want to cut a specific chunk for hls... [17:05] <Diogo> how can i cut this chunk ? [17:08] <Compn> Timothy_Gu , ubitux : i gave up learning japanese. i dont have the ability to discern all of the syllables. [17:09] <Compn> spanish seems a little slower , maybe easier :) [17:09] <Compn> also i never got started on the kanji, too many characters. [17:14] <Compn> and also there are emphasis on different syllables in japanese, so oba-san vs oBAsan vs OHba-san can meant aunt or grandmother :\ [17:17] <Compn> knew i was in trouble when i couldnt even subtitle a raw anime episode. which i think was the first time i looked for mplayer. because it was the only player that supported rm files and could slow down playback. [17:19] <Compn> so yeah probably no one say that [17:19] <Compn> saw* [17:19] <Compn> derp [18:29] Action: durandal_1707 should license smptehdbars AGPL3 [18:35] <kierank> lol [19:01] <cone-843> ffmpeg.git 03James Almer 07master:d3e9bee6f7e5: lavf/mp3enc: don't abort if audio packets can't be buffered [19:25] <cone-843> ffmpeg.git 03Timothy Gu 07master:25cb697d0c86: bktr: Fix Fabrice's name [20:07] <cone-843> ffmpeg.git 03Michael Niedermayer 07master:7c51f5bd3903: swr: aarch64 audio_convert and neon clobber test [20:41] <cone-843> ffmpeg.git 03James Almer 07master:7997d7864379: lavc/libwebpenc: use WebPMemoryWriterClear() [20:51] <cone-843> ffmpeg.git 03Luca Barbato 07master:ff771f79b55a: vc1: Initialize start_code_found to 0 [20:51] <cone-843> ffmpeg.git 03Michael Niedermayer 07master:a59d922d2488: Merge commit 'ff771f79b55a346b4163d814b58ee4c98114745e' [21:02] <cone-843> ffmpeg.git 03Luca Barbato 07master:404731bd20e1: pulse: Set the wallclock option as default [21:02] <cone-843> ffmpeg.git 03Michael Niedermayer 07master:42196083fe63: Merge commit '404731bd20e1df5880e6fe381e975ba48afc75b2' [21:07] <cone-843> ffmpeg.git 03Diego Biurrun 07master:1f156af4274d: x86: xvid_idct: Drop unused definitions [21:07] <cone-843> ffmpeg.git 03Michael Niedermayer 07master:3ff5ca89fc87: Merge commit '1f156af4274dc72d588620f6bedb4e9e66023c92' [21:16] <cone-843> ffmpeg.git 03Diego Biurrun 07master:dcb7c868ec7a: cosmetics: Make naming scheme of Xvid IDCT consistent with other IDCTs [21:16] <cone-843> ffmpeg.git 03Michael Niedermayer 07master:b3b05a11d31d: Merge commit 'dcb7c868ec7af7d3a138b3254ef2e08f074d8ec5' [21:33] <J_Darnley> If I want to submit a patch which makes some minor language corrections to comments in libavutil/frame.h, what should I prepend to the commit message? Would "lavu/frame.h" be okay? [21:34] <nevcairiel> Drop the .h IMHO [21:34] <J_Darnley> Although I did think that specifying the filename would be a little redundant [21:34] <Compn> either way sounds fine to me :P [21:49] <llogan> J_Darnley: similar typos in libavcodec/msvideo1enc.c and libavcodec/mpeg12enc.c [22:12] <J_Darnley> Stupid find! Why would I want to do different -execute options at the same time? [22:25] <cone-843> ffmpeg.git 03Diego Biurrun 07master:8d27bf1cff35: x86: xvid: K&R formatting cosmetics [22:25] <cone-843> ffmpeg.git 03Michael Niedermayer 07master:5403a288a7fa: Merge commit '8d27bf1cff35be406b0fd89d832e1852d4c573bc' [22:25] <J_Darnley> llogan: I think there is also one in doc/fftools-common-opts.texi [22:38] <J_Darnley> I hope this will be correct useage of --annotate [22:52] <cone-843> ffmpeg.git 03Michael Niedermayer 07master:7a2c380e93f8: avformat/aviobuf: remove uses of deprecated url_feof() [22:54] <jamrial> ^ wonder how i missed those two [23:13] <jamrial> michaelni: is "[PATCH] Adds support for setting aq_mode in libvpx encoder" ok to apply? James Zern hasn't commented yet [23:18] <michaelni> jamrial, if you are sure they are ok sure apply, otherwise ask james zern [00:00] --- Thu Aug 28 2014

1 0

ffmpeg.log.20140827
by burek 28 Aug '14

28 Aug '14

[02:45] <rcombs> anyone seen libx264 giving very low bitrates for the first several seconds of an encode using crf and maxrate? It looks like I'm seeing the first couple scenes (which are rather low-complexity and should be able to be encoded within the VBV boundaries at the given CRF) get forced to a higher CRF, probably because the VBV system isn't performing properly before its buffer fills [03:09] <Idlus> Hi, I want to know whether the use of stdin as input disables the use of other input files? [03:10] <rcombs> Idlus: nope! [03:10] <Idlus> too bad... [03:10] <Hello71> no, but you won't be able to use the keyboard controls. [03:10] <rcombs> &why is that bad? [03:10] <Idlus> I can't pipe audio in and mux it with a video file? [03:10] <rcombs> yes, you can do that [03:11] <Idlus> how then? [03:11] <rcombs> "nope!" meant "no, using stdin as input does not disable use of other input sources" [03:11] <Idlus> ah ok! [03:11] <sacarasc> Idlus: One -i for every input you want. [03:12] <Idlus> well I try with x11grab as video actually, and end up with audio only [03:12] <Idlus> I'll try harder now I know it should work [03:13] <rcombs> post your full ffmpeg command line? [03:15] <Idlus> ok here it is: http://pastie.org/9505566 [03:15] <sacarasc> And the output? [03:16] <Idlus> sorry, hold on [03:16] <llogan> it's obligitory to be asked for command and console output and provide only the command [03:16] <llogan> as per tradition [03:17] <rcombs> llogan: I actually only asked for the command [03:18] <rcombs> so blame me on that one [03:20] <Idlus> hope that's good: http://pastie.org/9505572#5 [03:21] <Idlus> this one actually :) http://pastie.org/9505576 [03:22] <rcombs> what player are you using to check the output? [03:23] <Idlus> vlc [03:23] <rcombs> also, you're re-encoding the audio twice there and that's kinda silly [03:23] <Idlus> oh, true [03:23] <Idlus> i was just testing and forgot to adapt this [03:23] <rcombs> what do you see in VLC? Black video? No video? [03:24] <Idlus> no video apparently, but audio playback [03:24] <rcombs> could you try ffplay, and/or try adding `-pix_fmt yuv420p`? [03:24] <Idlus> wow, indeed! [03:24] <Idlus> for ffplay at least [03:25] <rcombs> sounds like your version of VLC, or its settings, are having trouble with yuv444p H.264 [03:25] <rcombs> (probably) [03:26] <sacarasc> VLC is poo. [03:26] <rcombs> there's advice in the ffmpeg output about this [03:28] <Idlus> yes, but I overlooked it since before, strangely enough, it worked without the audio [03:28] <rcombs> huh, interesting [03:28] <rcombs> might be a VLC bug [03:29] <rcombs> which you might want to report to them [03:29] <Idlus> I'll sure do, for now I try with `-pix_fmt yuv420p` [03:29] <rcombs> cool [03:35] <Idlus> well, I can't seem to understand why, but the `-pix_fmt yuv420p` parameter is accepted without the piping and not with, returning "option not found" [03:35] <Idlus> I'm making a pastebin [03:36] <sacarasc> Idlus: It's an output option, not an input one. [03:36] <rcombs> ^ this is probably your problem [03:37] <rcombs> anyone have any ideas on my VBV issue? (wherein libx264 uses very low bitrates until the VBV buffer is full) [03:37] <Idlus> I do not understand what you mean. I try two commands identic save the piping part. [03:38] <rcombs> post your new entire command [03:39] <rcombs> with the piping [03:39] <rcombs> and output [03:40] <Idlus> here: http://pastie.org/9505594#1,3 [03:40] <sacarasc> Put -pix_fmt just before out.mkv [03:40] <rcombs> yes, move `-pix_fmt yuv420p` to between `-i -` and `out.mv` [03:41] <Idlus> ok I understand, sorry for the stupid issue [03:41] <Idlus> "an output option" indeed [03:43] <Idlus> well, vlc still doesn't account for the video, while ffplay does. [03:43] <Idlus> anyway, that's all I needed to know, thank you! [03:44] <rcombs> hmm, interesting [03:44] <rcombs> either way, blame them [05:54] <Eftekhari> hi [05:55] <Eftekhari> i have ffserver, when feed is larger than 2 GB, if ffmpeg that feed the ffserver exit, re executing the ffmpeg to feed the server not working, with the error : Error reading write index from feed file '/tmp/feed1.ffm': Resource temporarily unavailable [06:12] <Eiam_> whats the "--non-free" flag actually mean when passing it into ffmpeg? [06:12] <Eiam_> I noticed that very few distro package managers include ffmpeg with most of the codecs compiled, only one I could find was arch linux.. [06:19] <klaxa> it enables decoders and encoders which will make the resulting binary undistributable, because of royalties afaik [06:41] <vladkansk> Hi all. I have some problem with ffmpeg. I'm trying to stream multicast FullHD, but I have a lot artefacts in output. When I'm trying to stream HLS - there is no problem, but with UDP Multicast - I have artefacts. Can anyone help me with it? [06:43] <vladkansk> ffmpeg -i udp://@xxx.xxx.xxx.xxx:1234 -vcodec copy -acodec copy -hls_time 5 -hls_list_size 5 /mnt/hls/test.m3u8 works fine, but ffmpeg -i udp://@xxx.xxx.xxx.xxx:1234 -vcodec copy -acodec copy -f mpegts udp://yyy.yyy.yyy.yyy:1234 has a lot artefacts [06:46] <vladkansk> with SD channels 720x576 no problems [06:47] <vladkansk> sorry, with SD channel garbage is lower, almost invisible, but with HD channels all picture is garbage [06:48] <rcombs> vladkansk: sounds more like a network issue than anything [06:49] <rcombs> vladkansk: if you're stream-copying, there's nothing in there that could sensibly introduce artifacts [06:53] <vladkansk> in same network Multicast streams works perfectly =( [06:53] <rcombs> alas, UDP is unreliable by design, and can get very messy, especially across networks [06:54] <rcombs> with TCP, if you drop packets regularly, they'll just get resent; with UDP, if you drop packets regularly, your video ends up looking like a huge mess [06:55] <vladkansk> I have really large network with working Multicast. This network is perfect for multicast. Server with ffmpeg has 10g fiber interfaces to Catalyst 4900M [06:55] <rcombs> can you check if it works locally, when playing back on the same machine? [06:55] <[mbm]> udp is actually prefered for realtime video streaming because it drops packets and doesn't require retransmissions [06:55] <rcombs> [mbm]: assuming the network is reliable, yes [06:56] <rcombs> but if it's not, then you'll have problems [06:56] <rcombs> but it sounds like vladkansk's network is rather reliable, so perhaps we have a bug in the UDP code [06:57] <[mbm]> stream to a pipe and netcat it and see if it behaves any different? [06:57] <rcombs> sure [07:01] <vladkansk> I'l try to watch this UDP stream on server.. [07:12] <vladkansk> anyone tryed to stream 1920x1080 in UDP Multicast? [07:15] <[mbm]> wired connection? [07:15] <[mbm]> (wireless tends to defer multicast to basic rates) [07:16] <vladkansk> no wireless [07:16] <vladkansk> 1gbps or 10gbps [07:17] <vladkansk> I have Ethernet controller: Intel Corporation I350 Gigabit Network Connection (rev 01) on this server [07:17] <[mbm]> vladkansk: I've got an hdhomerun here that will dump the atsc feed to multicast, that works but I haven't tried ffmpeg lately [07:17] <vladkansk> and Ethernet controller: Broadcom Corporation NetXtreme II BCM57810 10 Gigabit Ethernet (rev 11) [07:18] <[mbm]> check the interrupt mitigation settings [07:24] <vladkansk> hm... ffprobe -show_frames -pretty -i udp://@yyy.yyy.yyy.yyy:1234 has no bugs... maybe really interrupts [07:25] <vladkansk> in the server no errors, but in PC directly connected to server there are a lot of [07:25] <vladkansk> [h264 @ 0x1fc5c00] Found reference and non-reference fields in the same frame, which is not implemented. Update your FFmpeg version to the newest one from Git. If the problem still occurs, it means that your file has a feature which has not been implemented. [07:25] <vladkansk> [h264 @ 0x1fc5c00] If you want to help, upload a sample of this file to ftp://upload.ffmpeg.org/MPlayer/incoming/ and contact the ffmpeg-devel mailing list. [07:25] <vladkansk> [h264 @ 0x1fc5c00] decode_slice_header error [07:25] <vladkansk> [h264 @ 0x1fc6360] reference picture missing during reorder [07:25] <vladkansk> Last message repeated 1 times [07:25] <vladkansk> [h264 @ 0x1fc6360] Missing reference picture, default is 65692 [07:25] <vladkansk> Last message repeated 1 times [07:25] <vladkansk> [h264 @ 0x22a7ee0] reference picture missing during reorder [07:25] <vladkansk> [h264 @ 0x22a7ee0] Missing reference picture, default is 65712 [07:25] <vladkansk> [h264 @ 0x22a7ee0] Reference 2 >= 2 [07:25] <vladkansk> [h264 @ 0x22a7ee0] error while decoding MB 7 1, bytestream 24424 [07:26] <relaxed> pastebin.com next time [07:26] <vladkansk> oops, sorry [07:27] <vladkansk> what can I do with this? anyone can help? [07:27] <vladkansk> ls [07:27] <tm512> has anyone else noticed that x11grab performs a lot worse with xorg 1.16? [07:29] <relaxed> vladkansk: are you using a recent version? [07:30] <vladkansk> ffmpeg version 2.3.3 Copyright (c) 2000-2014 the FFmpeg developers [07:30] <vladkansk> built on Aug 27 2014 12:16:55 with gcc 4.8.2 (GCC) 20140120 (Red Hat 4.8.2-16) [07:30] <vladkansk> configuration: --prefix=/usr --enable-gpl --enable-libfaac --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libtheora --enable-libvorbis --enable-libx264 --enable-nonfree --enable-version3 --enable-pic --enable-ffplay [07:32] <tm512> x11grab makes X11's CPU use go up quite a bit more than I remember, and it causes what I'm trying to capture quite a bit of lagginess [07:35] <tm512> can anything be tried to make it perform better? [07:39] <relaxed> tm512: what is your command? [07:41] <tm512> http://hastebin.com/onodugises.bash [07:43] <tm512> it seems to be an X11 issue. the x11grab code doesn't seem to have changed much recently [07:44] <relaxed> try adding -r 60 before the -i :0.0 [07:45] <tm512> what effect will that have? [07:46] <relaxed> sets the input framerate [07:46] <tm512> it seems to have no effect [07:47] <tm512> it also severely desyncs the audio [07:54] <tm512> I wonder if it could be related to shared memory [08:00] <relaxed> simplify your command to see if it makes any difference [08:02] <tm512> I'll cut out the audio but it doesn't really get simpler than that, I don't think [08:08] <tm512> that doesn't help [08:09] <relaxed> what about when you're not streaming it? [08:09] <tm512> I'm not streaming, I have rtmpurl set to an flv file in a tmpfs [08:32] <tm512> I restarted X [08:32] <tm512> now I get [x11grab @ 0xa7a0c0] Could not open X display. [08:35] <tm512> but I'm not typing the command in to the console [08:37] <tm512> I'll put set -x in my script [08:38] <tm512> http://hastebin.com/ohosegedex.txt [08:41] <tm512> ah. here DISPLAY has changed to :1 [08:41] Action: tm512 shrugs [08:43] <tm512> Xorg added systemd support, so I think arch linux has brought that in [08:43] <tm512> why I am using this distro still, I have no idea. upgrades are an endless source of frustration [08:44] <relaxed> Debian is nice [08:45] <tm512> I've been thinking of switching to LFS with the stable releases of pkgsrc [08:45] <K4T> I am trying to use pullup filter to convert video from 29.970 to 23.970 fps. Is it a good way to do it? I asking cause my output video have minimal stutter [08:45] <K4T> after conversion [08:46] <K4T> -vf pullup,scale=1280:-2,pad=1280:720:0:(oh-ih)/2,fps=fps=24000/1001 [08:46] <K4T> here is my filter settings [08:46] <K4T> are* [09:01] <tm512> interesting. it's only doing this when trying to record fceux, so far [09:06] <reti> so I am trying to batch rip subtitles off of these videos, I have thise script and the output is after, it's saying it failed to set -c? http://pastebin.com/SnPTUWdX [09:13] <spky_> myffplay "rtmp://xxxx timeout=3" hi ,guys, i set a timeout for socket, and if av_read_frame return timeout, i want to reconnect it. Q1. when timeout happens, av_read_frame return eof, no flag specify the timeout happens, this will be fine, whatever, i reconnect it Q2. then i call avio_close(format_ctx->pb), avio_open(&format_ctx->pb, filename ,AVIO_FLAG_READ), if open success, then i call av_read_frame, but it just dump the metadata [09:16] <spky_> av_read_frame dump the metadata of rtmp, then stucks, can anyone give me a little help please [09:17] <K4T> how can I check if video is interlaced or not? Is it possible to do this with ffmpeg? [09:41] <K4T> can I use CUDA with ffmpeg? [10:38] <Vish> Hi - when running ffmpeg - I get error audio stream missing - Can some one help? Details of the problem on this link - http://pastebin.com/r2mCZS4f [10:45] <Vish> if i put mpegvideo which is without audio per documentation then it does not recognize the format... so really struck up.. [10:55] <Vish> can anyone help - Unfortunately I am also not allowed to make a new post on forum.. [10:58] <relaxed> Vish: what are you doing with ffserver? [10:59] <Vish> so that I could access the stream through web browsers... [11:10] <Vish> relaxed: do you see any concerns in my approach? [11:15] <relaxed> You should see if ffmpeg can do what you want. [11:15] <relaxed> Not many people use ffserver. [11:17] <saste> Vish, NoAudio is missing [11:22] <Vish> saste: I put no audio and I was able to move forward. Now it throws different error.. [11:22] <Vish> image2 @ 0x7fa5e380a600] Could not get frame filename number 2 from pattern 'mjpeg' (either set updatefirst or use a pattern like %03d within the filename pattern) av_interleaved_write_frame(): Invalid argument frame= 2 fps=0.0 q=1.6 Lq=1.6 size=N/A time=00:00:00.08 bitrate=N/A video:11kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: unknown Conversion failed! [11:23] <saste> Vish, you can't use image2 with a pipe [11:24] <Vish> saste - ./getJpgStream | ffmpeg -f image2 mjpeg -i - http://localhost:9090/feed1.ffm -> here getJpgStream gives me mjpeg format how should I pipe it to ffmpeg? [11:25] <saste> Vish, did you try with image2pipe?? [11:25] <Vish> let me try that [11:28] <Vish> saste - Great help! yes now I am able to receive it.. But when I open in browser as http://localhost:9090/feed1.ffm -> I get error that feed is already being received? [11:39] <Vish> saste - sorry my bad I had to call the stream - http://localhost:9090/test1.mpg - chrome opened quicktime but still cant see the video.. let me try some thing else.. [13:08] <Ol1> Hi - Does anyone know if it's possible to set a mov_text subtitle to forced? [15:18] <K4T> ffserver sometimes not creating VIDEO stream, onlu audio - why? [15:18] <K4T> is this known bug? [15:18] <hendry> when i try playback videos i've convert to mp4 from my iphone they are upside down. is it mplayers fault for not auto-rotating them or ffmpeg? https://trac.ffmpeg.org/ticket/515 [15:20] <mistawright> hi guys I need some help. I have been trying to figure out how to make transitions happen between each image in a slideshow that was created with ffmpeg. currently. I have 12 images displayed for 5 seconds a piece. Am I able to use a kenburns transition and a fade in/out with a slideshow created from images? [15:20] <mistawright> I have not been able to figure it out [15:20] <spaam> easy fix: rotate the screen while watching the video :D [15:21] <klaxa|work> i don't think ffmpeg has transition video filters [15:21] <klaxa|work> is there metadata for video rotation? like exiv data for images? [15:22] <K4T> nope [15:22] <klaxa|work> if not use the proper rotate video filter while encoding [15:22] <spaam> klaxa|work: in the .mp4 container i think there is some metadata for it [15:23] <klaxa|work> oh yes, see: http://stackoverflow.com/questions/9408542/video-captured-from-iphone-gets-… [15:23] <spaam> and yes there is some metadata for it [15:24] <klaxa|work> apparently apple specific [15:25] <klaxa|work> people should stop recording vertical videos in the first place [15:26] <spaam> klaxa|work: yes. but they got this VVS thing. Vertical Video Syndrome. https://www.youtube.com/watch?v=Bt9zSfinwFA [15:34] <mistawright> how would you guys suggest I get the transitions working? [15:35] <mistawright> how would i go about using the blend option between image [15:37] <klaxa|work> oh that actually looks like what you want, maybe just try out the examples from the documentation and change the parameters until you like it? [15:37] <klaxa|work> http://ffmpeg.org/ffmpeg-filters.html#Examples-39 [15:40] <mistawright> klaxa|work, I am extremely new to ffmpeg and barely getting chaining etc. so I am not sure how I would have this effect apply to my initial input of images that it makes the video from [15:41] <klaxa|work> maybe try using two pictures first, like this: [15:42] <klaxa|work> ffmpeg -i img_0.jpg -i img_1.jpg -filter_complex "[0][1] blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'" out.mkv [15:42] <klaxa|work> but that still lacks some framerate filtering, not sure how to apply that [15:43] <Maverick|MSG> if I have a wav that has incorrect header information (like missing data size) is there a way to make ffmpeg repair/ignore this and continue processing the file? [15:44] <klaxa|work> maybe like this: ffmpeg -r 1/5 -i img_0.jpg -i img_1.jpg -filter_complex "[0] fps=30 [0_edit];[1] fps=30 [1_edit];[0_edit][1_edit] blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'" out.mkv [15:44] <klaxa|work> i'm not sure if that works even [15:48] <mistawright> hmm. i'll play around with it [15:51] <mistawright> klaxa|work, http://pastebin.com/fJW9DNHt that is the command i am using. I am unclear of how would I reference my first input which is the series of images. [16:02] <klaxa|work> well what's the ffmpeg output? [16:03] <klaxa|work> cc: mistawright [16:11] <mistawright> klaxa|work, http://pastebin.com/7vSzT4vN here is the initial output to create the slideshow with overlays [16:11] <Maverick|MSG> to answer my own question: is there a way to make ffmpeg repair/ignore this and continue processing the file? [16:11] <Maverick|MSG> you use the -f setting [16:11] <Maverick|MSG> in my case: -f s16le [16:11] <Maverick|MSG> hopefully that helps someone [16:14] <klaxa|work> ah now i see, you haven't used the blend thing yet. i'm not sure you can do that by referring to the images as one video stream, as the blend filter needs two video streams [16:14] <klaxa|work> you will probably need to use every image as a single input video, you might want to write a script to build a command line for that [16:18] <mistawright> to affect the images for the slideshow would i refer to the [0:v] label to use with the blend option [16:20] <klaxa|work> like i said, that will not work, blend needs two inputs to generate one output [16:22] <mistawright> what would be my alternative then. i was hoping to use a crossfade between each image [16:23] <klaxa|work> (16:14:57) klaxa|work: you will probably need to use every image as a single input video, you might want to write a script to build a command line for that [16:31] <mistawright> yeah probably gonna have to use imagemagick to create the basic slideshow with fades etc and then use ffmpeg for overlays and text and final output [16:32] <darsie> hi [16:32] <darsie> How do I list the streams in a file with avconv? [16:33] <darsie> or possibly other tools [16:33] <darsie> I want to know what there is so I can specify it and decide what to do with it. [16:35] <klaxa|work> try #libav [16:35] <klaxa|work> or get ffmpeg. ffprobe can list information about mediafiles [16:36] <darsie> thx [17:05] <Diogo> <Diogo> hi i have a mpegts file and i want to cut a specific chunk for hls... [17:05] <Diogo> 16:05:43 <Diogo> how can i cut this chunk ? [20:04] <mistawright> I havent found a way yet and have been searching for about a week. I need to assemble images into a slideshow and also have transitions, fade in fade out, between the images. How can I go about this? [20:04] <mistawright> i have tried blending unsuccessfully [20:20] <xavery> the 5 files on my DVD contain 3 titles. I want to rip only the first title, which takes up whole VTS_01_[1-3].VOB, and a part of VTS_01_4.VOB. how can I limit ffmpeg to process only the first title? [20:34] <llogan> Diogo: -ss and -t or -to [20:35] <llogan> xavery: you can try to use concat demuxer in combination with -ss and -t or -to [20:37] <xavery> llogan, thanks, I'll do that. [20:38] <llogan> mistawright: as klaxa mentioned you will probably have to treat each image as an individual input instead of a sequence or pre-process the images with [image|graphics]magick or similar [20:39] <llogan> or maybe you can do something with sendcmd filter, but i haven't really even looked at that (I should) [20:39] <llogan> xavery: http://ffmpeg.org/ffmpeg-formats.html#concat [20:39] <llogan> https://trac.ffmpeg.org/wiki/How%20to%20concatenate%20%28join,%20merge%29%2… [20:40] <xavery> llogan, thanks. I guess that's the best solution, especially when I use lsdvd for obtaining information about the title's length. [20:40] <llogan> also https://trac.ffmpeg.org/wiki/Seeking%20with%20FFmpeg [20:45] <Diogo> llogan: width -ss and -t or to the accurate of seek is not very good.. [20:45] <Diogo> to generate a specific ts file for hls ... [20:45] <Diogo> playlist [21:47] <vlatkozelka> hi .. is it possible to stream multiple files onto one stream ? [22:06] <llogan> vlatkozelka: what do you mean exactly? [22:08] <vlatkozelka> lets say i have 2 .mp4 files , i need to creat a stream with 2 programs , one for each file [22:08] <vlatkozelka> soory for delayed reply , i was reading some docs [22:09] <llogan> you can place multiple streams into a file, but maybe i don't correctly understand what you want [22:09] <vlatkozelka> something like [22:10] <vlatkozelka> ffmpeg -i input1 -i input2 ... onto one stream so Program 1 would play input1 and Program 2 would play input2 [22:11] <tm512> it seems like the x11grab lag can be fixed by switching to the new glamor acceleration method [22:11] Action: tm512 shrugs [22:14] <xavery> vlatkozelka, seems like what you want is -map. have a look : http://ffmpeg.org/ffmpeg.html#toc-Advanced-options [22:14] <vlatkozelka> yeah i read it and tried stuff [22:14] <xavery> keep in mind that your output format must support multiple "programs" inside one file. [22:14] <vlatkozelka> -f mpegts [22:14] <vlatkozelka> on udp [22:14] <vlatkozelka> yeah [22:14] <vlatkozelka> but [22:15] <vlatkozelka> i know how to use -map to extract a program from an already made stream ... not the opposite :/ [22:18] <xavery> vlatkozelka, google seems to suggest that it isn't supported. refs : http://ffmpeg.org/pipermail/ffmpeg-user/2012-April/005943.html , http://forum.videohelp.com/threads/360536-How-do-I-use-FFMPEG-to-output-an-… [22:19] <vlatkozelka> ah [22:19] <vlatkozelka> so it only works the opposite way it seems [22:19] <vlatkozelka> i think the other way around is hardware [22:20] <xavery> I'm pretty certain that you could do it in software. after all, there exists software that does just that. it's just something that's not supported by ffmpeg. [22:20] <vlatkozelka> i know how to code it myself in JAVA [22:20] <xavery> *cough* [22:21] <vlatkozelka> but that'l take a month of coding :/ [22:21] <vlatkozelka> meh... i suppose i should get to work lol [22:22] <xavery> vlatkozelka, you could use mpegtsmux from the GStreamer project in order to get multi-program Transport Streams. however, it will require you to write a custom application around the element anyway, since it requires passing a pointer to a GstStructure as a property in order to obtain stream->program mappings. [22:22] <vlatkozelka> yea [22:22] <vlatkozelka> i have some other API like that called Xuggler [22:23] <vlatkozelka> ill have to create a container and write metadata to it then add streams [22:23] <xavery> use whatever you're comfortable with. [22:23] <vlatkozelka> but thats like re-inventing the wheel , i hope ffmpeg would save me like it saved me on my last project [22:24] <vlatkozelka> anyways [22:24] <vlatkozelka> thanks for ur help :) [22:24] <xavery> vlatkozelka, you're always welcome to write multi-program TS muxing yourself. [22:25] <vlatkozelka> not from scratch lol [00:00] --- Thu Aug 28 2014

1 0

ffmpeg-devel.log.20140826
by burek 27 Aug '14

27 Aug '14

[00:01] <wm4> looking at the code, maybe not [00:01] <pross-au> wm4: do they still forbid mp3? [00:01] <wm4> pross-au: no idea if that's a thing of the past [00:02] <wm4> Timothy_Gu: so I suppose you don't really need to, but checking it anyway might be better style [00:03] <Timothy_Gu> I am asking because there are _tons_ of unchecked avio_tell() calls in matroskadec.c [00:04] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:2a85826e5753: avcodec/vp9: Use av_malloc_array() [00:04] <cone-798> ffmpeg.git 03Christophe Gisquet 07master:6ee7681723a4: huffyuvenc: write last odd sample [00:04] <cone-798> ffmpeg.git 03Christophe Gisquet 07master:f75baa6c9b17: huffyuvdec: decode the last odd sample [00:04] <kepstin-laptop> pross-au: as of fedora 20, they still don't include mp3 decoding. [00:04] <Timothy_Gu> pross-au: at least we have Fluendo now [00:06] <wm4> Timothy_Gu: then maybe it would be better to adjust the documentation to say it never fails [00:07] <Timothy_Gu> Does it not? [00:07] <Timothy_Gu> Well this was my original question: does avio_tell() ever makes avio_seek() fail? [00:07] <Timothy_Gu> *make [00:08] <beastd> I would rather say. Think what would happen if it fails -- what would the code of the mkv demuxer do? [00:08] <beastd> Especially if it returns either error or position [00:10] <Timothy_Gu> ... do weird stuff and not demux properly? [00:10] <beastd> Might be OK to ignore errors, but only if you can say the code will still behave safe in that case. [00:12] <Timothy_Gu> For example, http://git.videolan.org/?p=ffmpeg.git;a=blob;f=libavformat/matroskadec.c;h=… passes a potential negative value into resync() [00:12] <Timothy_Gu> which uses the value in avio_seek(), and seek backwards instead of forwards [00:13] <Timothy_Gu> And all kinds of weird problems may appear [00:31] <beastd> Timothy_Gu: that could possibly checked inside the internal matroska_resync . also other cases are just for error messages (though i am not sure they are treated as unsigned for error output). you could also write a mail to ffmpeg-devel and ask if others have better ideas how to proceed. i would examine each instance, categorize and then do fixes if needed [00:32] <Timothy_Gu> Thanks! I am a little bit short on time right now, but I will do what you suggested once I find time. [00:33] Action: beastd too [00:34] <beastd> Will go to sleep soon. [00:39] <cone-798> ffmpeg.git 03Alexander Strasser 07master:db85d11d9d88: libavformat/ftp: Do not leak memory in routine ftp_features [01:20] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:e96109f93ce9: ffplay: use av_malloc_array() [01:20] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:1f7e6c07130c: doc/snow: remove possibly wrong todo item [01:49] Action: cbsrobot_ wonders what is so patentable in this "Dreamworks Lossy Compression" [01:50] <cbsrobot_> https://github.com/openexr/openexr/blob/master/OpenEXR/IlmImf/ImfDwaCompres… [01:50] <cbsrobot_> looks like just another jpeg implementation [01:55] <Daemon404> of course [01:55] <Daemon404> it is patented after all [01:55] <Daemon404> of course it is not novel. [01:56] Action: llogan prefers lena over flower [01:58] <wm4> so who will implement native openexr in ffmpeg? [01:59] <Daemon404> exr.c is a thing [01:59] <Daemon404> just needs the mode [02:00] <Daemon404> im impressed how large this file is [02:00] <Daemon404> for what it does... [02:12] <wm4> lol no more lena [02:17] <Daemon404> wtf is even patented in this [02:22] <llogan> the 1970's [02:22] <Daemon404> my only guess its its dumb ad-hoc quant method [02:22] <Daemon404> is* [02:26] <Timothy_Gu> Seems like Andreas forgot my copyright... [02:29] <michaelni> llogan, if you dont lie flower.pnm, suggest another image [02:29] <michaelni> maybe bikeshed.pnm ;) [02:30] <llogan> who knows...maybe i'll actually take a photo of a bikeshed [02:30] <wm4> but what color will the shed have [02:30] <wm4> *should [02:34] <michaelni> Timothy_Gu, where did he forget it ? in git or that patch ? anything i can fix ? [02:35] <Timothy_Gu> In debian/copyright [02:35] <michaelni> ahh ok, i cant help with that i think [02:35] <Timothy_Gu> Is Lena still alive? [02:36] <michaelni> dunno, she was a few years ago though AFAIK [02:52] <cone-798> ffmpeg.git 03ThomasVolkert 07master:96b2ba68c4ae: avformat/rtpdec: support for HEVC/H.265 RTP payload format (draft v6) depacketizing [03:46] <jamrial> michaelni: what libvpx version do you have? VP9E_SET_AQ_MODE is available in 1.3.0, the first one with vp9 en/decoder [04:26] <rcombs> anyone around who can tweak fflogger? It's not zero-padding color codes to 2 digits, so names that start with [0-9A-Fa-f] throw off coloring [04:27] <llogan> that's burek but he's MIA [04:27] <rcombs> and a name starting with ",2" or something (which would be pretty weird, but not unheard of) would probably screw up the background color as well [04:27] <llogan> burek021 at gmail [04:27] <rcombs> alright, I'll mail him [04:27] <llogan> thanks. [04:28] <llogan> also, i don't think you need to cc yourself to your own bug reports. i think it will automatically email you any updates [04:28] <rcombs> I'm not sure, as it varies between trackers; there was a checkbox to CC myself, so I clicked it [04:28] <llogan> no big deal. just letting you know. [04:29] <rcombs> figured it wouldn't be there if it wasn't necessary [04:29] <rcombs> yeah, gotcha [04:29] <llogan> burkes builds have been halted since 16 july. asked him twice if he could resume them. no reply, so once relaxed starts providing 32 bit builds i'll update downloads [04:30] <rcombs> 01,00hmm, is this default or not [04:30] <rcombs> nope [04:36] <rcombs> well, I won't hold my breath for a response, but sent [04:49] <michaelni> jamrial, iam not sure but ive a source tree for v1.2.0-4277-gc73e441 here so maybe that [04:53] <jamrial> ah, a git snapshot. makes sense there's stuff missing [04:53] <jamrial> we should probably update the configure check, then. make it check for VP9E_SET_AQ_MODE instead of VP9E_SET_SVC [05:53] <cone-798> ffmpeg.git 03James Almer 07master:c2c56d54ee90: configure: update libvpx_vp9_encoder check [09:23] <ubitux> michaelni: [09:23] <ubitux> Applying: tests: switch to a test image that is under public domain [09:23] <ubitux> fatal: corrupt patch at line 3322 [09:30] <ubitux> michaelni: could you make flower.pnm available somewhere? [10:27] <saste> how can I know which metadata fields are supported in a given format? [10:28] <saste> unsupported fields are simply ignored at the moment [10:48] <ubitux> saste: wm4 asked something similar yesterday [10:48] <ubitux> currently we store random tags in a given format from metadata even though they sometimes don't make sense [10:48] <ubitux> are you talking about that? [10:49] <saste> ubitux, and my answer is, there is no automatic way [10:49] <saste> ubitux, yes more or less [10:51] <saste> also I'm discovering that sometimes output metadata is silently discarded [11:10] <michaelni> ubitux, http://ffmpeg.org/~michael/flower.pnm [11:10] <ubitux> thx [11:11] <ubitux> cool ok [11:29] <nevcairiel> lena is not free to use? everyone uses it, we even used it in university :p [11:32] <ubitux> not "DFSG-free" apparently [11:33] <nevcairiel> they wouldnt accept free beer in a pub if it doesn't come with a written letter that you relinquish all rights to it [11:33] <ubitux> it seems to be potentially problematic for debian [11:34] <nevcairiel> apparently so [12:24] <cone-840> ffmpeg.git 03Diego Biurrun 07master:ab56fabe6294: vfwcap: Add fallback define for HWND_MESSAGE [12:24] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:d5ee74e57d44: Merge commit 'ab56fabe6294524e99815451ad01e4ff50c6d734' [12:34] <pross-au> a flower? [12:36] Action: ubitux sends https://www.youtube.com/watch?v=AULG4MoYxQk to pross-au with love [12:48] <J_Darnley> > vfwcap: Add fallback define for HWND_MESSAGE [12:48] <J_Darnley> > Some obsolete versions of the MinGW32 runtime (<4.0.0) lack the definition. [12:48] <J_Darnley> Wasn't that the point of dropping the fallback? [12:49] <nevcairiel> he didnt expect it to actually break anything :p [13:06] <cone-840> ffmpeg.git 03Luca Barbato 07master:a4d3c2003594: vc1: Fix the skip condition [13:06] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:ba650ea11825: Merge commit 'a4d3c20035946cbc1509aec2dc28d51c2a2f9a8e' [13:32] <Eftekhari> hi [13:34] <Eftekhari> i have ffserver, when feed is larger than 2 GB, if ffmpeg that feed the ffserver exit, re executing the ffmpeg to feed the server not working, with the error : Error reading write index from feed file '/tmp/feed1.ffm': Resource temporarily unavailable [14:13] <cone-840> ffmpeg.git 03Luca Barbato 07master:4e9e6fa99f3f: mpeg: Write H264 streams at offset 2 [14:14] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:cf0e8e7ad488: Merge commit '4e9e6fa99f3ff83cedbf6c378d62065ae419a3b9' [14:47] <durandal_1707> doesnt libav use lena? [14:51] <nevcairiel> Probably, its been part of fate for ages [14:55] <michaelni> Eftekhari, ask reynaldo / wait till he is here [15:30] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:85f0bde4f07d: avcodec/snowenc: remove out-commented assert [15:30] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:bf16872feca8: avformat/nsvdec: fix out-commented asserts so the function names exist [15:30] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:46ad2c4aeddf: avformat/utils: remove assert that tests the same condition as the if() directly above [15:42] <ubitux> lol blackberry mail [15:46] <nevcairiel> Fun [15:48] <Daemon404> what is this ridiculous lena removal patch? [15:49] <Daemon404> oh. debian. [15:49] <Daemon404> fuck them. [15:49] Action: Daemon404 sighs at freetards [15:52] <nevcairiel> They are the best kind of people [15:57] <Daemon404> let's not use Lena, the standard test image, because IT IS NOT FREE ENOUGH [15:57] <ubitux> let's hide it in a xor-ed .c [15:57] <ubitux> and generate it on the fly [15:57] <Daemon404> it seems beyond stupid to change the regression testing image for DFSG... [15:57] <Daemon404> users dont get that [15:57] <Daemon404> and the debian daemons have no internet [15:57] <Daemon404> so... [15:57] <Daemon404> (a run with just asynth/vsynth is kinda dumb) [16:00] <ubitux> i guess our packager is getting bored after a few months waiting for the ftp-masters to get the shit done [16:00] <nevcairiel> Yeah its not even distributed by debian [16:00] <nevcairiel> Whz do they care again [16:00] <nevcairiel> Why* [16:01] <ubitux> it's only present in the sources but they do actually distribute the sources [16:01] <Daemon404> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=758442 [16:01] <Daemon404> freetards gonna freetard [16:01] <Daemon404> it is utterly inconsequential [16:01] <nevcairiel> We should tell them that there is hundreds of patents on the things avcodec implements [16:01] <nevcairiel> See the chaos ensue [18:48] <Daemon404> ugh [18:48] <Daemon404> i see we dont use pkg-config for e.g. vorbis either [18:48] <Daemon404> sigh. [18:58] <ubitux> wanna try to switch to? [18:58] <ubitux> ;) [18:58] <ubitux> i'm fighting for months to get it for libx264 [19:05] <Daemon404> ubitux, i dont feel like dealing with an autistic person [19:20] <Daemon404> man i feel silly [19:20] Action: Daemon404 uses -c:v libxvid [19:47] <llogan> Daemon404: why libxvid? [19:50] <Daemon404> reasons [19:50] Action: llogan leaves to have more engaging discussion with a wall [20:06] <wm4> so I guess this API mailing list idea failed... because of Libav [20:07] <Daemon404> of course [20:08] <llogan> maybe the formatting of the message was not correct [20:15] <cone-840> ffmpeg.git 03Hii 07master:604c4eab2bf6: libx264: fix -b_qfactor and -chromaoffset [20:25] <jamrial> wm4: what did you expect? they know they control the api as long as ffmpeg tries to be compatible because debian. why would they play ball unless they are forced to? [20:29] <kierank> wm4: of course because nobody talked to libav about it... [20:30] <kierank> which was my exact point [20:30] <kierank> the other day [20:30] <kierank> jamrial: they control the api because ffmpeg merges from libav [20:30] <kierank> not because of debian [20:31] <wm4> kierank: I mentioned it on irc, but nobody cared [20:31] <kierank> because it is an ffmpeg initaitve [20:32] <kierank> my point to ubitux is you have to do these things as a joint initiative [20:32] <wm4> but Libav was always blaming ffmpeg for not making a step towards cooperation [20:32] <kierank> which is would ideally be a result of a face to face conversation [20:32] <kierank> wm4: libav is blaming micahel [20:32] <kierank> which is largely true [20:33] <wm4> in this case ubitux just made a suggestion, how else would he even start a "joint initiative"? [20:33] <wm4> but Libav remained silent [20:33] <kierank> by turning up in a neutral environment (vdd) and proposing it [20:33] <kierank> ubitux was saying MLs are the best place for these things etc [20:36] <wm4> and this is your proof that it doesn't work? [20:39] <kierank> the opposite [20:39] <cone-840> ffmpeg.git 03ThomasVolkert 07master:e15824e75b55: avformat/rtpdec_h261: Fix sanity checks [20:39] <kierank> ubitux was saying all the issues could be solved on the ML and not face-to-face [20:39] <kierank> and he said what kinds of things could be discussed face to face [20:39] <kierank> i proposed an api mailing list [20:39] <kierank> he said he's send an email [20:39] <kierank> he'd* [20:39] <kierank> and he did and libav ignored it as expected [20:41] <llogan> and what did we learn here? never try. [20:41] <wm4> and you're claiming it magically works better in the real world [20:41] <j-b> ,agic [20:41] <wm4> the same where Diego's appearance could apparently silence the whole table (lol) [20:42] <j-b> wm4: tbh, being face to face improves communication. [20:42] <j-b> wm4: so, when do we start our fork? [20:42] <wm4> fork of what? [20:42] <j-b> ffmpeg :) [20:43] <wm4> yes you should totally fork it! [20:43] <kierank> I hope you say both ffmpeg and libav and avconv are deprecated [20:43] <j-b> kierank: ah, that's sure. [20:43] <j-b> kierank: I would remove a lot of things. [20:43] <wm4> like what? [20:43] <llogan> kierank: and "no longer developed anymore" [20:43] <kierank> like what? [20:44] <j-b> libavfilter, libavresmaple, libpostporc, libwsresample, libwscale [20:44] <j-b> for a start [20:44] <kierank> lol [20:44] <wm4> but not libavdevice? [20:44] <j-b> then, ffmpeg, ffserver, ffprobe and the rest [20:45] <kierank> j-b: not libavformat? [20:45] <j-b> wm4: it seems too tangled with libavformat [20:45] <j-b> kierank: idem [20:45] <wm4> and the removed stuff will actually end up in separate repos, right? [20:45] <j-b> so, a core repo with libavcode+liavformat+libavutil [20:45] <j-b> wm4: yes. [20:45] <wm4> sounds sensible [20:45] <j-b> wm4: so, the tools will not violate the internal symbols [20:46] <j-b> and libav/ffmpeg developers will finally eat their own shit about APIs [20:46] <wm4> one small problem is that you actually need all repos from the start, because the FATE depends on all the things [20:46] <j-b> and have ffmpeg and ffplay to work with 2 or 3 versions of libavcodec and libavfilter [20:46] <wm4> oooh you're cruel [20:46] <nevcairiel> I think the .v files are pretty much empty by now, at the very least [20:47] <wm4> that would force the devs to actually care about compat [20:47] <j-b> wm4: so? it's not hard to split. and use subfilter. [20:47] <j-b> oh, and tools also would split. [20:47] <nevcairiel> But it also means they can't use any new APIs, which is not very good for innovation [20:47] <j-b> Bullshit. [20:47] <j-b> it's called ifdef [20:48] <cone-840> ffmpeg.git 03Diego Biurrun 07master:4c8bd8ddb049: os_support: Adjust an outdated #endif comment [20:48] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:7206221d563e: Merge commit '4c8bd8ddb049950347a5018fecbca7ee25d48c44' [20:48] <j-b> and you should know that "innovation" is a bullshit word. [20:48] <nevcairiel> If you want half the thing not working with some old version, sure. :p [20:48] <j-b> half the things? [20:48] <j-b> exageration much? [20:49] <j-b> we manage to compile VLC with lavc HEAD and a version from debian stable. [20:49] <wm4> what j-b suggests sounds pretty much how things ideally should be [20:49] <j-b> without too much hassle [20:49] <j-b> AND, we compile against 3 forks [20:50] <j-b> so please, no "innovation blabla" BS, to hide laziness and go on abusing symbols. [20:50] <iive> 4 [20:50] <nevcairiel> There is 3 forks now? [20:50] <iive> ffmb [20:50] <j-b> and sorry, for once, I'm not smooth. [20:50] <j-b> iive: 4? [20:50] <wm4> and of course half a dozen of open or closed projects with their own changes [20:51] <j-b> yeah [20:51] <nevcairiel> Those mostly don't go revising API much [20:51] <iive> bcourdier have its own fork. gpl probably still uses the original api. [20:52] <iive> imho, if libav wanted ffmpeg to stop merging, they should relicense it as GPL only. [20:52] <j-b> and, as I said, one maintainer for each submodule. [20:52] <j-b> iive: they don't wnat ffmpeg to stop merging, they just ignore and not care about ffmpeg. [20:53] <iive> and please steamline the api. no more than 4 function. open, close, decode, encode. [20:53] <cone-840> ffmpeg.git 03Gabriel Dume 07master:56a721f02027: doc: fix a typo [20:53] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:ffa90d99fd89: Merge commit '56a721f020273d69daa8dcb0d99e42a43a0a0d4d' [20:54] <nevcairiel> I always have to grin when they talk about an issue long solved in FFmpeg but it would never even occur to them to check [20:54] <j-b> nevcairiel: indeed. [20:54] <j-b> nevcairiel: however, usually their fixes are better. [20:54] <wm4> j-b: how to make sure that the fork won't fall into obscurity from the start? [20:54] <wm4> it's hard to force others to their luck [20:54] <j-b> I have more security issues on FFmpeg than libav, for example [20:54] <iive> wm4: simply, by replaceing both libav and ffmpeg in debian! [20:55] <j-b> wm4: that's the hard part :) [20:55] <wm4> and who will do the grindwork of cherry-picking patches from ffmpeg/libav (and distributing them over 6 repos ro so) [20:55] <wm4> *or [20:55] <j-b> someone paid fulltime [20:55] <iive> you can ask michael :P [20:56] <j-b> I already said that at the time of the move to git [20:56] <j-b> and at the time of the split [20:56] <j-b> so, it has not changed. [20:56] <nevcairiel> That doesn't sound like a fun job, merging stuff for a living [20:56] <JEEB> I at one point wanted to fix a thing in the MXF (?) demuxer or so, and of course wanted to have it fixed in both, so I decided to first update it into libav and then it would flow into FFmpeg. Then I noticed that FFmpeg had widely updated that stuff. And then I learned that the backport would most probably be nope'd, so I just gave up [20:57] <JEEB> lörs laeraed [20:57] <j-b> JEEB: use VLC and the BBC demuxer [20:57] <j-b> nevcairiel: Linus does it. [20:57] <JEEB> probably a good idea [20:57] <JEEB> but at this point I don't even care [20:57] <JEEB> I will at some point finish my earthsoft dv rewrite [20:57] <wm4> j-b: actually, if you find such a person who wants to play maintainer for such a large project, you can just ask him to take over ffmpeg maintainership [20:57] <j-b> until then, more hentai playback on VLC! :D [20:57] <JEEB> (welcome to me saying this since 2013 or so) [20:58] <nevcairiel> j-b: he also has frequent outbursts at devs to balance himself [20:58] <j-b> nevcairiel: of course, but he does it correctly, with justice. [20:58] <wm4> and then we'll see if michaelni's offer is real [20:58] <j-b> wm4: as soon as I get 500k¬ per year more [20:58] <j-b> I'd hire 2. [21:03] <nevcairiel> Hrm why does vaapi need to link against something, vdpau doesn't. :( this makes distributing binaries more annoying [21:05] <wm4> because there are major vaapi api calls in the decoder [21:06] <nevcairiel> I wonder if libva is at least installed on most distros [21:06] <wm4> probably [21:06] <nevcairiel> My Debian seemed to have it on a fresh install [21:07] <wm4> even the emulation wrappers are often installed (like vdpau libva backend) [21:07] <nevcairiel> Well I'll discuss it with a colleague [21:08] <nevcairiel> I would prefer to start with vdpau myself, but we sell some hardware with an Intel GPU, so that's high on the priority list [21:09] <wm4> there's also a wrapper that emulates vdpau via libva [21:09] <wm4> and opengl [21:10] <nevcairiel> Do such wrappers work? [21:10] <nevcairiel> And work well? [21:10] <wm4> you'll have to see for yourself [21:11] <wm4> they probably incur quite some overhead [21:12] <ubitux> <+kierank> ubitux was saying all the issues could be solved on the ML and not face-to-face // no, i'm defending the point that if they can't communicate with ML, face-to-face will not solve anything [21:13] <kierank> ubitux: it's the other way round altogther [21:13] <kierank> that's my point [21:13] <ubitux> they're hiding behind this "they are autistic and can't talk face to face so we won't talk virtually" [21:14] <ubitux> and we are saying if they can't talk virtually then there is no point in trying to communicate with a worth medium (at least that's my opinion) [21:14] <kierank> no they just can ignore your emails [21:14] <kierank> whereas face to face they can;t [21:14] <ubitux> if they want to ignore the mails then they don't want to communicate [21:14] <ubitux> then there is no way i'm going to force them [21:14] <ubitux> or play their game irl [21:24] <wm4> IRL you just can turn it into a fist-fight!!!1 [21:52] <cone-840> ffmpeg.git 03Gabriel Dume 07master:0a024268261d: libxvid: K&R formatting cosmetics [21:52] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:9c41b59423bb: Merge commit '0a024268261d05ccdcf7e03c85fb78d22037a464' [22:08] <cone-840> ffmpeg.git 03Vittorio Giovara 07master:e87f5e4e5f2e: h264: fully check cropping amount from sps [22:08] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:570397c73115: Merge commit 'e87f5e4e5f2e2e36b0b7826d708cda7049877af0' [22:22] <jamrial> michaelni: did Thomas Volkert ever send the hevc rtp patch to the ml? [22:22] <jamrial> can't find it, and wanted to point him to http://fate.ffmpeg.org/log.cgi?time=20140826140938&log=compile&slot=x86_64-… [22:35] <michaelni> jamrial, dunno, there was a pull request on github [23:09] <llogan> i always forget about those [23:11] <wm4> so there are things that are reviewed (by some) and pushed, which never make it through the ML? [23:12] <llogan> https://github.com/FFmpeg/FFmpeg/pulls [00:00] --- Wed Aug 27 2014

1 0

ffmpeg.log.20140826
by burek 27 Aug '14

27 Aug '14

[00:01] <benlieb> tnx! [00:01] <benlieb> can anyone recommend a way to do what I need: I need to go through 2000+ videos and choose an accurate thumbnail frame down to the .xx decimal place. [00:01] <benlieb> Ffplay works fine for this [00:01] <benlieb> but it's tedious to manually copy paste this value from the terminal [00:02] <benlieb> is there a way once the video is paused to somehow write to a file with the filename and location? [00:08] <llogan> benlieb: you mean you want to make a screenshot wherever you pause in ffplay? [00:08] <llogan> doesn't vlc allow you to do just that? [00:09] <benlieb> I don't want a graphic, I want text, but it needs to be done from a program where I can pause with accuracy to the ss.mm level [00:10] <benlieb> in an ideal world I would have a file with lines like: filename 33.22 [00:10] <benlieb> http://stackoverflow.com/questions/25495112/how-can-i-write-to-file-with-vi… [00:10] <benlieb> llogan: ^ [00:19] <llogan> stackoverflow is for programming questions only. super user is probably the better place [01:15] <woof-woof> Hi! [01:31] <Camusensei> llogan: I managed to make the ass filter work using the information from the thread you linked me, thanks :) now I'm trying to get the subtitles larger ^^ [01:32] <Camusensei> it works \o/ [01:34] <Baked_Cake> what would happen if i used aac_he but set it to 128k instead of 64k [01:34] <Baked_Cake> would that b better that aac_lc [01:35] <Baked_Cake> i dont really need vbr at such low biit rates [01:36] <llogan> Camusensei: how did you deal with the issue? [01:37] <llogan> Camusensei: as for font size: http://stackoverflow.com/a/21369850/1109017 [01:38] <Camusensei> llogan: I already solved the font size issue [01:38] <Camusensei> llogan: I followed camelotmsl's post (http://ffmpeg.zeranoe.com/forum/viewtopic.php?f=10&t=318&start=20) as well as setting the FONTCONFIG_FILE variable, and it worked. I didn't bother trying out which one was the relevant one. [01:39] <llogan> i see. how about the font size? [01:40] <Baked_Cake> for subs? [01:40] <llogan> yes [01:41] <llogan> Baked_Cake: if you want 128k you might as well use AAC-LC [01:41] <Baked_Cake> u can edit sub fonts in aegisub [01:43] <Baked_Cake> llogan is there some drawback to using aac_he aside from limited playback? [01:45] <Camusensei> llogan: default font size suited me well. I had previously set my video size (1080p) into the ass file, which resulted in very tiny subs. removing these values got me the default font size back [01:46] <llogan> Baked_Cake: it's designed for lower bitrates. if you want to use it outside of that usual application then you're using the wrong tool. [01:46] <Camusensei> llogan: (I had set PlayResX: 1920 and PlayResY: 1040) [01:46] <Baked_Cake> ic [01:47] <Baked_Cake> how does the audio quality compare to aac_lc at 128k? it seems about the same to my ears [01:48] <llogan> if it sounds the same to you then use whatever you want to use. some people claim to like the way aac-he sounds. [01:50] <llogan> Baked_Cake: some info here https://trac.ffmpeg.org/wiki/Encode/AAC [01:50] <Baked_Cake> ya ive read thru that badboy [01:51] <llogan> also https://trac.ffmpeg.org/wiki/Encode/HighQualityAudio but i don't know if any of that is substantiated [01:51] <Baked_Cake> it does say on there that i can go upto 160k with it [01:52] <Baked_Cake> o ill check out that other link [02:02] <Baked_Cake> hmm its says the aac_he profiles cant reach transparency [02:04] <Baked_Cake> i guess ill toss that idea out the window [02:05] <Baked_Cake> aand stick to lc_aac [02:07] <maujhsn> Is there an ffmpeg command that will bring up a video image in vlc media player? [02:09] <Baked_Cake> hmm but the other page that says version 1 can go upto 160k has been updated more recently [02:09] <am0rphis> hi117, how to disable window with equalizer graphics in x when i listenning ffplay? [02:10] <Baked_Cake> ive never used ffplay [02:17] <am0rphis> i like to listening radio in background with it [02:18] <maujhsn> Is there an ffmpeg command that will bring up a video image in vlc media player? [02:18] <maujhsn> Is there an ffmpeg command that will bring up a webcam video image in vlc media player? [02:18] <Baked_Cake> looks like i need to do some reading on SBR [02:22] <Baked_Cake> hmm the idea is pretty cool, but i can see why its not considered transparent [02:22] <Baked_Cake> i guess ill just have to listen for my self some more [02:28] <Baked_Cake> o this is interesting [02:28] <Baked_Cake> Again, this is not a limitation of the standard, it's an encoder design decision. I thought I already made this clear in another thread once, but I'm happy to repeat: there's a mode called "downsampled SBR", in which you can move the SBR start frequency above half the input signal bandwidth. For example, with 44.1- or 48-kHz audio input, you could let SBR code only the frequencies above 16 kHz or so. Such a setting will be trans [02:28] <Baked_Cake> parent for many people (assuming the core bit-rate is high enough). Fraunhofer's encoder supports downsampled SBR, but that mode might not (yet) be available in Winamp, I don't remember. [02:29] <am0rphis> -nodisp hide the window ._. [05:11] <aho> i'm trying to join some images and it's dropping 2 frames [05:11] <aho> why is it doing that? [05:18] <aho> upgraded to the most recent build. works now. [05:26] <aho> or not. drops half the frames if i try to save it as webm [05:31] <aho> http://superuser.com/questions/452542/ffmpeg-drops-frames-when-encoding-a-p… [05:31] <aho> that works [06:52] <djdduty> what am I doing wrong here? http://puu.sh/b8dpu/3a19691647.png [07:09] <tbarletz_> Is there a way to dump the PMT content of a TS file using ffprobe? [07:10] <tbarletz_> dump=show the programs and pids in a human readable form [07:10] <aho> djdduty, try %2d (or 3 if there are up to 3 digits) [07:11] <djdduty> aho: they are up to 3 digits, but there are no precursing 0s on the first ones [07:11] <blazer420> Greetings. I'm having trouble piping rtmpdump to ffmpeg. Currently testing with unprotected rtmp source to make it easier, code is here.. http://pastebin.com/a9J5Me9G [07:12] <blazer420> Any assistance is appreciate. Thank you. :) [07:12] <aho> you'd have to use %02d if there are leading zeros [07:12] <aho> do you get the same error with %3d? [07:12] <djdduty> aho: yeah [07:14] <djdduty> aho: got it, there was a space in the front [07:14] <djdduty> a leading space* [07:14] <aho> ah [07:14] <aho> because i just tried %d and it worked just fine :> [07:18] <blazer420> Oh im sorry. Also, the error message I get is "pipe:: Invalid data found when processing input" [08:11] <ArM23> hi [08:14] <ArM23> hi guys how do i record RTMP live stream [09:16] <Baked_Cake> what do u guys think about libopus 5.1 vs aac 5.1 [09:16] <Baked_Cake> or just libopus vs libfdk_aac [09:25] <K4T> hi [09:26] <K4T> can someone provide me 30fps film sample, but no timeelapse or music video? [09:26] <K4T> I searched over one hundred videos on vimeo and found nothing [09:35] <Baked_Cake> looks like opus-tools is the only way to go [09:37] <K4T> ok, found one [09:43] <K4T> can someone give me some hints or link to informations about converting 29fps video to 25? [09:43] <K4T> I hope I can convert it with ffmpeg [09:44] <K4T> and get smooth playback [10:35] <K4T> can I use FFSERVER as an alternative to Wowza Media Server? I need to stream video card output (Declink Mini Monitor) to multicast, flash players on the web (like JWPlayer), android devices (via RTSP I think) and iOS devices [10:58] <K4T> no? :p [12:04] <alinescoo> Hi all. I have a question: is it possible to use android port of ffmpeg and add a watermark image over video ? [12:17] <spaam> using the overlay filter? should work [12:20] <alinescoo> Thanks @spaam, do you know how reliable is ffmpeg for android ? [12:21] <alinescoo> I need to learn a lot on this video matter and I don't want to invest time in something which has compatibility issues [12:22] <spaam> alinescoo: i have no idea. i know it builds :) [12:44] <K4T> is it possible to use ffserver on Windows? [13:03] <devnode> goooood ...day. [13:03] <devnode> i missed the morning :D [13:19] <Eftekhari> hi [13:21] <Eftekhari> i have ffserver, when feed is larger than 2 GB, if ffmpeg that feed the ffserver exit, re executing the ffmpeg to feed the server not working, with the error : Error reading write index from feed file '/tmp/feed1.ffm': Resource temporarily unavailable [13:27] <bigzed> Hello, I have a 16channel audio file. How do I convert this to stereo? [13:42] <bigzed> Or more precise, how can I drop everything except the first two channels and mix them to stereo? [13:54] <Mavrik> bigzed, you'll have to use audio filters in sequence [13:55] <Mavrik> bigzed, see channelsplit filter (for splitting) and amix filter (for remixing back) [13:59] <bigzed> Mavrik, Is there maybe an howto about it? [13:59] <Eftekhari> i have ffserver, when feed is larger than 2 GB, if ffmpeg that feed the ffserver exit, re executing the ffmpeg to feed the server not working, with the error : Error reading write index from feed file '/tmp/feed1.ffm': Resource temporarily unavailable [13:59] <Mavrik> bigzed, look on ffmpeg wiki and example on ffmpeg documentation [14:00] <Mavrik> I bet there's a remix example somewhere there [14:15] <blight> hi guys [14:17] <blight> i am thinking about opening a file twice (two format context's) and use one to read the audio and one to read the video so i can treat both as independent streams which would make my code more simple i think... does anybody know how much overhead this will have or if there is any reason why it could be a bad idea? [14:17] <blight> *contexts [14:18] <blight> (i always open files, no streams or anything) [14:37] <Hello71> sounds like you might have issues with A/V sync [14:40] <Mavrik> both audio and video frames have absolute timestamps so that's not an issue really. [15:02] <blight> the timestamps are not always 100% correct are they? and when handling the interleaved packets the code is more complex isnt it? [15:03] <blight> iirc i have a file where the audio stream has some strange timebase and the timestamps have an accumulated rounding error [15:05] <Mavrik> ugh [15:05] <Mavrik> the timestamps HAVE to be correct [15:05] <Mavrik> how else does the presenter know when it's time to present a frame or set of audio samples? [15:07] <blight> i am sure you know a lot more about ffmpeg internals, but do audio timestamps have to be sample correct? [15:07] <Mavrik> PTS has to be correct and granular enough to allow playback. [15:08] <Mavrik> note that PTS might not be in the same timebase for each stream [15:08] <blight> like if you decode a stream from some position you get some audio samples, then you seek before that position and you get samples with timestamp, will the delta between the timestamps be the exact sample count so the 2 pieces of samples can be matched with sample accuracy? [15:10] <Mavrik> depends on format [15:10] <Mavrik> but it should be for most sane formats. [15:10] <blight> hmm i am thinking about decoding an audio stream in reverse... by decoding blocks from back to front and then putting them together [15:10] <Mavrik> note that packet and data loss does exist in real world :) [15:11] <blight> well i always work with files but some of them are not perfect of course :( [16:29] <Sashmo_> here is a good question..... if my source is 59.94...... and I am encoding it to you tube live.... and they say to use a GOP of 2 seconds.... what do I set? should be -g 120 or should is be like -g 119 ? [16:39] <BtbN> flash can't play 59.94 or 60 fps anyway, so you better encode it as 30 fps [16:51] <chchjesus> Hey, I was wondering if anyone has experience with transcoding to webm from mkv? [17:01] <c_14> chchjesus: What's your question? [17:03] <chchjesus> http://pastebin.com/bnmQ9qXQ [17:03] <chchjesus> Here's a paste [17:04] <chchjesus> I was wondering what an appropriate way would be to transcode an mkv file to a webm file [17:04] <chchjesus> I'd like to make a clip out of part of the mkv file [17:04] <c_14> vpx does not have a preset setting [17:04] <chchjesus> but the resulting clip I'd like to be in webm [17:04] <chchjesus> ok [17:05] <c_14> -threads 0 is also a noop for vpx [17:06] <chchjesus> c_14: Yet theok [17:06] <chchjesus> oops [17:06] <chchjesus> It encodes it to 7 minutes instead of the 17 seconds I asked it for [17:06] <chchjesus> And the size of the file doesn't start counting up until ffmpeg has at least gotten to 12 seconds [17:09] <chchjesus> When I play the file [17:09] <chchjesus> It plays 17 seconds [17:10] <chchjesus> But it's the wrong 17 seconds, and the file says it's 7 minutes long [17:10] <c_14> Which 17 second range is it? [17:10] <c_14> about [17:11] <chchjesus> Instead it starts at what is 00:16:21 in the original mkv file [17:13] <chchjesus> Let me give you another paste [17:13] <c_14> try setpts=897 instead of the PTS+897/TB, if that doesn't work, do 897/TB [17:15] <chchjesus> What is pts? [17:15] <c_14> Presentation Time Stamp [17:15] <chchjesus> I've used it previously only to get the subtitles to the right position [17:15] <c_14> I'm guessing it has something to do with the setpts filters anyway, you can try the command without the -vf just to see if it cuts correctly. [17:16] <c_14> I know what you're using it for, I just don't know what else could be causing the difference. [17:17] <chchjesus> I suspect it's a difference in the formats? [17:17] <chchjesus> And I'm just not transcoding it correctly? [17:18] <c_14> Shouldn't be. I've done similar things rather often without issues. [17:19] <c_14> You can try using -ss and -t as output options to see if that helps. [17:19] <c_14> Maybe add -loglevel debug and see what ffmpeg has to say about the commandline. [17:20] <chchjesus> http://pastebin.com/gnUgbGSg [17:20] <chchjesus> Ok [17:20] <chchjesus> I'll try those [17:21] <chchjesus> Where do I put ss and t for output options? [17:21] <chchjesus> Do I just need to add them after the input file? [17:21] <c_14> Just on the other side of the input file. [17:23] <chchjesus> Oh crazy [17:24] <chchjesus> It's given me a 17 second file now [17:24] <chchjesus> But it's still the wrong 17 seconds [17:24] <chchjesus> But [17:24] <chchjesus> Lemme give you the output [17:24] <chchjesus> Hold up [17:25] <chchjesus> http://pastebin.com/RpVzXmjr [17:25] <chchjesus> There you go [17:25] <chchjesus> c_14 [17:30] <misitawright> hi guys i need some help. I am evaluating how to go about making a bunch of videos using ffmpeg [17:30] <c_14> I have no idea why it would be cutting incorrectly, does this happen with every file you try or just the one. [17:31] <misitawright> I would like to be able to have a text overlay that would read options: blah and have it fade in and out each option while leaving the text option in place. how would i go about doing something like this? [17:31] <chchjesus> c_14: It's happened with others I think [17:32] <chchjesus> c_14: So, the only problem now is that it's offset wrong [17:32] <c_14> Is it a static offset every time? (ie the same offset) [17:33] <chchjesus> What do you mean? [17:33] <chchjesus> The same time? [17:33] <c_14> The same start/end points. ie instead of from 14:57 it's from 16:30 every time. [17:33] <chchjesus> Oh, yesd [17:34] <chchjesus> It does mention in the output about a delay [17:34] <c_14> Ok, my cheap answer would be to find that offset and just apply it to the seek time. The slightly more intelligent answer would be to test with a recent git build (you can try a static build) and if it still doesn't work with that, check the bug tracker and if you can't find a similar bug, make a new one. [17:35] <chchjesus> Yeah, I was thinking that too (about finding the right offset) [17:35] <chchjesus> But sure [17:36] <chchjesus> I'll build from the AUR soon if this doesn't work [17:36] <c_14> (if you don't like building stuff) [17:36] <chchjesus> Ah [17:37] <chchjesus> c_14: Oh, btw, there's a long pause at the start of running it where the time nor filesize go up [17:37] <chchjesus> They both remain at 0 [17:37] <c_14> That's (probably) while it's seeking. [17:40] <chchjesus> It might just be easier to install ffmpeg-git from the user repos [17:43] <chchjesus> c_14: Weird. It just replaced both libx264 and x264 with the x264-git pacakge [17:43] <chchjesus> maybe there was something wrong with the x264 codec [17:43] <chchjesus> It's compiling it now, anyway. [17:54] <Vish__> Hi - I am getting a mjpeg stream from android device which is connected via USB through firing a terminal command. Android itself runs fmpeg and provides the mjpeg data. How do i get that data to ffserver on my mac? Any ideas? [17:56] <chchjesus> Vish__: Can you stream it to an ffm file? [17:58] <Vish__> chchjesus: thats what I am not able to do.. when I fire a terminal command eg: adb shell "ffmpeg -options" , it starts printing out put in console which is mjpeg format and accurate one. I know that. How do I pipe that output to ffserver? [18:00] <chchjesus> I'll give you a command I use [18:00] <chchjesus> First I get ffserver running [18:00] <chchjesus> with [18:00] <Vish__> chchjesus: part of problem is - ffmpeg runs on android while ffserver runs on mac [18:00] <chchjesus> It shouldn't matter [18:00] <chchjesus> ffserver -f /etc/ffserver.conf [18:00] <chchjesus> sudo ffmpeg -i $1 -vf "subtitles=$1, scale=1280:720" -b:v 256K http://localhost:8090/feed1.ffm [18:01] <chchjesus> ffmpeg -i input_file http://localhost:8090/feed1.ffm [18:01] <chchjesus> So, you want to have ffserver running on your mac, on 0.0.0.0 [18:01] <chchjesus> Then you use ffmpeg to stream to the ip address of the mac [18:02] <chchjesus> You'll also need to properly set the permissions of the feed in ffserver config [18:03] <Vish__> sorry but did not understand. You fired 2 ffmpeg on android itself.. on andorid there is no ffserver. so ffm does not exist there.. [18:07] <chchjesus> Yes, you have the ffm file running on the mac [18:07] <chchjesus> Then you stream from the android with ffmeg to the mac [18:08] <chchjesus> You want to send stuff to ffserver from ffmpeg, correct? [18:08] <chchjesus> ffmpeg being on the android? [18:13] <Vish__> yes correct.. but I guess unless ffmpeg and ffserver are not running on same machin localhost thing is not going to work... [18:13] <chchjesus> Yes, change localhost to 0.0.0.0 [18:14] <chchjesus> Are you transmitting on the same network? [18:14] <chchjesus> Or across the internet? [18:15] <chchjesus> Eh [18:30] <chchjesus> i tried to help them [19:39] <davis> hello [19:39] <davis> i pulled the ffmpeg sources using git and did a make install, but ffplay did not install. it appears its not built. Where do I get the ffplay sources? [19:40] <JEEB> then you didn't have the x11 stuff [19:40] <JEEB> check the output that configure gave [19:40] <JEEB> you can re-run configure if you want to [19:41] <davis> ok let me see what it says [19:41] <davis> hmm ./configure | less does not show any entries for ffplay [19:43] <klaxa> pastebin config.log maybe? [19:44] <JEEB> davis, it should tell you if it did find or do anything related to x11 [19:45] <davis> is there a pastebin site which allows you to upload a file? [19:45] <davis> the log is long and will be difficult with copy and paste [19:45] <davis> i do see x11 lines here, nothing with an error pops out [19:47] <davis> maybe a better question is, on the www.ffmpeg.org website, is there a place which lists all the libraries/packages necessary to build ffmpeg ? [19:48] <brontosaurusrex> slightly offtopic: i did a simple cmx 3600 plotter, but i can't make adobe to import it properly, any ideas? seems valid and all [19:48] <brontosaurusrex> i mean it wont import edl [19:51] <brontosaurusrex> example: http://paste.debian.net/plain/117692 [19:56] <llogan> davis: ffplay requires sdl [19:56] <davis> yes, i have found a ffmpeg-user post which seems to imply that. [19:57] <davis> i'm trying to get ubuntu to install it. sadly it looks like I've got broken dependencies. [19:57] <llogan> http://trac.ffmpeg.org/wiki/CompilationGuide/Ubuntu [20:04] <davis> llogan: many thanks [20:05] <davis> JEEB: many thanks [20:06] <llogan> were you successful? [20:06] <davis> i pulled down the libsdl-dev [20:07] <davis> i rerand configure and now I am dong a new build [20:07] <llogan> i wonder why someone added --dsiable-opencl to the x264 instructions [20:16] <davis> yes, it now works. many thanks again. [20:43] <drawesome> Is there a command line switch to set ffplay to be fullscreen when it opens? [20:44] <llogan> you can press 'f' while playing, but i guess that's not what you want exactly [20:45] <llogan> oh, -fs [21:02] <mistawright> hi guys i need some help. I have a png i am trying to overlay on top of a video i am making from images and an audio file. When i add the overlay the overlay shows at an offset. the image is 640x360 and the video is 640x480. how can i scale the image input and have it overlay directly on top? [21:04] <Fjorgynn> aha [21:04] <VeL0X> Hey guys [21:06] <VeL0X> as i tried to run my new music files, i got some errors. i was messing around, but at least i'm now at ffmpeg, as i try to run it i got the error "Output file #0 does not contain any stream" [21:06] <VeL0X> what to do? [21:11] <Hello71> how the ass do you run a music file [21:11] <VeL0X> with my player (moc / mocp) :D [21:12] <llogan> and can you provide the output image so i can see what you are describing? [21:12] <VeL0X> but thats showing an error like "Could not find codec parameters (err -543....) [21:13] <VeL0X> yep. what was the exact pastebin command? [21:18] <llogan> ok, next! [21:19] <mistawright> fflogger, This is the command i am using. http://pastebin.com/Mg6a6N8G [21:23] <llogan> mistawright: you forgot the complete console output. [21:28] <mistawright> llogan, http://pastebin.com/kX944PBs [21:28] <mistawright> here goes the console output [21:28] <mistawright> https://www.youtube.com/watch?v=-RBLvgQMlKU&feature=youtu.be [21:29] <mistawright> you can see my issue here [21:30] <mistawright> the bar should be at the very bottom of the screen. the width is fine. but the image is 120 pixels shy of 480. I have tried scaling but am just barely beginning to use ffmpeg [21:31] <llogan> can you provide frame.png? [21:32] <mistawright> https://docs.google.com/file/d/0B4rh1bWqHxtLUEFlYUEycDE5WkU/edit?pli=1 [21:33] <mistawright> I havent been able to figure out positioning. I will be using that with two different overlays based on a situation and will also need to have text draw on top [21:37] <llogan> mistawright: overlay=W-w:H-h [21:39] <mistawright> with the command i used how should i have gone about getting this image in place properly? [21:41] <llogan> mistawright: ffmpeg -framerate 1/5 -i image%03d.jpg -i audio.mp3 -i frame.png -filter_complex "[0:v][2:v]overlay=(W-w)/2:H-h,format=yuv420p[vid]" -map "[vid]" -map 1:a -c:v libx264 -c:a aac -strict experimental -r 30 outAudio.mp4 [21:42] <llogan> if you're just uploding to youtube then you can stream copy the audio instead of re-encoding it (-c:a copy) [21:43] <mistawright> thats good to know definitely helpful [21:44] <llogan> in your case W-w looks the same as (W-w)/2, but (W-w)/2 centers it in case the frame.png is smaller (if you don't scale it first) [21:44] <llogan> alternatively you could use drawbox filter to make the red and black bars instead of an image file [21:45] <llogan> also, add -crf 18 if you're uploading to youtube. since they re-encode you want to give it as high quality that is practical for you to upload. [21:45] <mistawright> that sounds like it may be easier. I have another transparent png that I needed overlayed as well. I have not seen an example of using multiple overlays. Once I have those overlays situated I will then go through and start adding text and transitions [21:46] <llogan> adding -shortest might be useful too to make sure the output ends when the shortest input ends [21:47] <llogan> multiple overlays aren't hard: [0:v][1:v]overlay[a];[0:v][a]overlay[b] [21:48] <llogan> drawtext also has a box drawing option, but i think it only is as tall and wide as the text (I may be wrong though) [21:48] <mistawright> the input for overlay a should that be written as -i inputa.png -filter_complex "blah" or should it be part of the existing filter complex options? [21:49] <llogan> all filtering can occur within one filtergraph (use one -filter_complex). you can chain filters in a row with a comma to make filterchains, and you can chain filterchains together with a semi colon [21:51] <llogan> as in this shitty example: [0:v]scale=640:-2,crop=iw/2,negate[scn];[0:v][scn]overlay [21:51] <llogan> you'll get the hang of it after a while [21:52] <mistawright> trying to figure out how to add the overlay to the command you provided. now [21:52] <llogan> you mean another image that you want to overlay? [21:53] <mistawright> yeah. trying to recreate what we currently have in an opensource solution minus windows servers [21:55] <llogan> something like: ffmpeg -framerate 1/5 -i image%03d.jpg -i audio.mp3 -i frame.png -i anotherimage.png -filter_complex "[0:v][2:v]overlay=(W-w)/2:H-h[v1];[v1][3:v]overlay,format=yuv420p[out]" -map "[out]" -map 1:a -c:v libx264 -c:a aac -strict experimental -r 30 outAudio.mp4 [21:58] <mistawright> I am lost where the v1 one label came from. [22:04] <llogan> mistawright: it's an arbitrary label name. you can name it anything. it's so the output from that filterchain can be referenced by other filters or via -map [22:08] <mistawright> makes sense. thanks for the help. i do appreciate it. trying to find concise documentation or examples has eluded me [22:09] <llogan> http://ffmpeg.org/ffmpeg-filters.html#Description [22:09] <llogan> https://trac.ffmpeg.org/wiki/FilteringGuide [22:15] <K4T> is ffserver still developed? [22:16] <llogan> K4T: it still receives rare updates. [22:17] <llogan> but i would not call it "active development" [22:19] <mistawright> llogan, not finding much on transitions between each image in my slideshow. I had read earlier today that kenburns and blend were available transitions [22:19] <mistawright> or at least allowing each image to fade into the next would be helpful, then kenburns would just be icing [22:19] <K4T> so it is better to not use ffserver for professional solutions, yes? [22:20] <llogan> i don't know. i've never used it. [22:20] <llogan> mistawright: i don't know the easiest way to place a blend or fade between sequential images [22:23] <llogan> but i do have a drawbox example: [0:v]drawbox=x=iw-w:y=ih-h:w=iw:h=48:c=black@0.9:t=h,drawbox=x=iw-w:y=ih-48-h:w=iw:h=5:c=red:t=h [22:25] <llogan> height could also be a percentage of the input size if you wanted it to be dynamic [22:25] <mistawright> interesting [22:25] <mistawright> time to open up pastebin so i dont forgot anything [22:26] <llogan> http://lists.ffmpeg.org/pipermail/ffmpeg-devel-irc/2014-August/thread.html [22:26] <llogan> channel is logged [22:28] <llogan> mistawright: using srt or ass subtitles to create hardsubs is a good alternative to drawtext if you want certain text to show up at certain times [22:29] <mistawright> I was going to do that for a list of options but wanted it to appear in a box about 100px by 20px [22:30] <mistawright> and figured an srt with the options would do fine. that way I could have be displayed easily [22:30] <llogan> aegisub is a good tool to make subs if you don't want to do it manually [22:32] Action: llogan will return in an hour or so [22:42] <vlatkozelka> hi [22:44] <vlatkozelka> when i stream from file to udp or any other protocol ( tho udp is what i need ) ... it gets encoded way too fast , like as fast as the cpu can read from the file [22:57] <c_14> Get a slower CPU? :P [22:57] <c_14> If you want real-time, use -re [22:57] <vlatkozelka> and [22:57] <vlatkozelka> i lose timestamps [22:58] <c_14> real-time as in stream 1s of video every second [22:58] <vlatkozelka> its a ts stream saved onto ts files [22:58] <vlatkozelka> i tried -copyts it ruined everything , started recording 1 second files [22:58] <vlatkozelka> i tried usu_wallclock .. also bad [22:58] <vlatkozelka> ill try this -re [22:59] <vlatkozelka> ok that solved the speed issue ... and dropped cpu usage :) thx [23:18] <KjetilK> I'm organizing a sports event where we will attempt to give spectators (we expect 250-300 in total) the possibility to view a video stream on their own devices using a Wifi network we set up using 7-8 access points. [23:18] <KjetilK> We think we have the wifi part rather well covered now, but we have maximum 70 Mbits/s into the arena, and so the plan we had to use livestream or ustream seems risky, since that seems to open one stream per client [23:20] <KjetilK> so, I was thinking of putting up a box with VLC, but then I just realised ffmpeg perhaps does the job better [23:21] <KjetilK> so, I can have a proxy on the LAN serving all the clients on the inside [23:21] <KjetilK> the problem seems to be that iOS clients should have HLS, whereas Android clients should have RTSP [23:22] <KjetilK> so, my question is just if I could have ffmpeg set up both a HLS stream and an RTSP stream? [23:22] <vlatkozelka> like run 2 processes at same time ? [23:23] <c_14> KjetilK: either use two outputs (this will encode everything twice) or see if you can get everything working with the tee muxer [23:23] <vlatkozelka> c_14 [23:23] <vlatkozelka> that -re thing worked [23:23] <vlatkozelka> but the timestamps always start from 00:00:00 [23:24] <vlatkozelka> idc about the timestamp in the files i need the real time stamp , like say windows clock [23:25] <c_14> use the setpts filter [23:25] <KjetilK> ok, so this should be doable, if I have one incoming stream, dunno which protocol, and then (re-)encode to get two streams out? [23:25] <c_14> yep [23:25] <KjetilK> cool :-) [23:26] <KjetilK> I may come back to haunt you with the details, meanwhile I'll report back to the rest of the pack that my evil plans are likely to work out :-) [23:26] <vlatkozelka> lol [23:26] <vlatkozelka> good luck :) [23:27] <KjetilK> thanks! :-) [23:27] <llogan> https://trac.ffmpeg.org/wiki/Creating%20multiple%20outputs#Teepseudo-muxer [23:27] <vlatkozelka> c_14 how to use it ? i know i should look at docs but ive been reading docs about stuff for about 2 months man ... this is the last bit [23:27] <KjetilK> llogan, ah, great, thanks! [23:28] <vlatkozelka> i made a monitor program in JAVA to monitor TV channels and record them and extract from them ... just need this last bit of code :v [23:29] <c_14> setpts=time(0) [23:29] <c_14> I think. [00:00] --- Wed Aug 27 2014

1 0