Ffmpeg-devel-irc

Download

ffmpeg-devel-irc@ffmpeg.org

August 2015

1 participants
62 discussions

ffmpeg-devel.log.20150731
by burek 01 Aug '15

01 Aug '15

[00:01:55 CEST] <cone-228> ffmpeg 03James Almer 07master:9dcaae70f2f3: x86/aacpsdsp: add SSE and SSE3 optimized functions [01:46:39 CEST] <cone-228> ffmpeg 03James Almer 07master:a0092cea462f: mips/hevcdsp: fix string concatenation on macros [02:07:48 CEST] <cone-228> ffmpeg 03James Almer 07release/2.7:a3d45dbcc144: mips/hevcdsp: fix string concatenation on macros [02:47:15 CEST] <cone-228> ffmpeg 03James Almer 07n2.5.8:HEAD: mips/hevcdsp: fix string concatenation on macros [03:28:31 CEST] <cone-228> ffmpeg 03Ganesh Ajjanagadde 07master:c1bfb99ff244: avformat/wavdec: add extra sample count check for G.729 files [12:04:17 CEST] <cone-782> ffmpeg 03Anton Khirnov 07master:a07979505271: fate/mp3: specify the number of output samples instead of filesize [12:04:18 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07master:814d4c4c94c7: Merge commit 'a0797950527120c85263c910eb6ba08fddcfdcb3' [16:00:53 CEST] <ubitux> michaelni: aw. [16:04:20 CEST] <Rathann> well he did a tremendous amount of work and I'm impressed he lasted so long with all that happened [16:04:46 CEST] <Rathann> michaelni: you deserve some kind of medal :) [16:08:25 CEST] <ubitux> well, maybe we need to see what libav developers want to do [16:12:58 CEST] <BtbN> I think the idea of a code review platform should realy be re-considdered now, so there's an easier way to manage patches. [16:13:26 CEST] <ubitux> i doubt that's the main issue [16:13:39 CEST] <ubitux> the main threat is obviously still the same, the toxicity between devs/projects [16:13:52 CEST] <ubitux> looks like a good opportunity to solve that [16:14:38 CEST] <nevcairiel> Good luck with that, people don't like their nice cozy bubbles getting popped [16:15:16 CEST] <ubitux> anyone willing to organize a session or something? [16:25:31 CEST] <j-b> wow [16:25:42 CEST] <j-b> Is it time for me to do an evil takeover? [16:25:46 CEST] <j-b> or is it too early [16:25:48 CEST] <j-b> :D [16:32:04 CEST] <ubitux> j-b: some help to avoid any "social" issues will be probably welcome [16:32:12 CEST] <ubitux> maybe the "problem" can be solved [16:34:47 CEST] <iive> that's the reason to rise this subject now? [16:56:33 CEST] <Compn> j-b : you should have kept quiet until we picked vlc as host [16:57:43 CEST] <j-b> Compn: haha :) [16:57:55 CEST] <Compn> ubitux : you want to merge back with libav ? [16:58:03 CEST] <j-b> ubitux: I already gave my views here and many times at that. [16:58:15 CEST] <ubitux> j-b: something just changed, right? [16:58:35 CEST] <j-b> maybe :) [16:58:38 CEST] <j-b> come at VDD :) [16:58:43 CEST] <j-b> and let's discuss about that. [16:58:49 CEST] <Compn> oh no, that discussion will take over vdd :( [16:58:54 CEST] <ubitux> Compn: well, i think it's a good opportunity to end the problem and finally stop the stress on our users [16:59:02 CEST] <Compn> and everyone whos not fffmpeg/libav will have opinion on what we should do :P [16:59:20 CEST] <j-b> Compn: the solution is simple: elect me and give me ALL your money : [16:59:29 CEST] <ubitux> i mean, we can obviously continue as is and see the project diverge and see one (or two) projects die [16:59:29 CEST] <durandal_1707> michaelni: so your commits stops right now or when? [16:59:40 CEST] <ubitux> and loose features/fixes/whatever on one or both side [16:59:55 CEST] <ubitux> or we could try to figure out a solution, because it's now only technical, isn't it? [17:00:09 CEST] <Compn> ubitux : every meeting libav said it was social issue [17:00:19 CEST] <nevcairiel> ubitux: once carl leaves too, then its only technical =p [17:00:24 CEST] <nevcairiel> the hate is strong in him [17:00:29 CEST] <ubitux> oh come on [17:00:36 CEST] <ubitux> now it's carl the problem? [17:00:39 CEST] <ubitux> please :) [17:00:50 CEST] <Compn> can we talk about this big api change coming in libav ? [17:00:56 CEST] <Compn> and how we arent going to apply that in ffmpeg ? [17:00:58 CEST] <nevcairiel> i didnt say that, its just that he will make it personal [17:01:01 CEST] <Compn> because it would break everything ? [17:01:16 CEST] <j-b> ubitux: it's not technical. [17:01:38 CEST] <wm4> <Compn> can we talk about this big api change coming in libav ? <- what "big API change", or are you FUDing again? [17:01:56 CEST] <Compn> wm4 : andreas was talking about it on the ml [17:02:08 CEST] <Compn> debian breakage [17:02:09 CEST] <nevcairiel> andreas opinion is unrealistic [17:02:20 CEST] <nevcairiel> we have to break downstreams at some point, or we can never ever introduce any new api [17:02:47 CEST] <Compn> wm4 : http://lists.ffmpeg.org/pipermail/ffmpeg-devel/2015-July/176439.html [17:02:48 CEST] <nevcairiel> and expecting us to maintain every single downstream is just naive [17:03:14 CEST] <Compn> yes nevcairiel , i saw your opinion on the list . [17:03:34 CEST] <nevcairiel> apparently its not only mine, but shared between most of those that cared to respond at all [17:03:49 CEST] <Compn> thats why i brought it up here lol [17:04:02 CEST] <Compn> because wm4 didnt see it or didnt respond to it on ffmpeg-devel [17:04:07 CEST] <nevcairiel> he saw it [17:04:15 CEST] <nevcairiel> its just not a "big api change" [17:04:31 CEST] <Compn> when you break compilation of multiple projects, i think its a big change, but whatever. [17:04:53 CEST] <Compn> doesnt the pix_fmt change just require a simple alias ? [17:05:03 CEST] <Compn> or is it all different :\ [17:06:48 CEST] <Compn> also i dont understand how creating a new api means removing the old one. but again, i'm not leader, so its not my problem :) [17:07:06 CEST] <Compn> j-b : get ready for a bunch of api changes :P [17:07:31 CEST] <durandal_1707> now if we get slashdoted site will die... [17:08:01 CEST] <Compn> durandal_1707 : one of our hosts was giving us grief because a false positive virus detection on a file :\ [17:09:04 CEST] <ubitux> durandal_1707: ppl are still reading /. ? [17:09:47 CEST] <durandal_1707> hmm, we will see... [17:20:28 CEST] <durandal_1707> So when we will vote for the new leader? What you need to candidate? [17:21:07 CEST] <wm4> lots of money and some mercenaries [17:21:56 CEST] <j-b> yep [17:22:02 CEST] <j-b> Compn: no way? [17:25:06 CEST] <Compn> durandal_1707 : anyone can be candidate of course [17:25:13 CEST] <Compn> any programmer, i'd say [17:25:30 CEST] <Compn> those wihtout experience, hopefully wont get the votes [17:54:08 CEST] <j-b> ubitux: trying to troll me on HN won't help, you know. [17:54:34 CEST] <ubitux> man, are you crazy? [17:55:02 CEST] <j-b> ubitux: or insulting me on IRC. [17:55:04 CEST] <ubitux> how do you think ffmpeg developers are you going to trust you for mediating any ffmpeg/libav meeting with all the fud you just did [17:55:24 CEST] <ubitux> i believe you are very unreasonable here [17:55:29 CEST] <ubitux> and not helping the situation [17:56:36 CEST] <j-b> ubitux: what FUD? [17:56:46 CEST] <ubitux> it's quoted [17:56:55 CEST] <ubitux> seems you don't read me [17:57:18 CEST] <j-b> I do read you. [17:57:36 CEST] <wm4> good start [17:58:43 CEST] <ubitux> i don't understand why you try now to take this opportunity to throw oil in the fire, while it's a very good opportunity to smooth things [17:58:57 CEST] <j-b> ubitux: I'm not. I'm quite objective. [17:59:05 CEST] <j-b> You are in the middle of one side and you are not. [17:59:13 CEST] <j-b> Deal with the truth: libav is cleaner. [17:59:19 CEST] <j-b> less features, but cleaner. [17:59:21 CEST] <ubitux> "the truth" [17:59:28 CEST] <ubitux> wth, what happened to you? [17:59:30 CEST] <wm4> libav isn't all that much cleaner [17:59:43 CEST] <kierank> libav h264 with > 1 thread or sliced threads is crash central [17:59:52 CEST] <kierank> a little bit of packet loss and it's segfault time [17:59:55 CEST] <kierank> fuzzing it is just sad [18:00:07 CEST] <kierank> ffmpeg's is now at least robust (perhaps ugly though) [18:00:28 CEST] <j-b> wm4: 2 prores decoders? 2 prores encoders? [18:00:43 CEST] <j-b> merging random patches that do not work like libstagefright? [18:00:51 CEST] <j-b> having 2 resampling libraries [18:01:23 CEST] <wm4> yes, but now this stuff can actually be cleaned up [18:01:30 CEST] <wm4> and most of them are fixed by rm [18:01:42 CEST] <j-b> wm4: indeed. [18:01:50 CEST] <j-b> wm4: noone said it was not cleanable [18:01:59 CEST] <j-b> wm4: but the current state is less clean [18:02:05 CEST] <j-b> and insulting me is not going to change it. [18:02:21 CEST] <wm4> you already told him you don't want the stuff he worked on [18:02:22 CEST] <bryno> it might. jk [18:02:41 CEST] <j-b> wm4: really? [18:03:04 CEST] <ubitux> j-b: you are overlooking many internal cleanups [18:03:06 CEST] <wm4> well at least you could read into it [18:03:18 CEST] <j-b> wm4: like? [18:03:50 CEST] <j-b> ubitux: sure. But elenril's work is very focused on being cleaner. [18:04:00 CEST] <atomnuker> If someone coverts the project to full C99/C11 I'll get him a case of beer and maybe a custom 'I am hero' tshirt [18:04:12 CEST] <bryno> wasn't one of those resampling libraries created after michael made the first one? :S [18:04:13 CEST] <kierank> atomnuker: they won't do it because of supporting ancient compilers [18:04:27 CEST] <atomnuker> I know, a man can dream you know [18:04:43 CEST] <kierank> atomnuker: I know - I always write for( int i = 0; out of habit [18:04:50 CEST] <kierank> and it fails to compile :( [18:04:56 CEST] <atomnuker> :( [18:05:23 CEST] <kierank> atomnuker: are you based in uk btw? [18:05:37 CEST] <atomnuker> yeah [18:10:00 CEST] <j-b> ubitux: and when even your blogpost started by attacking VLC. And you had just joined the project. [18:10:02 CEST] <bryno> j-b: ya swresample initially added 2011-09-19; avresample was 2012-04-25 :S [18:10:32 CEST] <kierank> bryno: that's a complete distortion of the truth [18:10:40 CEST] <Compn> ubitux : btw its going to get flamey for a few weeks ... i am sorry to say [18:10:42 CEST] <bryno> does git lie? [18:10:53 CEST] <Compn> ubitux : there will be a power vacuum , lots of angry people ... [18:11:01 CEST] <Compn> "utter chaos" [18:11:20 CEST] <ubitux> j-b: what? [18:11:30 CEST] <j-b> ubitux: and what about xvmc that you refused to remove [18:11:31 CEST] <ubitux> where did i say anything bad about vlc? [18:11:46 CEST] <j-b> ubitux: what about "While MPlayer community is very small, VLC seems to have enough developers to afford for example the cost of writing a better© support for some formats than libavformat from FFmpeg." [18:11:46 CEST] <kierank> bryno: swresample came out of nowhere, no mailing list, no review [18:12:18 CEST] <Compn> j-b : thats a compliment afaict [18:12:18 CEST] <ubitux> j-b: how can you interpret this as an attack to VLC? [18:12:38 CEST] <Compn> because vlc has better support than lavf for some formats [18:12:46 CEST] <Compn> probably rm and asf at least. [18:12:50 CEST] <ubitux> libavformat is known to be poorly appreciated, so i'm saying the vlc project does a better job at writing demuxer [18:12:57 CEST] <ubitux> at least for their needs [18:13:15 CEST] <ubitux> while other projects have less manpower and keep using libavformat [18:13:29 CEST] <Compn> ubitux : j-b took it as "vlc is writing their own crap and lavf is better" i guess [18:13:34 CEST] <j-b> of course [18:13:37 CEST] <j-b> like everyone else. [18:13:45 CEST] <ubitux> i'm not attacking anyone in this part, i'm describing how the projects revolve around one and another [18:13:54 CEST] <j-b> and the next sentence is: "Many other projects make use of these libraries (XBMC, Chromium, ...), you get the point." [18:14:03 CEST] <Compn> i didnt take it as insult, but i'm a native english speaker... [18:14:16 CEST] <ubitux> yes, the point is to say that the libav* libraries are used by many projects [18:14:21 CEST] <ubitux> totally unrelated to VLC [18:14:22 CEST] <j-b> ubitux: anyway, I keep my point, libav is cleaner, especially on libavcodec/libavformat. [18:14:26 CEST] <j-b> way less features [18:14:26 CEST] <ubitux> i'm mentioning API user, wtf [18:14:33 CEST] <j-b> but cleaner. [18:14:40 CEST] <Compn> ubitux : many people hate your blog post :( [18:14:43 CEST] <Compn> esp libav... [18:14:51 CEST] <j-b> of course [18:14:59 CEST] <j-b> and many people hate lu_zero blogpost [18:15:03 CEST] <ubitux> Compn: well it seems there is huge misinterpretation [18:15:05 CEST] <j-b> because both are biased. [18:15:12 CEST] <j-b> extremely biased. [18:15:24 CEST] <Compn> and factual errors... ahem [18:15:43 CEST] <ubitux> i have no idea how you came up with the idea that this section was actually bashing vlc etc [18:15:59 CEST] <ubitux> while i'm purely describing relationship between api users [18:16:12 CEST] <ubitux> well i guess i could reword, but that's the first time i hear this [18:16:20 CEST] <j-b> I already told you. [18:16:23 CEST] <j-b> Years ago. [18:16:31 CEST] <ubitux> do you have the log? [18:16:38 CEST] <j-b> probably, yes. [18:16:40 CEST] <ubitux> because i probably didn't realize that was your request [18:16:44 CEST] <ubitux> please show me [18:16:47 CEST] <Compn> i dont remember such thing, but sounds like long ago... [18:17:05 CEST] <j-b> and anyway, calling me now FUD is unjust. [18:17:22 CEST] <j-b> features vs clean is a good summary about the situation [18:17:24 CEST] <ubitux> it's not, and i maintain that it's pure fud to say that the project is a mess [18:17:37 CEST] <j-b> Your codebase is messier than libav. [18:17:44 CEST] <j-b> It's not unfixable [18:17:47 CEST] <j-b> but it's a fact. [18:17:51 CEST] <ubitux> proportionally i wouldn't say so [18:18:01 CEST] <j-b> proporitonally to what? [18:18:06 CEST] <ubitux> we have about 130k additionnal lines of code [18:18:29 CEST] <ubitux> so the amount of "trash" or whatever needs to be compared to that [18:18:40 CEST] <jamrial> and you think making libav's codebase be as robust and complete as ffmpeg's would be easier than making ffmpeg's codebase cleaner? [18:18:42 CEST] <j-b> On the common libavcodec/libavformat code, I doubt you have 130k more LoC [18:18:47 CEST] <jamrial> it would be quite a lot of work both ways [18:18:57 CEST] <j-b> jamrial: I _really_ don't know about that [18:19:01 CEST] <ubitux> j-b: ffmpeg is not composed on only this [18:19:03 CEST] <j-b> jamrial: and that was never my point [18:19:14 CEST] <j-b> jamrial: both ways are very complex. [18:19:19 CEST] <ubitux> and as i said, we made many internal cleanups [18:19:22 CEST] <j-b> ubitux: sure, but that's what people use. [18:19:38 CEST] <ubitux> j-b: people use the cli, and the cli covers the filter as well [18:19:46 CEST] <ubitux> players are not the only users of the libs [18:19:46 CEST] <j-b> sure. [18:20:00 CEST] <j-b> players have vastly more users than the cli [18:20:13 CEST] <j-b> libavcodec is used everywhere [18:20:20 CEST] <ubitux> that needs to be proven, but the scope of this function is poorly defined [18:20:23 CEST] <j-b> it's the most beautiful open source project [18:20:26 CEST] <ubitux> s/function/statement/ [18:20:32 CEST] <j-b> it's in ALL smart TVs [18:20:39 CEST] <j-b> almost all Android phones [18:20:44 CEST] <j-b> it has billions of users. [18:21:00 CEST] <j-b> and the glory of that goes to michaelni [18:21:15 CEST] <j-b> he did the fucking core parts, aka MPEG-2/MPEG-4/H264 decoders [18:21:23 CEST] <j-b> it's fucking amazing [18:21:54 CEST] <j-b> that does not change that on those core parts, libav is cleaner. [18:22:03 CEST] <ubitux> not for sure [18:22:15 CEST] <j-b> surely. Sorry. [18:22:17 CEST] <Compn> ubitux : probably wont be able to convince j-b about this , without detailed diffs or something. [18:22:21 CEST] <ubitux> maybe on a biased point of view [18:22:21 CEST] <jamrial> didn't kierank say above that h264 on libav's side is unstable? [18:22:25 CEST] <kierank> jamrial: yes [18:22:29 CEST] <j-b> jamrial: it is. [18:22:32 CEST] <nevcairiel> Knowing the features and clean-factor of both projects, personally I would favor ffmpeg for the features [18:22:35 CEST] <kierank> for > 1 thread or sliced threads libavh264 is a disaster [18:22:41 CEST] <kierank> I had many many many crashes on air with libav [18:22:45 CEST] <kierank> and then with ffmpeg too [18:22:51 CEST] <kierank> but michael fixed all the ffmpeg issues [18:22:54 CEST] <kierank> for some definition of fixed [18:22:58 CEST] <kierank> depending on your viewpoint [18:22:59 CEST] <j-b> I agree with kierank on h264 [18:23:06 CEST] <j-b> ubitux: biased? [18:23:18 CEST] <j-b> ubitux: did you not keep xvmc code? [18:23:32 CEST] <j-b> ubitux: do you not have multiple prores decoders? encoders? [18:23:42 CEST] <Compn> j-b : please go easy on our devels , its a bad time right now. i think [18:23:43 CEST] <j-b> ubitux: do you not have multiple resampling libraries in your codebase? [18:23:55 CEST] <Compn> do you not have vlc work to do ? :P [18:24:04 CEST] <j-b> Compn: sorry, but one cannot call me a liar, and me not reacting [18:24:25 CEST] <j-b> ubitux: look at your massiver MPEG12EncContext, etc... [18:24:29 CEST] <nevcairiel> j-b: actually that last point is kinda libavs fault, not ours, not merging libavresample would've made a mess on api support [18:24:44 CEST] <nevcairiel> and they re-invented the library after ffmpeg already had one [18:24:53 CEST] <nevcairiel> but agreed on the other parts [18:24:53 CEST] <j-b> nevcairiel: probably. But maybe not merge the other one then. [18:25:16 CEST] <j-b> ubitux: we use FFmpeg in VLC. We already said so. [18:25:24 CEST] <j-b> ubitux: and that's the features reason [18:25:29 CEST] <jamrial> j-b: regarding the resampling libraries, libav NIH'd the second, and ffmpeg merged it to remain compatible [18:25:37 CEST] <jamrial> the alternative was to not be compatible at all [18:25:53 CEST] <j-b> jamrial: actually it's not totally true, since libav started its before, but merged after. [18:26:10 CEST] <j-b> jamrial: but agreed, fault is shared and libav attitude on that was stupid. [18:26:32 CEST] <j-b> ubitux: Debian and Arch use FFmpeg by default, because of features. [18:26:43 CEST] <j-b> so, I agree that features are very useful [18:26:46 CEST] <jamrial> libavresample being pretty much the same as swresample, including michaelni's copyright, but with slight differences says otherwise [18:26:50 CEST] <bryno> (stability/security too) [18:26:52 CEST] <ubitux> it's not about features [18:27:00 CEST] <ubitux> i believe we are also way more stable generally speaking [18:27:11 CEST] <j-b> ubitux: I have to disagree. [18:27:17 CEST] <jamrial> swr came first out of nowhere. libav then paid someone to rewrite parts of it and slap a new name on top [18:27:18 CEST] <j-b> 2.1.0 was libav, 2.1.1 was FFmpeg [18:27:20 CEST] <jamrial> check the timeline [18:27:20 CEST] <j-b> on Windows [18:27:39 CEST] <ubitux> and again, i can look at random cleanup commits i've done over the years [18:27:45 CEST] <ubitux> and other devs as well [18:27:52 CEST] <ubitux> we just didn't advertize them [18:27:55 CEST] <j-b> ubitux: the number of crashes on libavcode_plugin.dll was a bit higher in 2.1.1 than 2.1.0 per million users. [18:28:05 CEST] <ubitux> like tauting about an internal private structure no user would give a shit about [18:28:22 CEST] <j-b> ubitux: sure, but most of elenril's commit are about cleaning. [18:28:35 CEST] <Compn> j-b : and the compile chain was all the same between those releases? thats all our fault? :P [18:28:39 CEST] <ubitux> j-b: and it's been merged [18:28:46 CEST] <ubitux> j-b: but we also did cleanups on our part [18:28:52 CEST] <j-b> ubitux: and some parts reverted [18:29:00 CEST] <ubitux> because they were wrong probably [18:29:03 CEST] <j-b> like to get low_res in a global one [18:29:11 CEST] <j-b> or to keep fucking xvmc support!!! [18:29:13 CEST] <ubitux> is this a problem for you? [18:29:27 CEST] <j-b> it's not a problem, I told you, I use FFmpeg [18:29:28 CEST] <ubitux> but like, i don't know, the fact that ass packets from mkv demuxer are not crafted with timestamps anymore [18:29:33 CEST] <j-b> and we host FFmpeg git! [18:29:35 CEST] <ubitux> don't you count this as a cleanup? [18:29:46 CEST] <j-b> yes, of course you did cleanups [18:29:47 CEST] <ubitux> (lowres is useful and requested) [18:29:52 CEST] <ubitux> and i'm not alone [18:29:52 CEST] <j-b> but libav did more. [18:29:55 CEST] <ubitux> maybe ! [18:29:58 CEST] <ubitux> but in the end [18:30:04 CEST] <ubitux> their cleanups ended up in ffmpeg [18:30:11 CEST] <ubitux> and we are talking about the current state of the code base [18:30:17 CEST] <j-b> no, they don't [18:30:19 CEST] <ubitux> not talking about who is the best [18:30:40 CEST] <ubitux> what do you mean "no, they don't"? [18:31:03 CEST] <j-b> last example: [18:31:09 CEST] <j-b> qsvdec_mpeg2 [18:31:25 CEST] <j-b> you merged a code that does not even flush the pipeline [18:31:32 CEST] <wm4> so I'm looking forward to j-b hiring a full time, paid ffmpeg maintainer, who is accepted by all the community _and_ Libav, and is in addition not a white caucasian male to foster diversity in the opensource community [18:31:53 CEST] <j-b> https://git.videolan.org/?p=ffmpeg.git;a=blob;f=libavcodec/qsvdec_mpeg2.c;h… vs https://git.libav.org/?p=libav.git;a=blob;f=libavcodec/qsvdec_mpeg2.c;hb=HE… [18:31:53 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07master:f2c58931e629: MAINTAINERS: Remove myself as leader [18:31:54 CEST] <cone-782> ffmpeg 03rogerdpack 07master:bed1d9ec7a25: dshow: show more debug timestamp info [18:31:58 CEST] <j-b> wm4: what about you? [18:32:10 CEST] <wm4> I'm already busy and everyone hates me [18:32:20 CEST] <nevcairiel> i gave up commenting on the qsv patches [18:32:21 CEST] <j-b> wm4: and don't joke, I officially went to see Youtube, Samsung and Dailymotion [18:32:29 CEST] <bryno> wm4: what's with that last part? lol [18:32:31 CEST] <j-b> to ask for money to maintain FFmpeg [18:32:32 CEST] <nevcairiel> the author of them is ignoring feedback [18:32:41 CEST] <j-b> nevcairiel: and it was merged nonetheless [18:32:52 CEST] <ubitux> j-b: so is this about "cleanup" or bugs? [18:32:55 CEST] <j-b> and it took some code that I FUCKING WROTE with lu_zero [18:33:09 CEST] <j-b> slapped the name on it and got it merged by michael [18:33:18 CEST] <nevcairiel> that guy is some corporate idiot [18:33:23 CEST] <j-b> when the patches from libav where on the mailing list for 3 weeks [18:33:35 CEST] <Compn> j-b : thanks for asking those companies to maintain ffmpeg. that was one of the ideas i had as well :) [18:33:45 CEST] <wm4> j-b: haha, was that so, I didn't notice [18:33:47 CEST] <j-b> Compn: I went to the Linux foundation [18:33:58 CEST] <wm4> (I didn't attempt to compare them) [18:34:00 CEST] <j-b> to register FFmpeg as a core infrastructure project. [18:34:03 CEST] <j-b> wm4: I did. [18:34:26 CEST] <j-b> ubitux: you can't say I prefer libav or FFmpeg. Both have different styles. [18:34:37 CEST] <j-b> ubitux: but a summary is "features vs clean" [18:34:41 CEST] <ubitux> i never said so [18:35:42 CEST] <ubitux> i'm just saying that ffmpeg being a huge mess in comparison to libav is extremelly biased and wrong depending on how you evaluate that [18:36:19 CEST] <wm4> ubitux: the main difference is not the actual code [18:36:21 CEST] <ubitux> the proportion of clean things is in my opinion larger in ffmpeg [18:36:21 CEST] <wm4> but the attitude [18:36:45 CEST] <ubitux> that is another debate, and i agree that it needs to be solved [18:36:45 CEST] <j-b> Anyway, back on cleaning: AVCodecContext [18:36:47 CEST] <wm4> basically, ffmpeg is too eager with merges, and often doesn't reject patches it should reject [18:36:59 CEST] <wm4> of course both projects contain major code sins etc. [18:37:07 CEST] <j-b> look at AVCodecContext in both projects [18:37:10 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07release/0.10:79bc4798b7bb: MAINTAINERS: Remove myself as leader [18:37:11 CEST] <wm4> and both projects have made some attempts at making things cleaner or easier [18:37:11 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07release/0.11:d3a1bf3579bc: MAINTAINERS: Remove myself as leader [18:37:12 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07release/0.5:b256bd3be348: MAINTAINERS: Remove myself as leader [18:37:14 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07release/0.6:0a849fe1f2d7: MAINTAINERS: Remove myself as leader [18:37:16 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07release/0.7:9526ad4f25f6: MAINTAINERS: Remove myself as leader [18:37:17 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07release/0.8:a064b4eb1297: MAINTAINERS: Remove myself as leader [18:37:19 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07release/0.9:36a3063aae33: MAINTAINERS: Remove myself as leader [18:37:21 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07release/1.0:e8bf189db283: MAINTAINERS: Remove myself as leader [18:37:22 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07release/1.1:7608871734c8: MAINTAINERS: Remove myself as leader [18:37:24 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07release/1.2:43919b692263: MAINTAINERS: Remove myself as leader [18:37:25 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07release/2.0:9b493887d502: MAINTAINERS: Remove myself as leader [18:37:26 CEST] <jamrial> and when some of those sins try to be cleaned, people ask for the deprecation to be postponed or cancelled [18:37:27 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07release/2.1:52fa2a2f6044: MAINTAINERS: Remove myself as leader [18:37:29 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07release/2.2:ac7d47dd7021: MAINTAINERS: Remove myself as leader [18:37:31 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07release/2.3:d88a85c49338: MAINTAINERS: Remove myself as leader [18:37:33 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07release/2.4:dfc147d24feb: MAINTAINERS: Remove myself as leader [18:37:35 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07release/2.5:af098adca39d: MAINTAINERS: Remove myself as leader [18:37:37 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07release/2.6:faed1deec9f2: MAINTAINERS: Remove myself as leader [18:37:38 CEST] <cone-782> ffmpeg 03Michael Niedermayer 07release/2.7:4afccfdc04f9: MAINTAINERS: Remove myself as leader [18:37:39 CEST] <ubitux> wm4: a huge amount of the mess can be solved in time if the toxicity between the two projects ends [18:37:48 CEST] <wm4> ubitux: I hope that too [18:37:55 CEST] <j-b> wm4: thanks. Maybe you formulate it better. [18:38:02 CEST] <wm4> at least Libav doesn't have the MiNi excuse [18:38:05 CEST] <wm4> (anymore) [18:38:10 CEST] <j-b> yep [18:38:18 CEST] <jamrial> j-b: again, the alternative was to drop compatibility. nobody likes all the hacks, duplicate defines, getter/setter functions and what not [18:38:26 CEST] <j-b> jamrial: xvmc... [18:38:28 CEST] <j-b> come on [18:38:32 CEST] <j-b> it never worked [18:38:34 CEST] <wm4> who on the ffmpeg side do they still really hate? maybe cehoyos if it comes down to that [18:38:41 CEST] <j-b> anyway, wm4, jamrial and ubitux: you are very invited VDD [18:38:46 CEST] <j-b> we can discuss that in persons [18:38:54 CEST] <j-b> We reimburse all your expenses [18:39:00 CEST] <wm4> when is vdd? [18:39:09 CEST] <j-b> http://vdd.videolan.org/ [18:39:26 CEST] <Compn> wm4 : sept 18-20 i thinks [18:39:27 CEST] <j-b> and Compn can be a good example that we don't eat anyone from FFmpeg or libav. [18:39:33 CEST] <Compn> lol i been to vdd [18:39:41 CEST] <Compn> j-b is nice [18:39:42 CEST] <j-b> Compn: please don't say we've tortured you [18:39:51 CEST] <Compn> very gracious host :) [18:40:00 CEST] <ubitux> jamrial: dropping compatibility is one thing, keeping ffmpeg as a superset is very important [18:40:03 CEST] <j-b> :D [18:40:03 CEST] <Compn> puts up with a lot of bearded nerds. [18:40:12 CEST] <ubitux> jamrial: because it makes possible a technical future reunification [18:40:34 CEST] <ubitux> so keeping 1 resampling, 1 prores, etc should be our priority [18:40:36 CEST] <Compn> michaelni : now that you are no longer leader, are you coming to vdd ? [18:40:38 CEST] <j-b> what I don't get is WHY resign NOW, when you have won? [18:40:40 CEST] <ubitux> but it must remains a superset in feature [18:41:03 CEST] <j-b> I mean, for example, libav has 0 subtitles support, 0 libavfilter (or almost) [18:41:05 CEST] <nevcairiel> i should figure out when i need to be in america so i can figure out if i have time for vdd [18:41:08 CEST] <ubitux> j-b: because so much work in such a toxic environment is not humanly viable maybe? [18:41:10 CEST] <Compn> j-b : it was never about winning. [18:41:14 CEST] <j-b> Compn: ah. [18:41:22 CEST] <Plorkyeran> huh, I could actually make it to vdd this year since I'll be in copenhagen the week before and it'd be easy enough to spend two days in france on the way home... [18:41:25 CEST] <j-b> nevcairiel: VDD is just after IBC [18:41:32 CEST] <Plorkyeran> not sure there's any point, though [18:41:36 CEST] <j-b> made exactly for this reason. [18:41:57 CEST] <ubitux> j-b: yeah i guess i will come again, even if i don't like it :) [18:42:09 CEST] <wm4> j-b: ok, that week I have something else to do... unless vdd happens somewhere close to italy? [18:42:16 CEST] <Compn> euro disney .. are there any arcades in paris ? rather go for a full street fighter tourney... any video gamers in vlc ? :P [18:42:26 CEST] <iive> j-b: what is your beef with xvmc? [18:42:27 CEST] <wm4> oh, paris it was [18:42:30 CEST] <j-b> Compn: disney is optionnal. [18:42:39 CEST] <j-b> wm4: Paris is close to Italy. [18:42:43 CEST] <j-b> iive: it never worked. [18:42:48 CEST] <iive> it still does [18:42:57 CEST] <wm4> not sure if close enough [18:43:01 CEST] <iive> well, last time i checked. [18:43:06 CEST] <wm4> in theory I'm in italy until satursday this week [18:43:09 CEST] <j-b> when was that? [18:43:17 CEST] <iive> few months ago. [18:43:18 CEST] <j-b> wm4: come. Serioulsy. [18:43:36 CEST] <j-b> iive: it never worked for me, and it's ridiculous outdated tech. [18:43:47 CEST] <nevcairiel> maybe i should try to organize my america trip to just stop in paris on my way back home [18:43:57 CEST] <j-b> wm4: I won't make you come in the same plane as lu_zero [18:44:04 CEST] <iive> and xvmc as api is still provided by fglrx and mesa-gallium drivers. no idea about nvidia [18:44:06 CEST] <j-b> wm4: and I won't make you use VLC. [18:44:07 CEST] <Compn> nevcairiel : where in america are you going ? [18:44:18 CEST] <nevcairiel> minneapolis [18:44:24 CEST] <Compn> ahh , couple states away [18:44:34 CEST] <j-b> wm4: and why don't you become the leader? [18:44:53 CEST] <Compn> he'd have to put his real name on something... [18:45:00 CEST] <philipl> man, i go away for a few days and all this happens? [18:45:14 CEST] <j-b> philipl: then, maybe this is your fault ? [18:45:23 CEST] <j-b> Plorkyeran: why not? [18:45:24 CEST] <Compn> philipl : welcome back to fiery planet of gasolene and more gas [18:45:29 CEST] <philipl> clearly [18:46:00 CEST] <philipl> i'm actually still on holiday. made the mistake of glancing at the log. [18:46:54 CEST] <wm4> philipl: all this happened just today [18:48:08 CEST] <iive> j-b: also, about swresample vs avscale . I'd like to see some proof that libav's one have been started before the ffmpeg one. I've heard this myth before, but I've never seen it confirmed. [18:48:28 CEST] <j-b> iive: you meant avresample. [18:48:44 CEST] <j-b> iive: I remember having seen the code before on github. But if you really want, I can dig. [18:48:45 CEST] <wm4> maybe now there's hope that we can hope that this libs could be merged [18:48:57 CEST] <andrewrk> michaelni, thank you for all your hard work over the past years. Here's hoping this resignation takes the pressure off you and allows the communities to repair [18:48:59 CEST] <iive> j-b: please do. [18:49:15 CEST] <j-b> iive: I will. [18:50:07 CEST] <jamrial> iive: swr showed up one day with no warning on ffmpeg's tree. then avr showed up several months later reusing most of its code on libav's tree [18:50:14 CEST] <j-b> for VDD, please come. [18:50:49 CEST] <iive> jamrial: yes, i also remember that ffmtech funded avr, while swr was already in git. [18:51:09 CEST] <durandal_1707> I will not come because I do not have enough bodyguards [18:51:23 CEST] <jamrial> https://lists.libav.org/pipermail/libav-devel/2012-April/026060.html [18:51:24 CEST] <Compn> durandal_1707 : i havent seen any fights at vdd yet [18:51:24 CEST] <j-b> durandal_1707: seriously? [18:51:40 CEST] <Compn> (everyone who fights does not get reimbursed for travel fees! ahhhhh) [18:51:46 CEST] <j-b> durandal_1707: it's very friendly and nice. [18:51:52 CEST] <Compn> otherwise fists would be flying! :P [18:51:58 CEST] <philipl> wm4: my timing is clearly impecable [18:52:04 CEST] <kierank> durandal_1707: I will bodyguard you [18:52:10 CEST] <j-b> and as Compn said, if there is any fight, no reimbursements [18:52:13 CEST] <j-b> durandal_1707: are you trolling or? [18:52:38 CEST] <wm4> jamrial: the truth is much more complicated, but let's not do this [18:52:52 CEST] <philipl> Well, i'm going back to the pool. see you all on monday. [18:52:53 CEST] <j-b> Compn: do you know we once refused to reimburse someone? [18:52:57 CEST] <Compn> j-b : so did linuxfoundation consider ffmpeg a core project? guessing nope. also corporate takeover failed ? [18:53:09 CEST] <j-b> Compn: well, they considered it. [18:53:17 CEST] <j-b> Compn: and then I had more pressing matters [18:53:20 CEST] <Compn> j-b : no, i only heard of one ragequit , but i wasnt there and dont like to spread rumors [18:53:31 CEST] <durandal_1707> nah, its to far away for me, organize it in some country more near Croatia [18:53:39 CEST] <j-b> Compn: and I thought I did not have the legitimacy. [18:53:45 CEST] <j-b> durandal_1707: Croatia is near. [18:54:03 CEST] <Compn> durandal_1707 : i'm travelling from usa to france, i'm farther than you are :P [18:54:03 CEST] <jamrial> wm4: i don't doubt it. i wasn't around back then i think. i'm just saying what i saw from emails and logs [18:54:20 CEST] <Compn> stupid non stop 7 hour flight :\ [18:54:38 CEST] <iive> Compn: trust me, you don't want it to stop :) [18:54:57 CEST] <Compn> i stopped to try flying virgin airways, it was the same as any other plane, i feel cheated :\ [18:55:05 CEST] <Compn> also virgin has terrible customer service so blah [18:55:36 CEST] <iive> why would you fly with them? because they are virgin? [18:55:49 CEST] <Compn> i thought they had more space between seats or something [18:56:04 CEST] <Compn> they keep showing pictures of people laying down in airplanes! [18:56:10 CEST] <Compn> i just want to stretch out a bit :( [18:56:33 CEST] <iive> you just need to be first class. [18:57:05 CEST] <Compn> http://www.aluxurytravelblog.com/2014/01/23/what-its-like-to-fly-upper-clas… [18:57:19 CEST] <Compn> nah first class looked the same as all other first class seats , little bigger [18:57:41 CEST] <Compn> first world problems [18:58:14 CEST] <Compn> hmm why is croatia to paris 5h ? [18:59:40 CEST] <kierank> Compn: it's 2hrs [18:59:42 CEST] <kierank> on ryanair [18:59:52 CEST] <kierank> air france too [19:00:32 CEST] <Compn> looking at bad flight websites then haha [19:01:28 CEST] <j-b> seriously, come at VDD [19:02:25 CEST] <ubitux> luca pasting post from 2011 on the hn list ;_; [19:02:28 CEST] <ubitux> it's going to be hard. [19:02:36 CEST] <ubitux> -list [19:03:08 CEST] <durandal_1707> hn list? [19:03:15 CEST] <ubitux> remove list [19:03:18 CEST] <ubitux> hn post [19:03:27 CEST] <ubitux> https://news.ycombinator.com/item?id=9981805 at the full bottom [19:03:46 CEST] <j-b> ubitux: downvote him. [19:04:01 CEST] <ubitux> whatever [19:04:17 CEST] <ubitux> it's not really the problem [19:04:30 CEST] <ubitux> i'm just saying it's going to be hard to have some aggreement [19:04:39 CEST] <ubitux> if they're still in the 2011 bubble [19:05:01 CEST] <j-b> luca is not they [19:05:24 CEST] <ubitux> (and the cleanup monkeys is something that was alreday disscussed so the post doesn't prove any point ;_;) [19:05:38 CEST] <ubitux> yeah sure, but well he's part of the group [19:05:58 CEST] <ubitux> i wish he was alone in this state of mind [19:06:05 CEST] <ubitux> but i'm not so sure about that [19:06:57 CEST] <iive> j-b: btw, how do you define clean code? [19:07:23 CEST] <iive> other than less features and more bugs ;) [19:09:25 CEST] <j-b> iive: many things, but not duplicating code is one good reason :) [19:10:52 CEST] <iive> you know that a lot of ugliness in ffmpeg comes from trying to be api/abi compatible with libav. [19:11:13 CEST] <j-b> iive: of course, I do. [19:11:25 CEST] <j-b> iive: but that does not change the result. [19:11:33 CEST] <Compn> did anyone even ask libav if they wanted to merge back ? [19:11:49 CEST] <Compn> i love peoples' 3rd-party ideas that the projects would simply merge back together [19:11:59 CEST] <Compn> but did they even ask any developers ? [19:12:00 CEST] <Compn> nope [19:12:48 CEST] <iive> don't worry. libav life support would end soon. [19:13:15 CEST] <j-b> Compn: well, what I'm sad is that there are less developers now if you add both sides, we have less than before. [19:13:49 CEST] <Compn> have to look at ohloh stats [19:13:53 CEST] <andrewrk> Compn, on the other hand, libav devs have said in the past that michaelni resigning would make them reconsider a merge [19:13:57 CEST] <Compn> "before" means ... 2010 ? [19:14:23 CEST] <iive> andrewrk: michaelni is ffmpeg. [19:14:49 CEST] <j-b> iive: so ffmpeg is dead now? [19:15:06 CEST] <iive> what? [19:15:23 CEST] <andrewrk> iive, I think you are using hyperbole to explain just how impactful michaelni is and how much work he puts in, and I agree that it is significant [19:16:05 CEST] <andrewrk> but if ffmpeg was *only* michaelni then his resignation would mean the project is dead. this is what j-b is saying [19:16:13 CEST] <Compn> look at the ops in this channel. ask how many of them have had commits within the last month... [19:16:23 CEST] <j-b> andrewrk: yes, thanks for explaining. [19:18:08 CEST] <iive> let me say it this way. If michael leaves ffmpeg today and creates his own fork, in 5 years that fork would be replacing ffmpeg. [19:18:23 CEST] <j-b> ah, ok, I see what you meant. [19:18:34 CEST] <Compn> i think that fork would replace ffmpeg immediately [19:18:37 CEST] <j-b> I understood in the other way. [19:18:49 CEST] <Compn> because a number of devels would switch to it [19:19:19 CEST] <j-b> well, the one who has the power is wm4. [19:19:22 CEST] <Compn> because i dont think there are many people who are happy with the current situation [19:19:33 CEST] <wm4> wut [19:19:35 CEST] <Compn> situation of code* [19:19:55 CEST] <Compn> if we could drop libav compat that would be nice [19:20:20 CEST] <Compn> maybe i should make a patch... [19:20:32 CEST] <durandal_1707> huh [19:20:38 CEST] <iive> give it a year or 2. [19:20:56 CEST] <Compn> iive : for the 5 projects using libav? :P [19:21:05 CEST] <Compn> er 5 projects using libav-compat code [19:21:10 CEST] <Compn> not talking about libav itself, my bad [19:21:33 CEST] <wm4> there's no libav compat code, except the shitty ABI hack, and libavresample being part of the repo [19:21:36 CEST] <Compn> (5 projects is also a guess, probably more like not many) [19:22:01 CEST] <Compn> which shitty hack? defining decoders to hex ? [19:22:17 CEST] <Compn> to avoid conflicts of decoder numbers [19:22:25 CEST] <wm4> ... [19:22:43 CEST] <Compn> anyone else want to get rid of it ? :) [19:23:55 CEST] <j-b> wm4: really. [19:24:22 CEST] <Compn> wm4 : j-b is probably picking you because you seem quite neutral on fork stuff [19:24:37 CEST] <Compn> and because you are a developer of mpv, which uses key features of ffmpeg [19:24:39 CEST] <j-b> no, because he is mpv's fork. [19:24:51 CEST] <j-b> no, because he is mpv's HEAD. [19:25:18 CEST] <Compn> you like mpv code then ? [19:25:25 CEST] <Compn> or style or something about it [19:25:37 CEST] <Compn> not knocking it just curious [19:25:46 CEST] <j-b> knowcking? [19:26:07 CEST] <Compn> as in i'm not "knocking it down" or insulting mpv [19:26:30 CEST] Action: Compn misses mplayer masters [19:26:36 CEST] <Compn> wonder if we can get arpi to take over leader [19:26:42 CEST] <Compn> or baptiste [19:26:45 CEST] <kierank> lol [19:26:46 CEST] Action: Compn waves at baptiste [19:26:51 CEST] <kierank> carl too, why not [19:26:53 CEST] <Compn> baptiste : nows the chance. [19:27:08 CEST] <Compn> kierank : well baptise has maintained ffmbc for years, as well as being ffmpeg dev for years [19:27:18 CEST] <Compn> so he has the skills, no doubt [19:27:40 CEST] <kierank> no offence to baptiste but I think he knows himself he's not going to be the unifying figure we need [19:27:44 CEST] <j-b> he does not have people's skill: he's French. [19:28:08 CEST] <j-b> and he's been a dick to the community, by forking with GPLv3 so noone can port it back. [19:28:16 CEST] <Compn> kierank : is there a true unifying person you could name ? [19:28:34 CEST] <kierank> merbanan probably [19:28:37 CEST] <j-b> +1 [19:28:44 CEST] <j-b> and kierank [19:28:44 CEST] <kierank> pengvado :) [19:28:47 CEST] <kierank> wbs [19:29:00 CEST] <j-b> of course, wbs, is the best guy in the world. [19:29:05 CEST] Action: Compn keeps saying kie-rank in his head [19:29:21 CEST] <Compn> pronouncing* [19:29:29 CEST] <Plorkyeran> I've actually seen a few people say mpv is the cleanest C codebase they've seen [19:29:45 CEST] <Plorkyeran> I'm not sure I'd agree, but it's not a crazy wrong sentiment [19:29:57 CEST] <wm4> not sure if I'd agree either [19:30:05 CEST] <j-b> come on, mpv is the best thing. [19:30:07 CEST] <Compn> yeah but double dashes? why [19:30:17 CEST] <Plorkyeran> it's certainly in great shape considering the starting point [19:30:48 CEST] <wm4> some mplayer stuff is still haunting me [19:30:55 CEST] <wm4> is ubitux completely out as leader? [19:31:07 CEST] <wm4> I think he's usually reasonable, but Libav folks hate him anyway [19:31:22 CEST] <j-b> I don't think he fits, no. [19:34:33 CEST] <j-b> wm4: what about you? [19:34:59 CEST] <wm4> no time, and I don't think I'm the right fit [19:35:06 CEST] <wm4> ffmpeg would become the most hated software project [19:35:14 CEST] <j-b> wm4: noone hates you. [19:35:17 CEST] <j-b> who hates you? [19:36:00 CEST] <Compn> j-b : why not french vlc fork of ffmpeg ? [19:36:10 CEST] <j-b> Compn: because people hate us. [19:36:11 CEST] <Compn> enough people working on it would draw developers [19:36:12 CEST] <kierank> mintelpeg [19:36:14 CEST] <Compn> who hates you ? [19:36:15 CEST] <baptiste> j-b, you are being dramatic :) [19:36:16 CEST] <kierank> SECAMPEG [19:36:32 CEST] <j-b> Compn: you know that I received death threats, I already told you. [19:36:45 CEST] <Compn> i thought that was over the santa hat thing :P [19:36:49 CEST] <j-b> yes, once [19:37:00 CEST] <JEEB> kierank: man SECAM... [19:37:23 CEST] <Compn> j-b : so no santa hats in vlc ffmpeg fork. [19:37:26 CEST] <Compn> its doable [19:37:29 CEST] <j-b> the other one was about intel/osx [19:37:31 CEST] <Compn> ah [19:37:38 CEST] <kierank> lol [19:37:39 CEST] <kierank> what? [19:37:47 CEST] <j-b> kierank: I never told you? [19:37:50 CEST] <kierank> no [19:37:55 CEST] <Compn> i dont remember also. [19:38:13 CEST] <j-b> kierank: that I went to the police because of some guy sending a white powder in snail mail to my parent's place? [19:38:21 CEST] <kierank> what.the.fuck [19:38:23 CEST] <Compn> whoa, never heard that [19:38:44 CEST] <wm4> being a public person (or almost) always exposes you to disturbed persons, I guess [19:38:51 CEST] <j-b> at the time where there was a fake Anthrax threads on the news? [19:39:17 CEST] <j-b> kierank: and once by an anonymous letter with words cut from newspaper asking me to die? [19:39:43 CEST] <iive> wtf? [19:39:52 CEST] <j-b> kierank: if I did not tell you that, I blame YOU. [19:39:57 CEST] <Compn> its a strange world out there :\ [19:40:03 CEST] <j-b> kierank: because you did not get drunk enough times with me :) [19:40:11 CEST] <durandal_170> perhaps they were not hapy with vlc [19:40:16 CEST] <j-b> durandal_170: probably :) [19:40:26 CEST] <wm4> vlc crashed in the middle of their porn session!?!!? [19:40:37 CEST] <Compn> vlc wouldnt play their spank porn... [19:40:47 CEST] <j-b> wm4: of course, the support for wmv3 is really bad in libavcodec. [19:40:52 CEST] <wm4> maybe it was because that 3D stereoscopic porn didn't play [19:41:00 CEST] <j-b> wm4: in BOTH forks, I have to say. [19:41:13 CEST] <wm4> aha, I would have guessed it was a Libav specific problem [19:41:22 CEST] <Compn> wm4 : you are right, we need all developers to have occulous rift displays so we can test this 3d stereo porn... [19:42:36 CEST] <j-b> wm4: nope. [19:42:49 CEST] <j-b> wm4: for Wmv3, FFmpeg supports more, but crashes more. [19:42:58 CEST] <j-b> wm4: (in VLC crash reports) [19:43:04 CEST] <wm4> business as usual [19:43:13 CEST] <durandal_170> samples? [19:43:28 CEST] <j-b> durandal_170: nope. crash stacktraces. [19:44:16 CEST] <Compn> durandal_170 : we are supposed to fix bugs without samples to verify if we fixed them or not . [19:44:29 CEST] <j-b> durandal_170: else, it would have been fixed, no? [19:44:51 CEST] <wm4> so the next practical issue that will pop up is that Libav merges have been stopped [19:45:07 CEST] <Compn> thats insignificant [19:45:21 CEST] <durandal_170> I can cherry pick them [19:45:28 CEST] <Compn> significant is that , from the email, it looks like half of the project will stop being maintained. [19:45:48 CEST] <j-b> then make it one project again. [19:45:49 CEST] <wm4> there are already 10 unmerged commits [19:46:20 CEST] <Compn> unless i read michaelni's email incorrectly. michaelni , are you still maintaining code or taking a break-break ? [19:48:41 CEST] <durandal_170> IIRC there is script for merging [19:48:53 CEST] <wm4> Compn: telling from his email, he's gone for now [19:49:02 CEST] <wm4> he says he might return, but not as leader [19:49:24 CEST] <ubitux> why do we need a leader? [19:49:37 CEST] <wm4> to make decisions, and to apply patches [19:49:44 CEST] <durandal_170> who will merge libav? [19:49:45 CEST] <wm4> without breaking the tree for everyone [19:49:45 CEST] <ubitux> i mean, michael was a "leader" because of meritocracy [19:49:57 CEST] <jamrial> anyone can apply patches, that's not the point of a leader [19:50:02 CEST] <wm4> even democracies have leaders [19:50:03 CEST] <ubitux> i don't think anyone in ffmpeg really stands as huge contributor [19:50:28 CEST] <wm4> jamrial: I'm pretty sure MiNi did a full fate run for every patch and merge [19:50:28 CEST] <ubitux> a leader is very useful to simplify the process of making decision [19:50:34 CEST] <gnaggnoyil> if there were no leader who could do code review then? [19:50:44 CEST] <ubitux> but for now, i think the main issue is solving the ffmpeg/libav issue [19:50:52 CEST] <ubitux> having a leader is purely future logistic [19:50:55 CEST] <Compn> ubitux : you werent around when we had democratic process and endless votes... [19:50:57 CEST] <Compn> i dont think [19:51:00 CEST] <wm4> ok [19:51:01 CEST] <jamrial> wm4: as should have whoever sent the patches to review in question [19:51:02 CEST] <ubitux> look, libav survived a long time without leader [19:51:02 CEST] <Compn> endless votes are ENDLESS. [19:51:05 CEST] <wm4> we still need merge slaves [19:51:22 CEST] <wm4> ubitux: didn't you complain about Libav's dev model? [19:51:39 CEST] <jamrial> but i agree with ubitux, we could try having a flat structure/hierarchy where nobody is has the last word but majority wins [19:51:40 CEST] <ubitux> i did, and i believe it's not viable, but i think it's not the priority right now [19:51:54 CEST] <j-b> libav have de-facto leaders [19:51:59 CEST] <j-b> all open source projects have [19:52:06 CEST] <j-b> even if the leadership is splitted [19:52:21 CEST] <ubitux> right now, in my opinion, the priority is to focus on making sure no divergence that will prevent a merge of the 2 projects [19:52:29 CEST] <ubitux> no divergence happens* [19:52:35 CEST] <ubitux> and start pourparler... [19:52:36 CEST] <j-b> Once again, I believe you should split the libraries, and you should have submaintainers [19:52:44 CEST] <wm4> does that mean strictly keeping Libav's API (as usual)? [19:52:44 CEST] <durandal_170> I can merge as cherry-pick only anybody against? [19:52:45 CEST] <j-b> like the linux kernel [19:52:46 CEST] <jamrial> what's libav's stance on all this? [19:52:56 CEST] <ubitux> jamrial: that's the question i'd like to answer [19:53:05 CEST] <wm4> j-b: splitting the libs is huge work, though [19:53:13 CEST] <durandal_170> they are not interested [19:53:17 CEST] <j-b> wm4: sure. And maybe not. [19:53:19 CEST] <ubitux> i think we should summarize what every dev want [19:53:39 CEST] <j-b> wm4: keep libavcodec/avf in the same, for now, because there is no other way. [19:53:52 CEST] <j-b> wm4: but the others can. [19:54:10 CEST] <Compn> durandal_170 : nothing against, but it was more than just merge, it was also reviewing before merge... [19:54:15 CEST] <nevcairiel> splitting on the dependency should be possible indeed [19:54:19 CEST] <wm4> ok, but splitting them is still work that _someone_ has to do [19:54:21 CEST] <ubitux> basically, i think libav wants "clean code", "not michael", and "forced review", right? on ffmpeg side, i believe most of the devs don't want the last 4 years of development to come to a waste [19:54:31 CEST] <ubitux> and on user side, i think no one wants to deal with the 2 forks anymore [19:54:34 CEST] <wm4> I certainly wouldn't reject a patch to do so [19:54:37 CEST] <Compn> durandal_170 : and then fate testing each merge..... [19:54:38 CEST] <ubitux> did i forget anything? [19:54:57 CEST] <wm4> ubitux: I think that's right [19:55:08 CEST] <ubitux> so, if we happen to merge both project [19:55:09 CEST] <j-b> who would be able to work on FFmpeg full time, if paid? [19:55:15 CEST] <j-b> s/able/willing [19:55:24 CEST] <ubitux> j-b: i am, but how much can you pay? [19:55:29 CEST] <Compn> j-b : depends who boss is and what their demands are [19:55:34 CEST] <Compn> of course :P [19:55:37 CEST] <j-b> ubitux: that's not your problem, yet. [19:55:44 CEST] <j-b> ubitux: how much money do you claim? [19:56:01 CEST] <ubitux> about my current salary [19:56:03 CEST] <jamrial> ubitux: even if libav's code base is used as "starting point" for the merge of both projects, ffmpeg's last four years of work could be slowly cherry picked [19:56:04 CEST] <ubitux> :/ [19:56:13 CEST] <j-b> what is your current salary + 2% ? [19:56:22 CEST] <jamrial> probably a lot of work, but eh, the git history will look nicer [19:56:28 CEST] <ubitux> jamrial: yes and i think it's fair to assume that the main claim ffmpeg devs will have is to use ffmpeg as the code base [19:56:37 CEST] <ubitux> jamrial: and the main discussion will be around code methods [19:56:49 CEST] <ubitux> jamrial: and eventually the project name (i guess a third name is pretty fine typically_) [19:56:57 CEST] <j-b> ubitux: what you need is clear rules of work [19:57:04 CEST] <wm4> would this work: we start with Libav, and then everyone posts his favorite feature as a patch? (I don't really believe in this) [19:57:09 CEST] <jamrial> ffmpeg-ng :p [19:57:14 CEST] <ubitux> j-b: yeah, this can be discussed, and we can come to an agreement [19:57:14 CEST] <j-b> CoC, extremely clear rules [19:57:24 CEST] <j-b> ubitux: who else? [19:57:27 CEST] <nevcairiel> wm4: there is way too much, and the libav review and merge process is way too slow, it would stall everything for 2 years [19:57:32 CEST] <kierank> nevcairiel: yeah [19:57:38 CEST] <j-b> not necesarrily [19:58:03 CEST] <wm4> nevcairiel: it'd imply new maintenance, not the current Libav regime [19:58:10 CEST] <durandal_170> ubitux: you are assuming too much; merge is unrealistic [19:58:11 CEST] <j-b> and everyone would need to do a step [19:58:12 CEST] <iive> ffmpeg development policy have proved to work better? why replace it with libav one that have proved otherwise? [19:58:17 CEST] <ubitux> durandal_170: why? [19:58:18 CEST] <j-b> merge is possible [19:58:20 CEST] <nevcairiel> wm4: still quite a shitload of patches [19:58:24 CEST] <j-b> with a lot of money [19:58:26 CEST] <wm4> nevcairiel: definitely [19:58:47 CEST] <j-b> wm4: would you be able to work on it fulltime, if paid? [19:58:55 CEST] <wm4> is there anything in ffmpeg that Libav would strictly reject? other than libswresample and duplicated codecs [19:59:25 CEST] <nevcairiel> wm4: feature-wise probably not, but stuff would have to be re-written massively i imagine [19:59:27 CEST] <durandal_170> Carl caf muxer [19:59:31 CEST] <j-b> we can ask. [19:59:32 CEST] <wm4> j-b: well, currently I have a job, so I'd have to quit that; doesn't seem likely (unless they throw me out soon) [19:59:34 CEST] <jamrial> wm4: fringe codecs/demuxers [19:59:48 CEST] <j-b> jamrial: I don't think they object to this? [19:59:54 CEST] <kierank> they don't [19:59:58 CEST] <kierank> they accept kostya patches [19:59:58 CEST] <wm4> they deleted some game demuxers after the fork or so [20:00:00 CEST] <kierank> for weird game things [20:00:03 CEST] <nevcairiel> well they have removed fringe codecs before [20:00:11 CEST] <j-b> DonDiego is gone, remember? [20:00:13 CEST] <Compn> wm4 : old asf demuxer [20:00:14 CEST] <nevcairiel> presumably because it was easier than to fix the code [20:00:15 CEST] <j-b> mru is gone too, remember? [20:00:23 CEST] <wm4> Compn: not really a huge issue [20:00:26 CEST] <Compn> what/ [20:00:31 CEST] <j-b> and both project need to go to a middle ground. [20:00:31 CEST] <ubitux> do you guys think we can start a sane online discussion with libav? [20:00:40 CEST] <j-b> online and sane? you are hopeful. [20:00:44 CEST] <wm4> Compn: ffmpeg devs weren't really against it, but thought it was premature [20:00:45 CEST] <jamrial> i don't know if we can, but i sure want to [20:00:48 CEST] <Compn> wm4 : new demuxer breaks all non wmv codecs [20:00:59 CEST] <Compn> and this includes g2mshit [20:01:04 CEST] <wm4> Compn: don't make this situation more complicated than it is [20:01:06 CEST] <Compn> which kostya spent years decompiling [20:01:09 CEST] <Compn> you asked. [20:01:18 CEST] <nevcairiel> personally i find the idea of working on ffmpeg/libav fulltime interesting, but as everyone else, I do have a job, and quitting is always a risky move [20:01:42 CEST] <nevcairiel> Compn: the new asf demuxer works fine for me after recent fixes [20:01:57 CEST] <Compn> nevcairiel : on the samples i linked to in that thread ? [20:02:16 CEST] <nevcairiel> not sure i re-tried all of them, but all i tried anyway [20:02:17 CEST] <Compn> i was reviewing changes to the demuxer but didnt see anything relevant [20:02:23 CEST] <Compn> ok , no problem then [20:02:42 CEST] <nevcairiel> the old one sucks at seeking massively [20:02:45 CEST] <nevcairiel> the new one works much better [20:02:57 CEST] <nevcairiel> which was a rather common complaint from my users :( [20:04:11 CEST] <Compn> thanks for testing [20:05:17 CEST] <wm4> I suppose elenril wouldn't have the time for being leader [20:05:33 CEST] <wm4> and he's, let's say, not interested in politics [20:06:27 CEST] <ubitux> a leader is not needed if the decision process is well defined and without dead lock [20:06:36 CEST] <ubitux> at least not at the beginning [20:07:13 CEST] <ubitux> the priority really is to make sure we don't lose anything in the future [20:07:33 CEST] <ubitux> i mean, if we stop merging libav, at least one project will lose something [20:07:35 CEST] <wm4> also, in case you consider someone from the Libav-only side, keep in mind that the leader could be for the glorious time when we attempt a merge (such as say by cherry-picking from ffmpeg into a new tree) [20:07:50 CEST] <Compn> j-b : i remember iive testing xvmc in mplayer a while back, when someone asked about removing it. mplayer supports a lot of old hardware :) [20:08:01 CEST] <ubitux> cherry-picking from ffmpeg is not viable at all [20:08:16 CEST] <iive> ffmpeg already have libav merged... [20:08:16 CEST] <ubitux> the new code base for a common project needs to be the ffmpeg [20:08:34 CEST] <j-b> Compn: yeah, but I don't believe it is relevant in thpse days [20:08:40 CEST] <ubitux> while many can agree to make trade off on the methods and project decisions and whatever [20:08:43 CEST] <Compn> fair enough. [20:08:53 CEST] <ubitux> for the code part, it can hardly be anything else than current ffmpeg code base [20:08:59 CEST] <jamrial> iive: the git history is a mess, that's the main complain really [20:09:00 CEST] <Compn> and i pretty much agree. [20:09:15 CEST] <iive> jamrial: that's how git works. [20:09:40 CEST] <wm4> iive: not really, the merge feature was completely abused [20:10:03 CEST] <Compn> ubitux : there are so many people using libavfilter now, that could never be removed. [20:10:23 CEST] <wm4> libavfilter isn't ffmpeg-only... [20:10:27 CEST] <kierank> Would people here be interested in making a statement [20:10:39 CEST] <wm4> that we want love and peace? sure [20:10:44 CEST] <iive> let me say it this way... history matters only if you want to blame somebody [20:10:44 CEST] <ubitux> Compn: i don't want the projects to diverge because it kills a reunification and it's a HUGE mistake for users. [20:10:45 CEST] <baptiste> kierank, yeah [20:10:57 CEST] <j-b> kierank: yes, you should [20:10:59 CEST] <Compn> statement from baptiste ? :) [20:11:00 CEST] <kierank> wm4: just one that thanks michael for his work, perhaps tells the world about some interim process [20:11:06 CEST] <j-b> kierank: use an etherpad [20:11:09 CEST] <kierank> and offers some form of olive branch [20:11:34 CEST] <wm4> kierank: sounds good [20:11:39 CEST] <j-b> and say that you are lookong for the future but not need a bit of time [20:11:44 CEST] <kierank> https://etherpad.mozilla.org/Cem9NXUdZw [20:11:54 CEST] <Compn> we could try working on a neutral 3rd party mailing list [20:12:02 CEST] <Compn> but i offered up that idea in 2011 and it didnt work then either [20:12:08 CEST] Action: Compn stops talking about 2011 now [20:12:13 CEST] <j-b> Compn: for leader [20:12:35 CEST] <j-b> kierank: and make Lydia read it before publication [20:12:40 CEST] <wm4> empty paper can be scary [20:12:55 CEST] <j-b> I got a master in BS [20:13:19 CEST] <Compn> wm4 : type 4 sentences and someone will come in with 2 pages of nonsense [20:13:21 CEST] <Compn> :P [20:13:42 CEST] <Compn> thresh : are russians going to take over and be leader of ffmpeg? [20:13:50 CEST] <Compn> thresh : we've ruled out the french already i think [20:14:06 CEST] <kierank> can you stick your name in please [20:14:21 CEST] <Compn> theres one ukrainian i want to ask about coming back [20:16:44 CEST] <Compn> although he fancies himself as a swede [20:20:02 CEST] <wm4> kierank: lol [20:20:25 CEST] <wm4> I think what will happen is that ffmpeg goes its way, probably with slightly more chaos [20:20:47 CEST] <wm4> and I _really_ hope that in some months we will have a proper solution, including getting Libav into the boat [20:21:44 CEST] <BtbN> Does it even have to be a single leader? [20:21:52 CEST] <wm4> no [20:22:06 CEST] <wm4> someone (j-b?) said each sub-lib should have its maintainer [20:22:16 CEST] <Compn> it would be nice if someone would talk to libav before issuing a statement [20:22:22 CEST] <kierank> we need an interim management [20:22:27 CEST] <kierank> for the next 6 weeks or so [20:22:29 CEST] <kierank> until vdd basically [20:22:53 CEST] <Compn> you mean michael-lite with the same policies ? [20:23:01 CEST] <Compn> in effect now, for 6 weeks ? [20:23:05 CEST] <kierank> roughly yes [20:23:15 CEST] <Compn> cant hurt. [20:23:23 CEST] <iive> i think we should ask michael to come back. [20:23:35 CEST] <Compn> iive : i think ffmpeg situation is killing him. [20:23:39 CEST] <Compn> literally. [20:23:44 CEST] <wm4> iive: come on, you can't want that... [20:23:51 CEST] <wm4> what Compn said [20:23:52 CEST] <iive> he needs a rest [20:23:52 CEST] <Compn> but i dont know him so i cant say ... [20:24:03 CEST] <iive> and he needs to stop libav merges. [20:24:27 CEST] <wm4> kierank: should we claim that this will be sorted out on VDD? [20:24:40 CEST] <kierank> I think we should [20:24:46 CEST] <Compn> i can assure you we wont be able to sort anything out in 2 days at vdd [20:24:57 CEST] <kierank> because otherwise we air all of this dirty laundry on ML on HN on reddit etc [20:25:45 CEST] <Compn> or there will be something bad like 3 ffmpeg devs at vdd (not enough for a consensus) [20:25:46 CEST] <kierank> who has access to ffmpeg security ml? [20:26:01 CEST] <Compn> admins , root@ffmpeg [20:26:03 CEST] <wm4> there's such a ML? [20:26:15 CEST] <Compn> we keep trouble like wm4 out of that list :P [20:26:22 CEST] <Compn> (i've also never used it , so cant help you there) [20:26:46 CEST] <Compn> root = reimar , beastd, uhhmmm [20:26:58 CEST] <Compn> we got some new roots too so i'm not sure. [20:27:18 CEST] <Compn> the people michaelni thanked when the server moved might be a good list [20:27:29 CEST] <cone-782> ffmpeg 03Paul B Mahol 07master:5870b3d2a3af: MAINTAINERS: update my maintership [20:28:20 CEST] <wm4> goddamn, calling Libav a fork is so political [20:29:13 CEST] <kierank> wm4: not my view [20:29:19 CEST] <kierank> but it will be political for others [20:29:24 CEST] <kierank> you know that as well as I do [20:29:30 CEST] <wm4> yes [20:30:03 CEST] <kierank> bear in mind there are people who left (mike) who aren't in either [20:30:21 CEST] <kierank> and others [20:32:38 CEST] <wm4> how about a cliched "this is not the end, but a new beginning" line [20:32:44 CEST] <kierank> LOL [20:34:13 CEST] <Compn> wm4 : mips stuff was rejected in libav iirc (if you still wanted to know stuff that was rejected in one and not other) [20:34:30 CEST] <kierank> big question: ask carl to sign or not? [20:34:38 CEST] <Compn> kierank : cant hurt. [20:34:48 CEST] <Compn> i dont know why people are scared of carl [20:35:10 CEST] <iive> he was vilified. [20:35:34 CEST] <nevcairiel> he did a lot of that himself with his aggressive comments against libav [20:35:53 CEST] <iive> he is actually quite nice and reasonable, if you sit and chat with him. [20:36:25 CEST] <wm4> he's a slight bit too persistent when it comes to issues he greatly cares about [20:36:30 CEST] <wm4> see: pkg-config [20:37:42 CEST] <wm4> is there a need to clarify that VideoLAN != vlc? [20:37:52 CEST] <wm4> because then the internet says vlc has taken over ffmpeg development [20:37:55 CEST] <kierank> ubitux: are you happy to run ffmpeg.org [20:38:14 CEST] <ubitux> you mean the server? [20:38:28 CEST] <kierank> yes [20:39:07 CEST] <Compn> wm4 : i wouldnt mention vlc in the statement [20:39:14 CEST] <Compn> because they arent doing anything yet :P [20:39:24 CEST] <wm4> "they" [20:39:36 CEST] <Compn> ? [20:40:25 CEST] <kierank> so, big questions [20:40:30 CEST] <kierank> are we doing the libav merge and who will do it? [20:40:42 CEST] <Compn> kierank : we still have to move servers [20:40:49 CEST] <kierank> Compn: do we actually need to? [20:41:01 CEST] <wm4> well, maybe you should ask Libav whether they'd accept such a merge [20:41:13 CEST] <kierank> wm4: huh? [20:41:16 CEST] <Compn> yes, ubitux's host is threatening to pull the plug because of a false positive virus alert on a rtmpdump build for windows [20:41:20 CEST] <nevcairiel> kierank: you mean continuing the merges we had so far? [20:41:24 CEST] <kierank> nevcairiel: yes [20:41:31 CEST] <wm4> kierank: oh, yeah, I got that wrong [20:41:32 CEST] <nevcairiel> i could volunteer to see about continuing those [20:41:33 CEST] <BtbN> Hello libav, we decided to merge you back, so, wellcome back guys. [20:41:41 CEST] <Compn> kierank : anyways, root is moving servers, so i'm just letting you know ... [20:41:43 CEST] <ubitux> BtbN: that will work for sure [20:41:51 CEST] <kierank> Compn: this is ridiculous [20:42:04 CEST] <Compn> what part ? [20:42:11 CEST] <Compn> theres a lot of shitty web hosts ... [20:42:14 CEST] <ubitux> kierank: yeah i guess i'm fine continuing to pay for the server [20:42:18 CEST] <nevcairiel> does anyone know if there is a script to do these merges? [20:42:21 CEST] <Compn> ubitux's temp host is one of them. [20:42:22 CEST] <nevcairiel> manual labor sounds excessive [20:42:23 CEST] <ubitux> kierank: i mean it's not really an issue for me [20:42:24 CEST] <kierank> ubitux: I mean we could use SPI money [20:42:31 CEST] <ubitux> kierank: lol it's 50¬/month [20:42:35 CEST] <ubitux> i don't care [20:42:37 CEST] <kierank> ok [20:42:42 CEST] <ubitux> i mean it's fine for now [20:42:44 CEST] <wm4> nevcairiel: yeah, that must be some awful work [20:42:52 CEST] <wm4> especially elenril's patches [20:42:52 CEST] <ubitux> i'm fine with not paying but it's not urgent [20:42:55 CEST] <kierank> but seriously, just going to some random bulgarian place is not a good idea [20:43:05 CEST] <Compn> kierank : there are two offers, we are taking both [20:43:08 CEST] <wm4> I can completely understand why MiNi got tired of this [20:43:15 CEST] <Compn> ones in sweden i think... [20:43:16 CEST] <wm4> and I'm impressed that he did this for 4 years or so [20:43:22 CEST] <Compn> the other is bulgaria [20:43:29 CEST] <kierank> Compn: based off who agreeing? [20:43:42 CEST] <kierank> the list was deadly quiet because they were two random hosts [20:43:49 CEST] <kierank> dreamhack isn't too bad [20:43:55 CEST] <kierank> but that bulgaria place is here today gone tomorrow [20:44:04 CEST] <Compn> seriously, this is not up for debate. i'm not going to say stop talking about it , but you are not root and root has made the choice. feel free to fork :P [20:44:13 CEST] <kierank> root resigned and then made the choice [20:44:17 CEST] <kierank> ergo has no authority to do so [20:44:19 CEST] <Compn> root is like 8 people [20:44:28 CEST] <Compn> you dont understand what root is even. [20:44:35 CEST] <Compn> root(a)ffmpeg.org if you want to email everyone [20:44:49 CEST] <kierank> ok whatever, still a joke [20:44:53 CEST] <Compn> ??? [20:44:57 CEST] <Compn> bulgaria you mean [20:44:58 CEST] <Compn> ok [20:45:01 CEST] <kierank> no [20:45:10 CEST] <kierank> the whole idea of going begging for servers [20:45:12 CEST] <kierank> when vlc has one right there [20:45:22 CEST] <wm4> heh [20:45:28 CEST] <iive> kierank: now that michael is gone from ffmpeg, you can assume direct control and move ffmpeg wherever you want. [20:45:31 CEST] <Compn> vlc is not neutral place , we may move there in the future... [20:45:45 CEST] <wm4> videolan at least claims to be neutral, and they host other projects [20:45:56 CEST] <wm4> projects which are used not only by vlc (not by a long shot) [20:46:11 CEST] <Compn> i have no objections to moving to vlc [20:46:16 CEST] <Compn> vlc had objections on us actually. [20:46:23 CEST] <Compn> because we arent secure enough :P [20:46:23 CEST] <wm4> really? [20:46:31 CEST] <kierank> j-b: ^ [20:46:32 CEST] <Compn> from what i've gleaned, yes [20:46:33 CEST] <kierank> wtf [20:46:34 CEST] <wm4> well, hosting technical details? [20:46:45 CEST] <Compn> i think j-b offered anyway [20:47:02 CEST] <Compn> like i said, root has some stuff to do (liek virtualization) and then we can finish moving to wherever [20:47:28 CEST] <jamrial> i don't remember them objecting. they offered even, but michael wasn't sure because of infrastructure reasons or something [20:48:52 CEST] <Compn> where were you guys for the last two weeks when we discussed this all publically on the mailing lists ? [20:48:57 CEST] <wm4> sounds like technical details to me [20:49:01 CEST] <Compn> including mplayer and ffmpeg mailing lists cc'd [20:49:06 CEST] <wm4> not a rejection [20:49:19 CEST] <Compn> i said objection not rejection. [20:49:25 CEST] <wm4> k [20:49:28 CEST] Action: Compn going to take a break [20:50:37 CEST] <jamrial> i remember BBB was very vocal in showing his support for videolan hosting ffmpeg [20:51:41 CEST] <BBB> they already do our git [20:51:47 CEST] <BBB> they are properly foundationed and funded [20:51:51 CEST] <BBB> and we are very friendly with them [20:51:54 CEST] <BBB> I see only +++ [20:52:02 CEST] <wm4> yep [20:52:19 CEST] <wm4> they provide a lot of stability and are very friendly with us [20:53:03 CEST] <wm4> I wouldn't have argued that much in favor of videolan when ffmpeg still had a maintainer who considers his project his baby [20:56:02 CEST] <BBB> is there a signed version pf pavgw? [20:56:06 CEST] <BBB> s/pf/of/ [20:56:19 CEST] <BBB> I cant find it, and it looks like pavgw doesnt work correctly when the input is signed [20:57:08 CEST] <peloverde> BBB: I don't think so, I wanted to use it for the iwht [20:57:17 CEST] <BBB> :( [20:57:44 CEST] <BBB> ohwell [21:00:39 CEST] <BBB> hi btw peloverde :) hows stuff [21:00:45 CEST] <BBB> </late> [21:03:00 CEST] <wm4> anyway, we really need to come to a decision how to handle merges from Libav, before it's too late and too much work has piled up [21:03:21 CEST] <wm4> I see some possibilities: [21:03:33 CEST] <wm4> 1. have a merge slave (someone would have to volunteer... not me) [21:04:15 CEST] <wm4> 2. let merge slaves volunteer on a by-case basis, so you can volunteer to merge specific commits (probably too slow and not reliable enough) [21:04:40 CEST] <wm4> 3. switch to cherry-picking, and let anyone who is interested send patches to ffmpeg-devel [21:05:15 CEST] <wm4> "merge slave" doesn't sound nice, but I bet it _isn't_ nice [21:05:16 CEST] <BBB> didnt michaelni merge daily? [21:05:22 CEST] <wm4> yes [21:07:41 CEST] <BBB> oh I see [21:07:50 CEST] <BBB> I only check the ML once a day or so [21:08:00 CEST] <BBB> hm& that sucks [21:08:50 CEST] <BBB> so uhm, isnt the main splitting point of the fork done then? as in, rather than picking a merge slave, lets consider a merge [21:09:08 CEST] <BBB> this whole libav vs ffmpeg thing is incredibly unhealthy for everyone involved [21:09:18 CEST] <Gramner> BBB: what if you add (1<<15) to both registers and then use pavgw? [21:09:50 CEST] <BBB> Gramner: well Im trying to decrease instructions for /= 2 [21:09:59 CEST] <peloverde> I agree with BBB, but I'm not very active in libav anymore [21:10:02 CEST] <Compn> its not that michaelni merged daily, its that he reviewed and fate tested everything when merging [21:10:04 CEST] <wm4> BBB: yes, but a merge will require long discussions etc. if it's even possible [21:10:20 CEST] <BBB> right now I have pcmpgtw zero, value; psubw value, zero; psraw value, 1 [21:10:26 CEST] <wm4> BBB: we're also hoping that VDD will contribute to reaching this goal [21:10:30 CEST] <Compn> so its kind of like 3 jobs all at once , daily. [21:10:34 CEST] <wm4> but it's some time until then [21:10:41 CEST] <BBB> Gramner: I was trying to do pcmpgtw reg, value, zero; pavgwsigned value, reg [21:11:00 CEST] <BBB> Gramner: and that obviously doesnt work, and adding stuff sorta defeats the point [21:11:06 CEST] Action: Compn gets off puter now. good luck, try not to screw everything up while i'm gone. [21:11:39 CEST] <BBB> is anyone still on their irc channel? [21:11:48 CEST] <peloverde> I did unpack, add, shift, pack. But it helped I was only using half the vector [21:14:16 CEST] <wm4> BBB: you mean libav? several of us [21:14:22 CEST] <iive> BBB: what is exactly that you want to do? [21:14:25 CEST] <wm4> but they aren't very excited [21:14:28 CEST] <BBB> oh :( [21:14:34 CEST] <BBB> ok [21:15:24 CEST] <BBB> btw, michaelni: :( - but I guess I understand a little& well miss you [21:15:35 CEST] <Gramner> i see. well, 3 instructions doesn't seem that bad. i mean one extra subtraction isn't very significant performance-wise [21:16:47 CEST] <BBB> Gramner: yeah, not too bad, was just trying to be nitpicky for a second& not a big deal I guess [21:19:06 CEST] <BBB> I feel like we need a linux-foundation-type-of-org funding some benevolent dictator, linus-style [21:19:14 CEST] <BBB> doing it for free is not going to work for anyone for very long [21:24:17 CEST] <wm4> BBB: j-b keeps talking about paying someone for full-time work on ffmpeg [21:24:29 CEST] <BBB> great! [21:24:33 CEST] <BBB> who is it? :) [21:24:39 CEST] <wm4> we don't know [21:24:44 CEST] <wm4> that's the hard part [21:24:48 CEST] <ubitux> anyone knows how to handle the fate samples? [21:24:52 CEST] <wm4> BBB: hey aren't you free? [21:25:08 CEST] <BBB> ubitux: uploading? Ive uploaded once [21:25:13 CEST] <BBB> I might still remember how that works [21:25:17 CEST] <BBB> its in my email somewhere [21:25:20 CEST] <BBB> that may be pre-fork though [21:25:53 CEST] <ubitux> yeah it's actually more about how it's handled on our cron [21:26:00 CEST] <BBB> wm4: uhm& not exactly, Im running my own little startup company [21:26:14 CEST] <wm4> see, everyone has his job [21:26:44 CEST] <wm4> so even if we had that money (actually we have it, going by j-b), who will we give it [21:26:49 CEST] <BBB> its how people pay rent and groceries [21:27:14 CEST] <BBB> nobody is waiting around for michael to resign so jb can start paying him to me michael_v2 [21:27:28 CEST] <BBB> because it dosnt happen all that often [21:29:19 CEST] <durandal_1707> I can work half time, rest I need for sleep [21:30:31 CEST] <J_Darnley> You could pay me to be Michael v0.1 [21:35:02 CEST] <durandal_1707> J_Darnley: how is asm going? [21:35:40 CEST] <J_Darnley> ha ha ha [21:35:45 CEST] <J_Darnley> not at all [21:47:09 CEST] <durandal_1707> J_Darnley: reason? [21:55:57 CEST] <j-b> wm4: sorry, but no. we can't replace michael with one person. [21:58:08 CEST] <thresh> also videolan can happily host ffmpeg if needed. [21:58:17 CEST] <j-b> yes [21:58:19 CEST] <j-b> and libav [21:58:22 CEST] <j-b> and fate [21:58:28 CEST] <j-b> and whatever crap you neeed. [21:58:36 CEST] <thresh> yes, we have enough power and bw and disks and whatnot. [21:58:42 CEST] <j-b> yeah. [22:02:51 CEST] <j-b> kierank: good etherpad [22:03:19 CEST] <iive> yeh, you need at least 3 people to replace michaelni. one to admin the servers, one to merge stuff and one to fix bugs... [22:04:04 CEST] <wm4> we now have a "nobody knows what this code does" situation (though Libav has been effectively working with that assumption before) [22:04:07 CEST] <iive> but given how much work he did, 10 people is more realistic estimate. [22:04:07 CEST] <thresh> I can assure that server administration of a project wouldnt take that much time. [22:04:20 CEST] <kierank> https://usercontent.irccloud-cdn.com/file/FG7wCtwM/irccloudcapture-77481956… [22:04:42 CEST] <kierank> atomnuker and me getting drunk [22:04:56 CEST] <kierank> j-b: ^ [22:05:03 CEST] <j-b> well, I know kierank hates me, because he never drunk enough with me. [22:05:37 CEST] <thresh> you always have your excuses j-b, like "I have to prepare for a talk in the morning" [22:06:17 CEST] <ubitux> durandal_1707, BBB, J_Darnley, jamrial, and all the other active devs (sorry i forgot many ppl but hey), can you start thinking about what you are willing to conceed/change in the project and what you think must be kept (or is important in general) as it is if a potential project reunification happens? [22:06:39 CEST] <ubitux> (both technical & social stuff) [22:06:45 CEST] <ubitux> i'll send a mail over the week end [22:07:06 CEST] <ubitux> but the idea is just to have an overview of the different expectations and what we can do about it [22:07:15 CEST] <ubitux> and have a longer discussion at vdd in september [22:07:21 CEST] <BBB> Im open-minded [22:07:46 CEST] <ubitux> it can be anything really [22:08:07 CEST] <ubitux> some ppl @ libav wants public excuses on ffmpeg.org, that kind of thing ;) [22:08:23 CEST] <wm4> lu_zero wanted it, from what I've seen [22:08:24 CEST] <ubitux> but joke aside, i think it will help understanding each other a bit more [22:08:29 CEST] <j-b> You can ask to have naked pictures of kierank for example :) [22:09:11 CEST] <j-b> kierank: you are not on this picture [22:09:14 CEST] <ubitux> j-b: thx [22:09:27 CEST] <ubitux> so are we getting naked kierank photos when all of this is done? [22:09:58 CEST] <kierank> https://usercontent.irccloud-cdn.com/file/pAJDrBQb/irccloudcapture717622604… [22:09:58 CEST] <kierank> There done [22:10:04 CEST] <kierank> Two of us [22:10:18 CEST] <ubitux> damn [22:10:54 CEST] <J_Darnley> durandal_1707: because I haven't been working on it [22:11:24 CEST] <j-b> kierank: damn. [22:11:30 CEST] <j-b> kierank: I'm beaten. [22:11:30 CEST] <ubitux> btw, i think it would be wise to hear about "downstreams" (still afraid to use that word without quotes) [22:11:43 CEST] <ubitux> so j-b, wm4 (okay you're also a ff/av dev), etc [22:11:59 CEST] <ubitux> that would be nice to hear what you expect from a project reuinification [22:12:08 CEST] <ubitux> as a user that it [22:12:09 CEST] <J_Darnley> ubitux: I don't think I care anymore about the social side of the "conflict". I think I once had a reason for staying here but I don't remember what it was. Perhaps I was just attached to the FFmpeg name. [22:12:10 CEST] <ubitux> is* [22:12:22 CEST] <ubitux> J_Darnley: it doesn't need to be social [22:12:40 CEST] <ubitux> J_Darnley: "i want to keep the name and my code" is a typical expectation [22:13:28 CEST] <ubitux> we'll get into a lot of arguments about stuff like "i want to maintain that stuff and no one can stop me from pushing stuff on my code" that sort of thing, but i don't think that's the core of the problem [22:13:34 CEST] <J_Darnley> as for the technical, I don't know what people might propose. [22:13:44 CEST] <ubitux> i believe we'll get more problems such as "i don't want to hear about Carl" [22:13:49 CEST] <J_Darnley> I would prefer to not have to use some javascript heavy web service to submit and review patches though [22:14:01 CEST] <ubitux> yeah, that sounds about right [22:14:09 CEST] <ubitux> anyway, not expecting answer like this now [22:14:17 CEST] <J_Darnley> fair enough [22:14:18 CEST] <ubitux> i don't know what medium we will use to gather all of this [22:14:28 CEST] <ubitux> but just think about it, because it's going to be discussed [22:14:29 CEST] <J_Darnley> I will look out for emails later [22:16:36 CEST] <BBB> what code do we use and what project name do we use is probably a tail question [22:16:51 CEST] <BBB> I mean, its important, but if we start there, I dont think well get anywhere [22:17:02 CEST] <j-b> just prepare your conditions [22:17:06 CEST] <j-b> each of you [22:17:10 CEST] <j-b> then find common conditions [22:17:12 CEST] <BBB> couples dont start dating by signing their prenups [22:17:20 CEST] <BBB> couples start dating by having fun together [22:17:33 CEST] <j-b> yeah, but you're not couple, your colleagues [22:17:51 CEST] <BBB> most companies Ive interviewed it tried to get me excited, not argue about my salary or vacation days [22:18:14 CEST] <ubitux> i can tell you one thing BBB [22:18:16 CEST] <ubitux> it's not going to be fun [22:18:19 CEST] <ubitux> :D [22:18:23 CEST] <j-b> +1 [22:18:24 CEST] <ubitux> at least not at the beginning [22:18:31 CEST] <j-b> the only fun thing is kierank naked [22:18:37 CEST] <j-b> when drunk :D [22:18:45 CEST] Action: iive is having flashbacks from the time after takeover and before forking. [22:19:49 CEST] <baptiste> pics ? L( [22:20:15 CEST] <j-b> :D [22:20:33 CEST] <j-b> baptiste: when the projects are merged and people are drinking together at FOSDEM. All pictures will be released. [22:20:49 CEST] <BBB> fosdem is dirty :( [22:20:52 CEST] <baptiste> neat, would be nice for me that is soo faarr :( [22:20:53 CEST] <BBB> drinking should be at vdd [22:23:15 CEST] <durandal_1707> my condition is simple leave my code /commits as is, no vertical aligments, thanks [22:23:56 CEST] <BBB> if I dont have to align my code myself, its fine with me [22:24:02 CEST] <BBB> I dont mind if other people align my code [22:24:07 CEST] <BBB> its free software after all [22:24:56 CEST] <durandal_1707> they are stealing my blame points, so I prefer to do it myself [22:25:13 CEST] <j-b> it's untrue [22:25:22 CEST] <j-b> git blame -C -C -M will give you blame points back [22:26:18 CEST] <durandal_1707> still va sucks [22:26:21 CEST] <ubitux> but what if you just sort the functions by alphabetical orders? [22:26:58 CEST] <j-b> durandal_1707: then list those in your conditions. [22:27:57 CEST] <durandal_1707> I told you already my conditions, leave my precious alone [22:28:11 CEST] <j-b> durandal_1707: that's all? [22:28:12 CEST] <BBB> I believe one of the highest-ranked posts about ffmpeg on stackoverflow is what is the difference between libav and ffmpeg [22:28:28 CEST] <BBB> so that alone should tell us how confusing this whole thing has been for end users [22:28:30 CEST] <j-b> BBB: the answer is: it's not libVLC [22:29:13 CEST] <BBB> yup [22:29:29 CEST] <BBB> http://stackoverflow.com/questions/9477115/what-are-the-differences-and-sim… [22:29:34 CEST] <BBB> 269 upvotes [22:29:38 CEST] <BBB> the next highest has 196 [22:29:45 CEST] <BBB> (in my list, all posts tagged with ffmpeg) [22:29:56 CEST] <durandal_1707> j-b: its lots of code [22:31:53 CEST] <ubitux> anyway, i will mail the list tomorrow with my own expectations etc [22:32:10 CEST] <ubitux> not sure how private other devs want to keep it until vdd, but mine will be public [22:32:33 CEST] <ubitux> and they won't be to be discussed as a flame bait, more like like examples to discuss [22:32:41 CEST] <ubitux> anyway, detaching, see you then folks [22:33:59 CEST] <BBB> bye [22:34:00 CEST] <jamrial> ubitux: i don't have conditions, just preferences/expectations [22:34:08 CEST] <jamrial> and ok, i'll think about it [22:34:30 CEST] <BBB> I wonder what michaelni will do next [22:34:45 CEST] <jamrial> ask him? :p [22:34:53 CEST] <BBB> like, lets say you worked on a startup for 10 years and you just sold it to Bigcorp[TM] and they didnt want you, just your company [22:34:57 CEST] <BBB> now what? [22:35:16 CEST] <BBB> its like, hey kids, havent seen you for 10 years, youve grown so big! [22:35:21 CEST] <BBB> (or so) [22:35:52 CEST] <j-b> BBB: I hope michaelni will find what makes him happy [22:36:14 CEST] <j-b> but I'm not sure fighting makes anyone here happy/ [22:37:54 CEST] <jamrial> BBB: i guess something that he actually enjoys doing. the merges weren't fun and he said before he would have rather used that time to do more useful and interesting stuff like writting an hevc decoder a year before the one we got ever got written [22:39:05 CEST] <BBB> our hevc decoder is also somewhat slow [22:39:23 CEST] <BBB> let me see if I can find real numbers [22:39:24 CEST] <JEEB> there IIRC is still plenty of intrinsics left non-ported from openhevc [22:39:29 CEST] <BBB> even with that [22:39:34 CEST] <JEEB> oh [22:39:48 CEST] <jamrial> i ported some, but idct still remains unported [22:40:55 CEST] <BBB> same quality, ffvp9 gets 115fps, whereas ffhevc gets only 70fps. openhevc gets 85fps [22:41:08 CEST] <BBB> ffh264 gets 105fps [22:41:17 CEST] <jamrial> but seeing how hevc aims at being used for 4k 10bit content (blu ray 4k standard) i doubt any software implementation will be up to the task [22:41:19 CEST] <BBB> its significantly slower than other codecs [22:41:49 CEST] <peloverde> for a long time ffh264 was notoriously slow, it can get better [22:42:41 CEST] <j-b> wm4: FYI, I'm kidding (ml) [22:46:05 CEST] <rcombs> drop the concept of a single leader, have a few trusted people with sudo on relevant boxes, set up a process by which merges happen that doesn't involve eating one person's time all the time [22:50:35 CEST] <rcombs> unless a libav/ffmpeg reunion is actually practical (I really don't understand enough about the "politics" to know if that's true), and that would seem like the best route [22:59:08 CEST] <BtbN> I realy think some reviewboard-like code review tool would be extremely helpfull. Never worked with one of these, but i think a lot of stuff could be automated with it. [22:59:26 CEST] <BtbN> It's realy easy for stuff to get lost on the ML [23:00:48 CEST] Action: BBB will think also [23:00:49 CEST] <BBB> bbl [23:01:04 CEST] <BtbN> A lot of people won't give up on plain text mails though [23:01:07 CEST] <jamrial> BtbN: philipl suggested that a month ago [23:01:51 CEST] <j-b> BtbN: do you like github PR? gitlab's? webkit review process? Qt's? [23:01:59 CEST] <jamrial> but there wasn't much interest, judging by the lack of replies his email got [23:02:15 CEST] <BtbN> github PRs would be a horrible mess for a project of the scale of ffmpeg [23:02:20 CEST] <jamrial> it's something worth considering now, though [23:03:04 CEST] <BtbN> reviewboard or gerrit, which is what Qt and android uses iirc, seems good though [23:05:00 CEST] <j45> we use review board in handbrake. it makes the review process easier. but i don't think it solves the issue of stuff getting "lost". patches still get ignored. [23:15:35 CEST] <j-b> gn * [23:48:03 CEST] <rcombs> BtbN: what problems do you tend to have with GH PRs? [23:48:21 CEST] <rcombs> just easy for things to get "lost" if they pile up? [23:49:16 CEST] <BtbN> For small projects they are fine i think, but once stuff gets bigger, it gets realy hard to keep track of what's going on, and so far in every bigger project using them there's a huge pile of open old pullrequests. [23:49:39 CEST] <rcombs> probably still better than the current ML setup [23:50:00 CEST] <rcombs> with a huge pile of old patches and no tracking of what's done [23:50:02 CEST] <nevcairiel> maybe, but there are definitely better solutions [23:50:05 CEST] <BtbN> yeah, i'd personaly prefer it over a ml, but it's far from perfect. [23:50:28 CEST] <rcombs> yeah [23:50:31 CEST] <nevcairiel> but yeah thats the main problem with the ML, stuff just gets lost unless the dev actively pings it [23:51:19 CEST] <BtbN> I think the pure ML based process also stops some people from joining in, as it's not modern anymore. [23:51:38 CEST] <rcombs> wm4: btw, re: sidx, yeah, I wrote one, but I'd like someone who knows the relevant code better than me to help work out whether or not it would leak mem in seek cases [23:52:42 CEST] <BtbN> the problem of github pullrequests could propably be addressed with some kind of bot that pings inactive PRs and closes them if they stay inactive. [23:53:18 CEST] <rcombs> sure [23:53:36 CEST] <rcombs> and then it's still easier to reopen an old PR than to find an old email chain to reply to (IMO) [23:54:08 CEST] <BtbN> I wouldn't even know how to easily reply to an old mail if i don't have it in my mail client anymore. [23:54:11 CEST] <beastd> IMHO ml is good for patches and review because it is stable and asynchronous and distributed. OTOH ml alone has problems regarding patch tracking. [23:55:55 CEST] <rcombs> there are also some workflow questions in either case [23:56:42 CEST] <rcombs> the age-old rebase-vs-merge amongst them [23:56:44 CEST] <beastd> and isn't the real problem more about people who participate in patch reviewing and taking the blame of pushing? regardles of the medium for posting and discussing patches. [23:59:04 CEST] <philipl> I'm still on my holiday but my offer to astand up a reviewboard instance still stands. github prs are horrible. [23:59:26 CEST] <baptiste> well we could use phabricator [00:00:00 CEST] --- Sat Aug 1 2015

1 0

ffmpeg.log.20150731
by burek 01 Aug '15

01 Aug '15

[02:07:15 CEST] <bwilson> When running ffmpeg on a mac, Im attempting to list supported/connected devices with `ffmpeg -f avfoundation -list_devices true -i ` but I get Unrecognized option 'list_devices. [02:07:22 CEST] <bwilson> using the binary release from the main site [02:07:29 CEST] <bwilson> anyone run into this? [02:07:53 CEST] <bwilson> according to the docs, thats how to do it https://trac.ffmpeg.org/wiki/Capture/Webcam [02:09:14 CEST] <chungy> Did you use the curly quotes on the command line? Try using straight " quotes only. [02:09:42 CEST] <bwilson> no, thats just my irc client. :) [02:10:11 CEST] <bwilson> it looks like the binary release isnt build with avfoundation nor qtkit? [02:10:16 CEST] <bwilson> s/build/built/ [02:10:46 CEST] <chungy> I don't use Mac, but the "" thing stood out to me. [02:17:36 CEST] <c_14> bwilson: if you check the configure line, you can see that qtkit is disabled and avfoundation isn't enabled [02:18:43 CEST] Action: bwilson built with llvm-gcc 4.2.1 (LLVM build 2336.11.00) [02:18:44 CEST] Action: bwilson configuration: --prefix=/Volumes/Ramdisk/sw --enable-gpl --enable-pthreads --enable-version3 --enable-libspeex --enable-libvpx --disable-decoder=libvpx --enable-libmp3lame --enable-libtheora --enable-libvorbis --enable-libx264 --enable-avfilter --enable-libopencore_amrwb --enable-libopencore_amrnb --enable-filters --enable-libgsm --enable-libvidstab --enable-libx265 --disable-doc --arch=x86_64 --enable-runtime-cpudetect [02:21:24 CEST] <c_14> Can you check ffmpeg -devices? [02:21:29 CEST] <c_14> Should be listed there if it's compiled in. [05:57:13 CEST] <k_sze> Hello. [05:57:42 CEST] <k_sze> Is the entropy in the image data supposed to have a huge effect on the encoding speed of the x264 and ffv1 codecs? [05:58:09 CEST] <k_sze> disregarding I/O, that is. [07:26:29 CEST] <satinder> Hi , Dear friends can we detect vocals with ffmpeg in any mp4 video ? if yes then how ? anybody have any idea about that ?? [08:44:44 CEST] <harisk92> good morning [08:48:31 CEST] <harisk92> I'm using ffmpeg to output hls and I'm trying to create my own hls playlist with the chunks that were created.But I'm having the hard time with getting duration of chunks so I came to ask you a bit about how does ffmpeg writes chunks to storage does he starts and finish writing one chunk or in the same time he's writing to 2 different chunks [10:08:05 CEST] <satinder> Hi , Dear friends can we detect vocals with ffmpeg in any mp4 video ? if yes then how ? anybody have any idea about that ?? [11:03:21 CEST] <mathieu> I have a problem building ffmpeg from source with the --enable-shared flag [11:03:41 CEST] <mathieu> It errors with: /usr/bin/ld: libavutil/color_utils.o: relocation R_X86_64_32S against `.rodata' can not be used when making a shared object; recompile with -fPIC [11:03:41 CEST] <mathieu> libavutil/color_utils.o: error adding symbols: Bad value [11:03:41 CEST] <mathieu> collect2: error: ld returned 1 exit status [11:07:36 CEST] <JEEB> mathieu: it means that some of the dependencies aren't built with PIC [11:08:29 CEST] <mathieu> but then the enable-shared option doesnt work at all in ffmpeg? [11:08:46 CEST] <JEEB> it most definitely does work [11:08:54 CEST] <JEEB> as in, in general [11:09:32 CEST] <JEEB> if you have before this built ffmpeg from those sources, then make sure that you have the source directory clean (you can use out-of-source compilation or git clean to clean it all up) [13:02:46 CEST] <niknok> hi! I have a content which has 2 audio languages I was able to tag each audio track with its proper language but everytime I am adding a language metadata to an audio track...the video stream got the same metadata language metadata [13:06:44 CEST] <KarlFranz> Knowing your exact procedure would help. [13:19:43 CEST] <niknok> here is the command I used http://pastebin.com/C0abkSSq [13:28:59 CEST] Action: KarlFranz checks [13:36:33 CEST] <AlicanC> Instead of using '-an', can I disable audio using '-c:a none' or something like that? [13:37:16 CEST] <KarlFranz> Is there a reason for doing that? [13:38:34 CEST] <AlicanC> I am generating the ffmpeg command programatically (JavaScript). I would like to be able to `-c:a ${codec}` without having an if statement to check if `codec` is none and then do '-an' instead. [13:38:46 CEST] <AlicanC> Just trying to make my cod 0.62% cleaner. [13:38:50 CEST] <AlicanC> *code [17:39:10 CEST] <nyuszika7h> is it possible to use the Intel QuickSync encoder with ffmpeg? [17:43:52 CEST] <durandal_1707> you mean qsv thing? [19:41:40 CEST] <diegoviola> I want to convert a mp4 file to mkv and attach a subtitle to it, how do I go in doing this? [19:41:45 CEST] <diegoviola> can I do it with ffmpeg? [19:41:50 CEST] <diegoviola> I don't want to re-encode [19:43:43 CEST] <pzich> never done it, but maybe something like ffmpeg -i input.mp4 -i subtitle.srt -c copy output.mkv ? [19:47:58 CEST] <JEEB> you couldn't be able to -c copy it [19:48:06 CEST] <JEEB> you would have to convert the srt to 3gpp subtitles [19:48:14 CEST] <JEEB> which is probably something like mov_text in ffmpegspeak [20:16:29 CEST] <chungy> srt works perfectly fine in a mkv container [20:36:45 CEST] <AAA> hey all. has anyone successfully compiled ffmpeg with the qsv codec? (Intel Media SDK) [20:37:39 CEST] <BtbN> I'm sure at least the people who sent the patches have. [20:38:11 CEST] <BtbN> I didn't feel like digging through Intels stuff, so i gave up when they wanted me to sign up just to download the sdk. [20:38:32 CEST] <AAA> it's a mess and the docs aren't super helpful [20:38:58 CEST] <AAA> make: *** [libavcodec/qsvenc.o] Error 1 [20:40:01 CEST] <AAA> mfxVideoParam has no member named mfx [21:04:03 CEST] <niknok> is it possible to mux in a subtitle using ffmpeg? [21:05:43 CEST] <chungy> yes [21:09:21 CEST] <niknok> I will try to mux in an Arabic subtitle...do i need to specify that the subtitle is in Arabic? [21:09:59 CEST] <c_14> No [21:10:13 CEST] <c_14> You can tag it as such to make it easier to tell in the muxed file. [21:11:16 CEST] <niknok> ok thanks I will try to make 1 [21:12:38 CEST] <chungy> "-metadata:s:s:0 language=ara" would be a way in mkv to tag it as Arabic, but it's not strictly required. [21:13:01 CEST] <chungy> the colon syntax reads odd, but it's stream:subtitle:0 (0 being the first subtitle stream) [21:25:42 CEST] <niknok> the samples of command line I had read on the web are for mp4 and mkv files...i guess it is the same with a ts file [21:32:44 CEST] <niknok> speaking of language tags...I had tried earlier to add tags in audio streams. I just do not know why it also add a tag in the video stream [23:12:06 CEST] <diegoviola> how do I convert a mp4 file to mkv and also include a srt as part of the stream? [23:12:33 CEST] <c_14> ffmpeg -i mp4 -i srt -c copy -map 0 -map 1 out.mkv [23:12:57 CEST] <diegoviola> I don't know what -map will do [23:13:23 CEST] <diegoviola> looking at the man page [23:13:29 CEST] <c_14> Makes sure all streams from the first file are copied as well as the subtitle stream [23:14:04 CEST] <diegoviola> thanks [23:14:14 CEST] <diegoviola> is that optional though? [23:14:53 CEST] <diegoviola> I mean, won't ffmpeg do that already if I don't specify -map? [23:15:47 CEST] <c_14> IT picks 1 audio, 1 video, and 1 subtitle stream by default. [23:15:51 CEST] <c_14> If you have more, it's dropped. [23:16:04 CEST] <diegoviola> ok [23:16:20 CEST] <diegoviola> I only have 1 mp4 and 1 srt [23:16:25 CEST] <c_14> s/it's/they're/ [23:16:59 CEST] <c_14> But the mp4 might have several audio streams. [23:17:13 CEST] <diegoviola> it works, thanks [23:17:22 CEST] <diegoviola> c_14: it has only 1 audio stream [23:17:35 CEST] <c_14> Then you don't need the maps [23:18:09 CEST] <diegoviola> ok so the maps are useful when you have more than 1 subtitle / audio stream? [23:18:33 CEST] <c_14> yep [23:18:49 CEST] <diegoviola> thanks a lot [23:18:52 CEST] <diegoviola> got it [00:00:00 CEST] --- Sat Aug 1 2015

1 0