Ffmpeg-devel-irc

Download

ffmpeg-devel-irc@ffmpeg.org

June 2015

1 participants
60 discussions

ffmpeg-devel.log.20150629
by burek 30 Jun '15

30 Jun '15

[00:36:14 CEST] <philipl> For everyone's general amusement: I managed to carefully double the height of various vdpau elements (video surface, decoder, mixer) while keeping the output size unchanged. Then if I render only the 'top field' of the hevc frame, it provides the decoded frame correctly. If course, this prevents 4k video from being decoded as the maximum decoder height is capped. [03:20:02 CEST] <Compn> wm4 : is asf-new the default asf demuxer in mpv now ? [03:20:31 CEST] <Compn> i forgot what the testing protocol was when mplayer was testing new lavf features. 3 months? 6 months? [03:20:47 CEST] <cone-242> ffmpeg 03George Boyle 07master:02bd4d93c9df: fate/api-tests: Added dependency on $(FF_DEP_LIBS) [03:20:47 CEST] <cone-242> ffmpeg 03Ludmila Glinskih 07master:ca3b27455234: api-h264-test: build with another api test [03:21:21 CEST] <Compn> mplayer users found lots of bugs in previous lavf codecs and demuxers. when those were made default in mplayer for testing, since ffmpeg is more stable and mplayer is more experimental [03:22:22 CEST] <Compn> i dont really care which is default, but this is a format with at least 15 years of testing. it would be like rewriting the avi demuxer and making the new one default without testing millions of files. [03:22:42 CEST] <Compn> oh well no one cares about compn [03:25:47 CEST] <Compn> "lavf codecs" derp [03:26:06 CEST] <jamrial> well, difference is that (afaik) unlike asf, people went and hacked every single kind of audio and video codec known to man into avi :p [03:26:35 CEST] <jamrial> so rewriting avidec is a monumental task if you plan to also support crap like that [03:27:08 CEST] <Compn> asf has a few codecs , not as many as avi of course [03:27:36 CEST] <Compn> how do i test the new asf demux ? [03:27:48 CEST] <Compn> -f asf-o ? [03:34:13 CEST] <jamrial> asf_o i think [05:17:15 CEST] <cone-242> ffmpeg 03Michael Niedermayer 07master:be4e1f28fd65: avformat/asfdec: Print packet_obj_size in case it is invalid [05:17:16 CEST] <cone-242> ffmpeg 03Michael Niedermayer 07master:4ccd2b31f0e9: avformat/asfdec: Allow packet_obj_size == 0 [05:49:50 CEST] <jamrial> i find the act of fixing the old demuxer to "favor" the argument about it being better than the new somewhat misguided [06:45:32 CEST] <Compn> none of that matters [06:45:43 CEST] <Compn> only matters to test both demuxers against millions of samples [06:45:51 CEST] <Compn> i have the lists still [06:46:05 CEST] <Compn> videos_wmvp.txt videos_wvp2.txt [06:46:10 CEST] <Compn> from picsearch [06:46:21 CEST] <Compn> there were some more too. [06:47:31 CEST] <Compn> the weird thing is that libav complains ffmpeg adds any feature without testing, now libav wants to add a feature quickly and ffmpeg wants to test, now we are the bad guys wanting more testing? weird situation :P [08:13:55 CEST] <cone-242> ffmpeg 03Paul B Mahol 07master:21cede9e970a: doc/filters: fix documentation bug in ssim filter [10:18:48 CEST] <cbsrobot_> j-b: is this your new VLC ad in japan ? https://www.youtube.com/watch?v=MxC_Rl5INpY&feature=youtu.be&t=7 [10:20:39 CEST] <j-b> cbsrobot_: W-T-F [10:35:46 CEST] <JEEBsv> Compn: it's not as if the demuxer was not tested at all, but I do see the point. was the actual list of working/non-working samples and the test results published? [11:45:55 CEST] <ubitux> michaelni: i didn't understnad your answer @ http://ffmpeg.org/pipermail/ffmpeg-devel/2015-June/174337.html [11:46:02 CEST] <ubitux> my point was about the 2 branches having the same code [11:46:24 CEST] <ubitux> so that was on purpose (!yuv444 needs to be changed later or something?) [11:48:35 CEST] <michaelni> YUV should not use a transform (it doesnt), RGB should use the ICT for lossy and RCT for lossless, this should be easy to implement, [11:50:38 CEST] <michaelni> see G.2 and G.3 in the j2k spec, its quite simple to implement [11:51:02 CEST] <ubitux> ok [11:52:43 CEST] <michaelni> ICT & RCT are basically RGB->YUV type transforms one losslessly reversible [12:24:43 CEST] <cone-006> ffmpeg 03Shivraj Patil 07master:d9deae04a78b: avcodec/mips: MSA (MIPS-SIMD-Arch) optimizations for pixblock functions [13:02:22 CEST] <BBB> maybe we should do a vote on asfdemux [13:02:31 CEST] <BBB> I mean, & [13:07:00 CEST] <wm4> sounds like an interesting idea [13:07:25 CEST] <wm4> but isn't only cehoyos against? and mini pretends to be impartial as usual [13:23:37 CEST] <kierank> lol [13:57:28 CEST] <ubitux> did anyone share samples that were supported by one but not another? [14:08:29 CEST] <durandal_1707> c-e found sample that didn't worked with old demuxer [15:20:48 CEST] <Compn> new feature vs replacement feature [15:21:09 CEST] <Compn> i mean, you understand how many millions of files the old demuxer has been tested on [15:21:27 CEST] <Compn> possibly billions. so many spanking videos for j-b [15:21:44 CEST] <Daemon404> iirc vlc doesnt even use out wmv demux [15:21:48 CEST] <Daemon404> neither does mplayer or mpv. [15:21:52 CEST] <Daemon404> because its so broken. [15:21:55 CEST] <Daemon404> they all use their own. [15:22:10 CEST] <wm4> mpv dropped the mplayer one and suffered the consequences [15:22:19 CEST] <Daemon404> constant bug reports? [15:22:20 CEST] <kurosu> Has any of them be fuzzed? That would matter to some (not all) users. But then, such scrutiny isn't applied to most pieces of code [15:22:34 CEST] <Daemon404> old one has of course been [15:22:39 CEST] <Daemon404> during the j00ru phase. [15:22:59 CEST] <Compn> wtf @ that youtube clip. cbsrobot_ [15:23:08 CEST] <wm4> Daemon404: complaints about seeking being borked are common, but mostly they are against stuff like .ts or DVD (as expected) [15:23:25 CEST] <Daemon404> wm4, i think top complains for ffms2 are: .ts seeking is broken, wmv seeking is broken [15:23:31 CEST] <Daemon404> by several orders of magnitude [15:23:47 CEST] <Daemon404> and this is even with an index. [15:24:15 CEST] <Daemon404> some people on doom9 and github had taken to patching ffmpeg with the new asf demuxer (before it was in) [15:24:20 CEST] <Daemon404> and distributing ffms2 builds with it [15:24:22 CEST] <Daemon404> to work around it [15:25:21 CEST] <Daemon404> i wonder how well this new demuxer works on problematic long-gop wmv files [15:25:27 CEST] <Daemon404> this was one of the reasons vlc hated it iirc [15:25:30 CEST] <Daemon404> (dont quote me) [15:25:33 CEST] <kurosu> too bad nobody has time or expertise to fix these broken demuxers (iirc kierank complained ts couldn't fit the ffmpeg model anyway) [15:25:42 CEST] <Daemon404> it doesnt [15:25:59 CEST] <Daemon404> it is a consequence of having a 'generic' demux lib. [15:28:36 CEST] <Compn> has anyone tried to add a wrapper for another ts demuxer , or everyone just uses their ts demuxer and lavc and dont care about lavf anyway [15:36:13 CEST] <BBB> kierank: btw, regarding the api thing we talked about yesterday [15:36:31 CEST] <BBB> kierank: Id be really delighted if (over time, very long term) this could lead to better documentation of how seeking works [15:36:59 CEST] <BBB> kierank: not api-wise, but implementation wise, like in if you use demuxer XYZ, seeking works like this XYZ and has these limitations XYZ [15:37:33 CEST] <Daemon404> BBB, gl;hf [15:37:35 CEST] <BBB> (most people dont care about any of this, but try to do frame-accurate seeking and share single code between ivf and annexb/hevc, and youre basically in hell) [15:37:39 CEST] <Daemon404> it would would need to be demxuer specific [15:37:44 CEST] <Daemon404> e.g. some seek based on pts, some based on dts [15:37:46 CEST] <BBB> absolutely [15:37:48 CEST] <Daemon404> and based on *neither* [15:37:52 CEST] <Daemon404> some* [15:37:56 CEST] <BBB> but Id like to know how it works so that I can write things ffmpeg.c doesnt do [15:38:02 CEST] <BBB> right now thats not possible [15:38:09 CEST] <BBB> and I bet half of our demuxer authors dont know either [15:38:11 CEST] <BBB> which is really bad [15:38:14 CEST] <Daemon404> its possible if you know the internals [15:38:15 CEST] <BBB> because it likely means its undefined [15:38:18 CEST] <Daemon404> ive hit that wall a few times. [15:38:21 CEST] <BBB> I know the internals [15:38:24 CEST] <BBB> thats what makes it so bad [15:38:33 CEST] <BBB> imagine someone trying to do what Im trying to do, but not knowing the internals [15:39:11 CEST] <Daemon404> BBB, relevant from me: https://twitter.com/daemon404/status/612288114195865601 [15:39:18 CEST] <Daemon404> i say this as a long time API user, of couse. [15:39:41 CEST] <Daemon404> though that was about ubitux [15:39:42 CEST] Action: Daemon404 runs [15:40:25 CEST] <BBB> :D [15:40:33 CEST] <BBB> I appreciate the cynisism [15:40:38 CEST] <BBB> its appropriate [15:41:05 CEST] <Daemon404> btw i count mplayer as internal [15:47:47 CEST] <ubitux> :( [15:48:11 CEST] <Daemon404> ubitux, long time api users have no remorese [15:48:15 CEST] <Daemon404> remorse* [15:48:16 CEST] <Daemon404> ;) [15:56:41 CEST] <wm4> mplayer doesn't really use the API correctly [15:56:58 CEST] <Daemon404> did it stop cannibalizing finally? [15:57:02 CEST] <wm4> no [15:57:22 CEST] <Daemon404> didnt think so [15:57:29 CEST] <wm4> you'd think with development slowing down they'd make changes to make maintenance easier [15:57:52 CEST] <BBB> I think they just stop touching code [15:57:56 CEST] <BBB> it works now" [15:58:00 CEST] <BBB> and then they expect api stability [15:58:08 CEST] <wm4> mplayer? kind of [15:58:20 CEST] <wm4> I suspect mplayer is much of the reason some old APIs don't get removed [15:59:06 CEST] <Daemon404> did we nuke lavc's deint yet [15:59:17 CEST] <wm4> it's easy to say that users should update their API usage, but if you have to fix MPlayer code yourself... then it's not easy [15:59:33 CEST] <Daemon404> it only took you 3 years [15:59:38 CEST] <wm4> lol [16:00:01 CEST] <BBB> no lavc deint is still under a macro enabled by default [16:00:02 CEST] <BBB> sadly [16:00:12 CEST] <Daemon404> :( [16:00:16 CEST] <Daemon404> what even uses it? [16:00:18 CEST] <Daemon404> and why? [16:00:20 CEST] <wm4> we also have 3 vdpau APIs [16:00:23 CEST] <BBB> mplayer from 7 years ago [16:00:29 CEST] <Daemon404> o. [16:00:40 CEST] <wm4> (granted, the 3rd is very similar/builds on the 2nd, and is rather recent) [16:01:22 CEST] <wm4> Daemon404: also there's this guy on the ML who is a fan of the old deint API [16:02:08 CEST] Action: Daemon404 ponders sending a rm patch [16:02:17 CEST] <wm4> automatic flamewar [16:02:29 CEST] <Daemon404> ... has that ever stopped me? [16:02:43 CEST] <wm4> if you do I'll reply with +1 [16:02:47 CEST] <BBB> I think it needs a lavc version bump [16:02:58 CEST] <BBB> so youd need to do that (Im fine with it) [16:03:04 CEST] <Daemon404> BBB, it was deprecated several verison bumps ago [16:03:05 CEST] <Daemon404> wasnt it [16:03:33 CEST] <wm4> right, you probably need stupid "remove on next bump" ifdeffery [16:03:48 CEST] <Daemon404> i never understoof the point of that [16:03:48 CEST] <wm4> if it isn't already [16:03:49 CEST] <BBB> ../libavcodec/version.h:#define LIBAVCODEC_VERSION_MAJOR 56 [16:03:54 CEST] <Daemon404> if the version is already bumped [16:03:55 CEST] <BBB> ../libavcodec/version.h:#define FF_API_DEINTERLACE (LIBAVCODEC_VERSION_MAJOR < 57) [16:04:09 CEST] <Daemon404> hmm ok [16:04:29 CEST] <wm4> we all know the FF_API_DEINTERLACE def will be updated on the next actual bump [16:04:39 CEST] <BBB> also see f4c444e17d137c786f0ed2da0e5943df505d5f9e [16:04:48 CEST] <BBB> -#define FF_API_DEINTERLACE (LIBAVCODEC_VERSION_MAJOR < 56) [16:04:49 CEST] <BBB> +#define FF_API_DEINTERLACE (LIBAVCODEC_VERSION_MAJOR < 57) [16:04:55 CEST] <Daemon404> what the shit? [16:04:56 CEST] <BBB> I havent backtracked further than that [16:05:08 CEST] <wm4> that was silly... it was a major bump except they didn't want a real major bump [16:06:00 CEST] <durandal_1707> audioconvert [16:06:39 CEST] <Daemon404> durandal_1707, ... is that not dead yet? [16:06:47 CEST] <Daemon404> (if it's what i think it is [16:06:49 CEST] <wm4> ffzombie [16:07:10 CEST] <BBB> the version before that is my original patch which put it under the version conditional [16:07:15 CEST] <BBB> so I guess major was never bumped since then [16:07:27 CEST] <BBB> 54b298fe5650c124c29a8283cfd05024ac409d3a [16:07:42 CEST] <wm4> over 2 years ago [16:07:45 CEST] <durandal_1707> michaelni: old deprecated stuff need to go [16:08:10 CEST] <BBB> Date: Sun Mar 3 08:23:08 2013 -080 [16:08:11 CEST] <BBB> indeed [16:08:58 CEST] <Daemon404> at least carl wont object to API removal [16:09:02 CEST] <Daemon404> since he can only comprehend the CLI. [16:09:14 CEST] <wm4> heh [16:09:21 CEST] <nevcairiel> he did object the old vdpau shit removal, since his beloved mplayer uses that [16:09:41 CEST] <wm4> mplayer supports both now [16:09:56 CEST] <wm4> because apparently there were some regressions [16:10:28 CEST] <wm4> do we keep a list of "stupid things that should be fixed on the next major bump"? I recall av_init_packet having stupid semantics [16:10:40 CEST] <wm4> or does everything need to be ifdefferied? [16:10:59 CEST] <nevcairiel> ideally you want a replacement in early so people that follow master can update [16:11:20 CEST] <Daemon404> wm4, well my issue with the packet stuff is [16:11:24 CEST] <nevcairiel> and changing semantic of an existing function is always difficult since it can be hard to notice such changes [16:11:26 CEST] <Daemon404> some is av_packet_thing [16:11:30 CEST] <Daemon404> some is av_thing_packet [16:11:36 CEST] <Daemon404> i constantly fuck it up [16:15:00 CEST] <wm4> av_dup_packet is also a thing that must die [16:17:06 CEST] <wm4> Daemon404: how much would you hate me if I added av_packet_init() [16:17:16 CEST] <wm4> which, unlike av_init_packet, would actually init _all_ fields [16:20:39 CEST] <BBB> yeah can we kill av_dup_packet? [16:20:43 CEST] <BBB> the doxy says it all [16:21:07 CEST] <BBB> Warning [16:21:09 CEST] <BBB> This is a hack - the packet memory allocation stuff is broken. The packet is allocated if it was not really allocated. [16:21:42 CEST] <BBB> wm4: please dont add function duplicates, just fix av_init_packet? [16:21:44 CEST] <BBB> or is that silly? [16:21:55 CEST] <wm4> BBB: complaints about subtly different semantics [16:22:12 CEST] <BBB> right but now every time I typo the function name, it works but behaves differently [16:22:13 CEST] <wm4> av_init_packet doesn't set the size/data fields, and some code might rely on this [16:22:17 CEST] <BBB> instead of giving me a compile error [16:32:49 CEST] <cone-006> ffmpeg 03Michael Niedermayer 07master:da8b22897740: avformat/avio: Move avio_delete() avio_move() to avpriv_ namespace [16:58:31 CEST] <cone-006> ffmpeg 03Rostislav Pehlivanov 07master:7c10b87b5744: aacenc: add support for coding of intensity stereo scalefactor indices [17:34:58 CEST] <nevcairiel> wm4: you mean some code sets the fields before calling av_init_packet? that would seem like a rather terrible control flow [17:36:51 CEST] <wm4> yes, but this is how this function was "designed" [17:37:22 CEST] <nevcairiel> unfortunately changing this is a bit iffy, since it would be a silent failure [17:37:31 CEST] <nevcairiel> such things are annoying [17:37:36 CEST] <wm4> yep [17:38:42 CEST] <wm4> it was done this way by Fabrice Bellard in 2003 [17:40:55 CEST] <wm4> and for no apparent reason [17:41:02 CEST] <cone-006> ffmpeg 03Carl Eugen Hoyos 07master:a876a4da4a92: lavf/img2dec: Autodetect dds frames. [17:41:03 CEST] <cone-006> ffmpeg 03Carl Eugen Hoyos 07master:45441eb0f643: Changelog: Add jpeg 2000 improvements before they get forgotten. [17:41:04 CEST] <cone-006> ffmpeg 03Carl Eugen Hoyos 07master:5a458420e2fb: lavf/asfdec: Reduce minimum header size. [17:41:05 CEST] <cone-006> ffmpeg 03Carl Eugen Hoyos 07master:dee794381904: lavf/img2dec: Improve detection of valid Quickdraw images. [17:41:06 CEST] <cone-006> ffmpeg 03Carl Eugen Hoyos 07master:77c0b149be12: lavf/mpegts: Return 0 if the probe function does not detect mpegts. [17:41:07 CEST] <cone-006> ffmpeg 03Carl Eugen Hoyos 07master:4b920d7b4a58: lavf/msnwc: Return 0 if the probe function does not detect msnwc-tcp. [17:41:08 CEST] <cone-006> ffmpeg 03Michael Niedermayer 07master:e55e5be982a6: Merge remote-tracking branch 'cehoyos/master' [17:48:05 CEST] <Daemon404> i wonder why the MIPS guys are sending so much stuff to every media project [17:48:16 CEST] <Daemon404> last ditch attempt to not die? [17:48:44 CEST] <BtbN> isn't the chinese CPU a MIPS one, and they are the ones sending all the patches? [17:48:53 CEST] <Daemon404> longsoon? lol [17:49:51 CEST] <BtbN> Yep, seems to be MIPS. [17:50:28 CEST] <Daemon404> yes it is [18:06:47 CEST] <markos_> hi all, speaking of dying arches, I've done a bunch of altivec patches that give a 2-7% on x264 decoding, I have them on a forked github tree, what's the usual procedure, should I send them one by one or combine them? they pass the regression tests and most of the videos I've tested, but this is a first time, so I might be missing sth [18:07:07 CEST] <markos_> (they are against current master btw) [18:08:27 CEST] <BtbN> you should propably ask on the x264 ml. Or are they against ffmpeg? [18:08:39 CEST] <markos_> https://github.com/markos/FFmpeg/commits/master [18:08:42 CEST] <markos_> against ffmpeg [18:09:11 CEST] <BtbN> Hm? What do they do then? x264 is primarily an external library [18:09:36 CEST] <J_Darnley> I think you mean H.264 decoding [18:09:45 CEST] <BtbN> oh, _de_coding [18:09:56 CEST] <markos_> sorry, yes, bad choice of words :) [18:10:59 CEST] <J_Darnley> it happens especially when one software dominates a particular area [18:12:14 CEST] <BtbN> I'd guess it happens because certain content providers put x264 in their filenames. [18:12:58 CEST] <Daemon404> J_Darnley, its worse when the scene picks it up and uses it incorrectly [18:13:02 CEST] <Daemon404> as they wont to d [18:13:02 CEST] <Daemon404> o [18:32:43 CEST] <cone-006> ffmpeg 03Paul B Mahol 07master:9842d6707fd6: avfilter/avf_showvolume: optionally display channel names [19:13:34 CEST] <Compn> markos_ : send a pull request [19:13:42 CEST] <Compn> i guess [19:13:46 CEST] <Compn> or push request [19:13:49 CEST] <Compn> whatever its called [19:14:38 CEST] <Daemon404> a patch to the ml [19:18:22 CEST] <Compn> that too [19:18:50 CEST] <markos_> the one generated by git format-patch like mentioned in the howto, I guess? [19:21:32 CEST] <jamrial> yes [19:22:03 CEST] <jamrial> if you can't get git send-email to work, just attach the file created by format-patch into an email [19:24:06 CEST] <markos_> yeah, that's my worry atm, I wouldn't want to fill the list with test mails just to make sure it works :) [19:47:45 CEST] <tizbac> hello, i'm trying to implement frame interpolation on libavfilter to allow converting videos from 24/30 fps to 60 fps, https://github.com/tizbac/FFmpeg/commit/ac73e47cf3b5b3b3a066425fc13db238d1b… , problem is i can't get a reliable motion estimation and as a consequence, the result is full of misplaced squares [19:48:28 CEST] <tizbac> does anyone have suggestions on how to do it correctly? [19:49:07 CEST] <nevcairiel> how do you expect to get motion information? [19:49:24 CEST] <tizbac> i've done it using diamond search [19:49:38 CEST] <tizbac> to get the motion vectors between 2 frames for each block [19:50:10 CEST] <nevcairiel> well motion detection is one of the larger problems for every video codec, you can try to benefit from their research, but its probably never going to be perfect [19:50:27 CEST] <tizbac> i know, but svp project has it working [19:50:32 CEST] <kierank> svp project? [19:50:36 CEST] <tizbac> so somehow it's possible to get a decent result [19:50:41 CEST] <nevcairiel> their quality is terrible, imho [19:51:39 CEST] <tizbac> it's still better than 24 fps [19:51:39 CEST] <nevcairiel> kierank: its a fancy installer and config GUI for various avisynth motion interpolation plugins [19:51:45 CEST] <kierank> ah [19:51:48 CEST] <kierank> mvtools is good [19:51:55 CEST] <Daemon404> http://i4.minus.com/ibjPtPu1AKZLlh.jpg [19:51:58 CEST] <Daemon404> yeah svp is great... [19:51:59 CEST] Action: Daemon404 runs [19:51:59 CEST] <kierank> I really dont understand people who do frame interpolation [19:52:10 CEST] <tizbac> kierank, yeah , problem with mvtools is that source code is no more readable than assembler [19:52:11 CEST] <kierank> I accept high frame rate [19:52:19 CEST] <wm4> mvtools got ported to vapoursynth [19:52:24 CEST] <kierank> but I don't accept creating frames out of nowhere [19:52:27 CEST] <nevcairiel> tizbac: well its avisynth, thats expected [19:52:40 CEST] <Daemon404> wm4, rewritten* [19:52:55 CEST] <wm4> right, not like that shit could actually be ported [19:53:08 CEST] <tizbac> wm4, is there any source code of the rewritten version? [19:53:17 CEST] <wm4> yes [19:53:39 CEST] <wm4> this one I think? https://github.com/dubhater/vapoursynth-mvtools [19:54:40 CEST] <Daemon404> you know mvtools just uses x264's code in some bits right? [19:54:47 CEST] <Daemon404> for ME [19:55:27 CEST] <tizbac> Daemon404, yeah i've seen that [19:57:31 CEST] <tizbac> kierank, compared to real 60 fps it sucks, but compared to 24 fps it's much better [19:57:45 CEST] <tizbac> especially on big screens [19:57:57 CEST] <Daemon404> i can only disagree [19:58:01 CEST] <nevcairiel> I wouldnt agree, personally [19:58:05 CEST] <nevcairiel> but everyone their own [19:58:06 CEST] <Daemon404> interpolated 24->60 looks awful to my eyes [19:58:13 CEST] <kierank> yes it looks too artificial [19:58:39 CEST] <nevcairiel> I used to enjoy the processing my TV offered for a bit, until I saw terrible artifacts in a couple scenes, and i turned it off :d [19:59:12 CEST] <tizbac> yeah there are artifacts on scenes like when james is on top of the speeding train in "Skyfall" movie [19:59:40 CEST] <nevcairiel> any high motion scenes or scenes with motion in several directions is going to give you issues [20:00:25 CEST] <rcombs> Daemon404: svp BURNING LOVEs you [20:00:44 CEST] <Daemon404> im afraid any anime references go over my head nowadays [20:01:04 CEST] <rcombs> wrong bote anyway [20:35:56 CEST] <llogan> https://wiki.debian.org/Debate/libav-provider/ffmpeg [20:39:22 CEST] <philipl> wow. That's quite a turnaround by some Debian folks. [20:40:30 CEST] <wm4> "Libav even dropped support for version 0.8, which is used in Debian Wheezy, already:" [20:40:36 CEST] <wm4> these arguments are harmful for the ffmpeg project [20:40:39 CEST] <wm4> but whatever [20:49:13 CEST] <durandal_1707> why? [20:49:37 CEST] <wm4> because it implies that maintaining crap forever is a thing worth doing [20:49:47 CEST] <wm4> which diverts important resources and effort [20:50:21 CEST] <wm4> if debian wants to keep really ancient stuff, why don't they maintain it? it's their self-declared job after all [20:50:45 CEST] <J_Darnley> They'll do another openssh if they do that. [20:57:54 CEST] <Daemon404> it wouldnt be that bad if ffmpeg was sane [20:58:01 CEST] <Daemon404> and had 1 or 2 point releases, maintained [20:58:02 CEST] <Daemon404> bot 40 [20:58:05 CEST] <Daemon404> not* [21:15:23 CEST] <Zeranoe> If I was to manually remove all source files that are GPL only and I compile without --enable-gpl, will the build compile and work? [21:17:48 CEST] <Daemon404> wat [21:19:29 CEST] <Zeranoe> Daemon404: Can a build be compiled that's LGPL only from source that contains only LGPL components [21:24:12 CEST] <JEEBsv> as long as you don't break the build system around it, yes [21:25:25 CEST] <JEEBsv> the build system should set which files are compiled [21:25:48 CEST] <JEEBsv> https://github.com/FFmpeg/FFmpeg/blob/master/libavcodec/Makefile [21:26:00 CEST] <JEEBsv> like the stuff here [21:26:04 CEST] <J_Darnley> I would guess that somebody might be interested in (and fix) any errors you encounter. [21:26:07 CEST] <Daemon404> whats the goal here though [21:26:12 CEST] <Daemon404> i dont get it [21:26:23 CEST] <J_Darnley> Paranoia? [21:26:27 CEST] <JEEBsv> if a feature is not selected, that .o file is not generated from the same named .c file [21:26:41 CEST] <JEEBsv> aka that file is not compiled [21:32:33 CEST] <cone-006> ffmpeg 03Michael Niedermayer 07master:47f4e2d8960c: avcodec/pngdec: Only allow one IHDR chunk [21:40:22 CEST] <durandal_1707> nevcairiel: you already use new add demuxer? [22:17:43 CEST] <nevcairiel> JEEBsv: there are one or two things that are GPL only in the middle of other files [22:17:50 CEST] <nevcairiel> iirc something in h264 [22:20:29 CEST] <JEEBsv> nevcairiel: yes, but those are generally within ifdefs, no? [22:20:55 CEST] <nevcairiel> sure, but finding and removing them might end up slightly more annoying than just excluding lavfi filters which are GPL, for instance. [22:21:06 CEST] <JEEBsv> sure [22:21:26 CEST] <JEEBsv> which is why generally you just trust that the system was more or less sane'ish [22:21:31 CEST] <nevcairiel> not that I get the point of even wanting to do that, but shrug [22:21:33 CEST] <JEEBsv> and if you configure it as LGPL you get LGPL shit [22:26:06 CEST] <Compn> seriously, asf_o demuxer ? [22:26:17 CEST] <Compn> seriously? people defending this? [22:27:58 CEST] <Compn> ffmpeg -f asf_o -i http://samples.ffmpeg.org/A-codecs/WMSP/wma9audio-notworkmplayer.wma out.wav [22:28:00 CEST] <Compn> huhuhu [22:29:41 CEST] <cone-006> ffmpeg 03Michael Niedermayer 07master:a1736926e9ae: avcodec/pngdec: Require a IHDR chunk before fctl [22:29:42 CEST] <cone-006> ffmpeg 03Michael Niedermayer 07master:f1ffa01dd3da: avcodec/pngdec: Copy IHDR & plte state from last thread [22:31:08 CEST] <JEEBsv> Compn: thank you for actually linking a sample instead of lolrhetoric [22:31:45 CEST] <Zeranoe> Looks like these are all the source files that contain GPL components http://paste.debian.net/plain/266857 [22:31:59 CEST] <Compn> JEEBsv : i'm sad no one else linked a sample :( [22:32:15 CEST] <Daemon404> of course not [22:32:20 CEST] <Daemon404> carl admitted he didnt even test it [22:33:32 CEST] <Compn> guys give it a rest will you? :P [22:33:39 CEST] <Compn> actually, dont care [22:33:44 CEST] <Compn> nevermind [22:34:32 CEST] <Compn> also sorry for late testing delay, i had to wait for nightly build :P [22:34:37 CEST] Action: Compn too lazy to compile [22:34:47 CEST] <durandal_1707> cropdetect is gpl? [22:35:09 CEST] <JEEBsv> I almost wanted to joke if some nightly builds actually worked for you [22:36:00 CEST] <durandal_1707> they dont? [22:37:16 CEST] <Compn> Zeranoe's builds work fine for me [22:37:20 CEST] <Compn> (i'm on windows atm) [22:39:16 CEST] <JEEBsv> durandal_1707: IIRC Compn was on ye olde NT5 for a very long time [22:39:19 CEST] <JEEBsv> and I don't mean 5.1 [22:39:23 CEST] <Compn> win2k [22:40:48 CEST] <Compn> [asf_o @ 031c0140] Could not find codec parameters for stream 0 (Video: mpeg4 (MP4S / 0x5334504D), n [22:40:48 CEST] <Compn> one, 640x480): unspecified pixel format [22:40:48 CEST] <Compn> Consider increasing the value for the 'analyzeduration' and 'probesize' options [22:40:48 CEST] <Compn> bug821-2.asf: could not find codec parameters [22:40:51 CEST] <Compn> argh bad paste [22:41:03 CEST] <Compn> -f asf_o http://samples.ffmpeg.org/mplayer-bugs/bug821/bug821-2.asf [22:41:05 CEST] <Compn> ehe [22:41:10 CEST] <Compn> er dont do the http [22:41:12 CEST] <Compn> download it first [22:41:19 CEST] <Compn> -f asf_o bug821-2.asf [22:49:23 CEST] <cone-006> ffmpeg 03Michael Niedermayer 07master:b54ac8403bfe: avcodec/pngdec: Check values before updating context in decode_fctl_chunk() [22:51:15 CEST] <Compn> -i...... argh i'm losing it today [22:52:39 CEST] Action: JEEBsv hands Compn a lollipop [23:02:46 CEST] <Daemon404> a a wild thierry appears [00:00:00 CEST] --- Tue Jun 30 2015

1 0

ffmpeg.log.20150629
by burek 30 Jun '15

30 Jun '15

[00:28:24 CEST] <anoop_r> http://www.qfs.mobi/f2403835 [00:28:55 CEST] <anoop_r> please check this and forward your valuable feed back to keepitsimplepltd(a)yandex.com [01:01:50 CEST] <user> hi channel [01:03:30 CEST] <Guest60595> May I ask a question related to using ffmpeg as web-cam simulator ? (I search a lot so far about my problem...with not success) [01:04:11 CEST] <c_14> web-cam simulator? [01:06:16 CEST] <Guest60595> I have done a setup with v4l2loopback and snd-aloop. FFmpeg serve as a video stream generator for simulating a web-cam. [01:06:29 CEST] <c_14> What's the problem? [01:07:45 CEST] <Guest60595> I use a static 800x600 png as video input source and -f v4l2 /dev/video0 as output. [01:08:54 CEST] <Guest60595> I want a lightweight as possible video stream with 1 fps. I can't motivate ffmpeg to deliver that. [01:09:37 CEST] <c_14> ffmpeg -loop 1 -i png -r 1 -f v4l2 /dev/video0 [01:13:01 CEST] <Guest60595> Ok problem 1. solved. Bug: -framerate 1 do not work -r 1 do. THANK YOU. [01:18:19 CEST] <Guest60595> 2. problem: The CPU consume is 151.2% (two cores). I tried to switch: -f rawvideo -pixel_format yuv420p No error messages but the status line state still rawvideo (RGB[24] Any idea ? [01:23:12 CEST] <c_14> -pix_fmt not -pixel_format [01:29:46 CEST] <Guest60595> OK. -pix_fmt change now the outout (I get this from some a howto, I should had verified it. apologies) But flash broadcast app does not recognize yuv420p and it not change the very high cpu usage. :( [01:31:06 CEST] <c_14> add -re before -i png [01:33:18 CEST] <Guest60595> 59% cpu. I can live with that. You are know this beast very well :-) Thank you. [01:41:14 CEST] <Guest60595> I will put 2-3 (working, tested) lines together as a example for using ffmpeg to create a nice virtual web-cam. I thing a good addition for the FFMprg Documentation. [01:41:28 CEST] <Guest60595> bye [02:23:03 CEST] <nashgul> hi [07:19:34 CEST] <wn33dk> Hi, does anyone have time for a noob question? [07:22:00 CEST] <wn33dk> I have a video format that has embedded subtitles (mkv). I would like to convert this to another format, but have the subtitles be on and part of the recording itself. Thanks in advance if anyone can help. [09:14:49 CEST] <brewtany> hi, how can i set ts file start time exactly? [09:24:32 CEST] <pentanol> brewtany ffmpeg -ss [09:25:07 CEST] <brewtany> pentanol: -ss cannot set the start time of ts [09:25:32 CEST] <brewtany> pentanol: i want to create ts files for m3u8 file [09:28:20 CEST] <brewtany> first I do this [09:28:21 CEST] <brewtany> ffmpeg -y -i 11.mp4 -c:v libx264 -f mpegts -filter:v setpts=PTS+10.0/TB 11.tst [09:28:54 CEST] <brewtany> then how can I set the start time exactly at 10seconds? [09:29:14 CEST] <pentanol> -ss 10 [09:31:04 CEST] <brewtany> but when : ffmpeg -y -i 12.mp4 -c:v libx264 -f mpegts -filter:v setpts=PTS+20.0/TB 12.tst [09:31:11 CEST] <brewtany> -ss 20 not work. [09:31:43 CEST] <brewtany> ffmpeg -y -copyts -ss 20 -i 12.tst -c copy -f mpegts 12.ts [09:31:58 CEST] <brewtany> Input #0, mpegts, from '12.ts': [09:31:59 CEST] <brewtany> Duration: 00:00:05.00, start: 27.866667, bitrate: 348 kb/s [09:34:16 CEST] <pentanol> brewtany ffmpeg show time what it plays [09:34:30 CEST] <pentanol> sorry I need to go [09:34:44 CEST] <brewtany> help~ [09:34:48 CEST] <brewtany> plz don't go [09:35:09 CEST] <brewtany> show time? [12:23:09 CEST] <sine0> exit [12:23:11 CEST] <sine0> fail [12:25:01 CEST] <maqr> i'm trying to force a key frame at exactly 0.05 seconds into my video stream, i thought that would be done with '-force_key_frames:v 00:00:00.05', but that doesn't seem to do it (as checked with ffprobe -show_frames)... any idea what i could be doing wrong? [12:27:48 CEST] <maqr> i think it might have something to do with my stream being generated from testsrc [15:37:56 CEST] <mateo`> hello there o/, i'm trying to mux raw h264 into an mp4 container using the avformat API. What i get from the decoder is the codec data, then the actual frames. Muxing the codec data then all the frames using av_interleaved_write_frame seems to work (with an error at the beginning of the stream "no frame!") but i'm wondering if i'm not supposed to set the codec data more specifically somewhere else (ie: in [15:37:58 CEST] <mateo`> AVCodec/AVCodecContext for example) ? [16:39:21 CEST] <BtbN> Has anyone ever seen a camera that outputs files which file identifies as FoxPro FPT? [17:07:17 CEST] <Anoia> I don;t recognise an fpt image ofrmat [17:07:45 CEST] <BtbN> FoxPro seems to be some kind of DBMS oO [17:08:57 CEST] <Anoia> it is [17:09:07 CEST] <Anoia> https://support.microsoft.com/en-us/kb/130686 says it's a "table memo" [17:09:18 CEST] <Anoia> if it is inded a .fpt file [17:12:42 CEST] <BtbN> it has no extension [17:12:54 CEST] <BtbN> file just says it is [17:13:05 CEST] <BtbN> but i'm not sure if it's right [17:14:31 CEST] <DHE> ffprobe ? [17:14:47 CEST] <BtbN> is unable to make anything of the file [17:14:55 CEST] <BtbN> might as well give it a bit of random noise [17:16:00 CEST] <janicez> lol [17:20:44 CEST] <Anoia> Windows won;t say what type it is if it didn;t have an extension (it doesn;t do magic number detection) [17:20:51 CEST] <Anoia> it does hide soem extensions though [17:22:55 CEST] <BtbN> Who said anything about windows? [17:24:34 CEST] <Anoia> You mentioned FoxPro, which is a Microsoft package (only on windows IIRC) [17:25:56 CEST] <BtbN> yes, because "FoxPro FPT" is litteraly what file says [17:37:08 CEST] <Anoia> oh, the "file" command? That's normally based on a magic number [17:50:53 CEST] <_Vi> How do I limit avformat_find_stream_info when calling FFmpeg tool? It reads and caches entire video and then I get a lot of " Delay between the first packet and last packet in the muxing queue is ...". [19:14:21 CEST] <Timster> Hey, guys - could somebody explain me the principle behind Keyframe Intervals / GoP-Length? I am asking in regard to visual quality. Is a higher, or a lower amount better? [19:15:27 CEST] <DHE> varies by codec. in H264 there's no harm in any setting except that any seek always lands on a keyframe/GOP boundary [19:15:41 CEST] <kepstin-laptop> in most recent codecs, longer interval between keyframes means smaller files for equivalent file size, at the expense of making seeking more expensive [19:15:46 CEST] <DHE> MPEG2 will suffer quality issues over the longer term and will usually suggest 2 keyframes per second (approximately) [19:16:13 CEST] <Timster> Ok, and if we talk about libx264? [19:16:13 CEST] <DHE> keyframes are always large though, so there's that [19:16:37 CEST] <DHE> that's H264. larger intervals is better for bitrate, but worse for seek accuracy. [19:17:03 CEST] <Timster> Ok, and what do you mean with "better for bitrate"? [19:18:42 CEST] <kepstin-laptop> keyframes are big. With x264, using fewer of them means more efficient encoding in most cases. [19:19:22 CEST] <kepstin-laptop> (x264 will still by default still put in keyframes on scene changes and stuff even with a large keyframe interval setting) [19:20:04 CEST] <Timster> Ok, thanks. Let's say I use -r 40 and -g 80 - would the keyframe interval be fixed to 2 seconds by that, or do I need to use something else to really fix key frames? [19:20:57 CEST] <Timster> *so force it to be 2 seconds [19:22:02 CEST] <kepstin-laptop> if you're encoding with x264 and want fixed keyframe timing, you have to disable scenecut detection. I don't remember the ffmpeg option for that... [19:22:43 CEST] <kepstin-laptop> (but I'm not sure why you would want that, unless you're doing something like segmented files for streaming) [19:23:00 CEST] <Timster> Yeah, it's for streaming. Maybe: -x264opts "keyint={{fps}*2}:min-keyint={{fps}*2}:no-scenecut" ? [19:26:58 CEST] <DHE> scenecut=-1 is the preferred syntax in the most recent versions (I ride Git though) [19:27:06 CEST] <maqr> if i know i'm going to need to take a screenshot at a very specific time, like 0.05 seconds into an x264 stream... can i force a keyframe to be exactly there? [19:28:32 CEST] <Timster> DHE: So in my case it would be -r 40 -x264opts "keyint=80;scenecut=-1" ? [19:28:46 CEST] <DHE> colon, not semicolon [19:29:17 CEST] <DHE> and that looks right [19:29:19 CEST] <maqr> when i try force_key_frames it doesn't seem to work [19:32:07 CEST] <Timster> Oh, OK - thanks - will try that. So you say with libx264 there is no visual difference between different amounts of keyframes - it's just about the bitrate as well as the seek accuracy? In my setup, the stream gets reencoded at a last stage where I need to have a 2 seconds Keyframe Interval - so does it matter how big or little it is in the first encoding before that? [19:32:45 CEST] <Timster> Like to get the best out of the second encoding? [19:33:17 CEST] <DHE> assuming disk space is a non-concern (we're talking a few percentage points anyway) I would say don't worry about it [19:33:33 CEST] <DHE> though transcoding a transcoded video is considered bad form [19:39:13 CEST] <Timster> Ok, thanks a bunch! [19:40:36 CEST] <Timster> Would you recommend -strict -2 ? [19:41:52 CEST] <DHE> only if it says you need it, and you accept the warning it gives you about it [19:42:49 CEST] <Timster> Could you explain what it does exactly and what the ad- and disadvantages might be? [19:44:32 CEST] <DHE> it unlocks codecs that are considered "experimental" [19:44:46 CEST] <DHE> and/or codec features of the same quality [19:46:30 CEST] <Timster> Ok, thanks. One last question - as I need to have CBR as well, -x264opts "keyint=80:scenecut=-1" will not work for me, right? So I need to stick to -g 80? [19:55:24 CEST] <maqr> oh, i see what i was doing wrong, heh [19:57:09 CEST] <maqr> if i put "-ss time -q:v 1 -vframes 1 foo.jpg" at the end of my ffmpeg command that uses testsrc, it can pull out a frame that won't actually be in the video's encoding... how can i tell the -ss command to use the encoded output as the input stream, instead of the source input stream? [19:57:51 CEST] <maqr> in other words, i'm saying "-ss 00.00.00:0.05" and it's giving me a frame with "5" on it, which i thought was right, but i'm at 30fps and that frame can't exist [19:58:25 CEST] <maqr> i could just change it to "0.0333", but i'm wondering if there's a way to tell -ss to use the encoded stream instead of the source stream anyway [20:01:24 CEST] <Nolski> Hey, has anyone had to loop a section of a video n times? [20:01:56 CEST] <Timster> Gtg, bye [20:15:27 CEST] <llogan> Nolski: one method here: https://trac.ffmpeg.org/wiki/Concatenate [20:15:34 CEST] <llogan> search for "loop a video" [20:29:01 CEST] <Nolski> llogan: the issue is if I want to loop a /section/ of a video [20:29:54 CEST] <Nolski> Would I have to rip the video out and then concat it inbetween two videos? [20:31:50 CEST] <llogan> yes, that would be one way. either use the concat demuxer or concat filter. [22:32:24 CEST] <rizotas> hi, i'm installing vlc from source, but I got this /usr/bin/ld: /usr/local/lib/libpostproc.a(postprocess.o): relocation R_X86_64_32S against `.rodata' can not be used when making a shared object; recompile with -fPIC /usr/local/lib/libpostproc.a: error adding symbols: Bad value, then I compiled ffmpeg from source with --enable-pic, but I get the same error [22:44:47 CEST] <Marcin_PL> Hello. What I am missing on vps (debian7), that avconv reading input stream prints metadata (icy-sth, StreamTitle etc.) on my Deb8 at home, and on vps it can't? [22:45:22 CEST] <JEEBsv> possibly just an older version in an older version of debian [22:45:51 CEST] <Marcin_PL> Huh, wait& [22:46:33 CEST] <JEEBsv> most people who use OSS multimedia build stuff themselves since you generally want only X things (something in between of 2 to 4) in addition to what is in the code base itself [22:47:25 CEST] <JEEBsv> like, libx264 and fdk-aac for AVC/AAC encoding and pretty much all the decoders are in the code base [22:47:36 CEST] <JEEBsv> or libvpx and libopus [22:47:48 CEST] <Marcin_PL> avconv version 11.4-6:11.4-1~deb8u1 mine, avconv version 0.8.17-6:0.8.17-1 on vps& Huh, 0.8 to 11.4? How it is possible? [22:47:54 CEST] <Marcin_PL> I'll try install [22:48:05 CEST] <JEEBsv> yes, old things have old versions :P [22:48:07 CEST] <JEEBsv> nothing new really [22:48:15 CEST] <JEEBsv> although quite ouch with 0.8 [22:48:43 CEST] <JEEBsv> also this stuff is from Libav, not FFmpeg [22:49:03 CEST] <Marcin_PL> Isn't that just a fork? [22:49:26 CEST] <JEEBsv> yes, libav and ffmpeg-new are both forks of ffmpeg-old [22:49:27 CEST] <Marcin_PL> Debian treats it that way [22:49:38 CEST] <Marcin_PL> Ah. [22:50:08 CEST] <Marcin_PL> I can't have those libraries from the repo& [22:50:08 CEST] <JEEBsv> anyways, as of very late you can get both ffmpeg and libav binaries packaged in debian [22:50:21 CEST] <JEEBsv> Marcin_PL: yes. you have to build. thankfully it's not really hard [22:50:30 CEST] <Marcin_PL> I see [22:50:43 CEST] <JEEBsv> git, build-essential(s), yasm (>=1.2) [22:50:50 CEST] <Marcin_PL> Perhaps I'd decide to upgrade server to deb8 [22:51:26 CEST] <JEEBsv> if that's an alternative for you, sure :P [22:51:43 CEST] <Marcin_PL> The point is, everything is working, but I want now to monitor stream name to auto start recording [22:51:44 CEST] <JEEBsv> I'm just going to say that in any case those packages are going to get old relatively quickly [22:52:02 CEST] <JEEBsv> so getting used to building OSS multimedia is not a bad idea [22:52:37 CEST] <relaxed> Marcin_PL: I have static builds if that would be easier. [22:52:42 CEST] <Marcin_PL> Well I must rethink that, cause the server is quite slow [22:54:29 CEST] <JEEBsv> not that you build these things daily [22:54:41 CEST] <JEEBsv> unless you are into keeping up with master HEAD [22:56:47 CEST] <Marcin_PL> Well, I'll try things out, thanks for help [23:16:49 CEST] <rizotas> thanks you are very useful [23:17:00 CEST] <Nolski> How do I concat 3 videos? [23:29:56 CEST] <relaxed> Nolski: there are numerous ways, which container? [23:41:46 CEST] <Nolski> relaxed: it depends. They will all be in the same container [23:53:04 CEST] <Nolski> It seems that 'ffmpeg -i "concat:beginning.mp4|loops.mp4|end.mp4" -c copy looped.mp4' throws No such file or directory [23:56:34 CEST] <llogan> don't use the concat protocol with those inputs. use concat demuxer or concat filter [00:00:00 CEST] --- Tue Jun 30 2015

1 0

ffmpeg-devel.log.20150628
by burek 29 Jun '15

29 Jun '15

[00:14:22 CEST] <cone-389> ffmpeg 03Michael Niedermayer 07master:29b09233249e: avcodec/h264_slice: Silence pointer type warnings [00:14:23 CEST] <cone-389> ffmpeg 03Michael Niedermayer 07master:6323647c3550: avcodec/h264_slice: Reformat IN_RANGE() uses to be readable [00:53:12 CEST] <cone-389> ffmpeg 03Rostislav Pehlivanov 07master:d71935f8833d: aac: add additional fields needed by the encoder for intensity stereo [01:07:20 CEST] <cone-389> ffmpeg 03Ludmila Glinskih 07master:86fb20324690: api-flac-test: Fix the bug of comparing zero bytes [02:57:31 CEST] <cone-389> ffmpeg 03Michael Niedermayer 07master:f01028bf58b8: avcodec/jpeg2000: Avoid one shift in ff_jpeg2000_ceildivpow2() [02:57:32 CEST] <cone-389> ffmpeg 03Michael Niedermayer 07master:cd1a9908b5de: avcodec/jpeg2000: support >32bit in ff_jpeg2000_ceildivpow2() [02:57:33 CEST] <cone-389> ffmpeg 03Michael Niedermayer 07master:d62657eb4abb: configure: Also create tests/api [04:57:27 CEST] <cone-389> ffmpeg 03Michael Niedermayer 07master:317be53ca626: avcodec/jpeg2000: Support 32 decomposition levels [11:20:10 CEST] <thardin_> hum-hum.. what do I need to build ffplay again? [11:24:34 CEST] <thardin_> sdl-gfx-dev and/or sdl-image-dev it seems [11:30:43 CEST] <cone-658> ffmpeg 03Andreas Cadhalpun 07master:f9883a669c3d: wavpack: use get_bits_long to read up to 32 bits [11:48:24 CEST] <cone-658> ffmpeg 03Anton Khirnov 07master:d3ea79e8a65d: FATE: drop the last truncated frame from the wma lossless test [11:48:25 CEST] <cone-658> ffmpeg 03Michael Niedermayer 07master:fb2eeebf662e: Merge commit 'd3ea79e8a65ddad4da11813bb43c46701295f68c' [12:07:05 CEST] <cone-658> ffmpeg 03Anton Khirnov 07master:441e8ae5efd6: FATE: drop the last truncated frame from the wmapro tests [12:07:06 CEST] <cone-658> ffmpeg 03Michael Niedermayer 07master:3d2267019cd7: Merge commit '441e8ae5efd681055e5af6f4317fb60110de9dd0' [12:07:07 CEST] <cone-658> ffmpeg 03Michael Niedermayer 07master:0940169743aa: tests/fate/wma: use the existing reference files for wmapro tests [13:05:27 CEST] <cone-658> ffmpeg 03Michael Niedermayer 07master:da0616985095: avformat/riffdec: Forward error code from avio_read() in ff_get_guid() [13:41:49 CEST] <cone-658> ffmpeg 03Alexandra Hájková 07master:b08569a23948: lavf: Replace the ASF demuxer [13:41:50 CEST] <cone-658> ffmpeg 03Michael Niedermayer 07master:30b8f3e7dcd0: Merge commit 'b08569a23948db107e5e6175cd4c695427d5339d' [14:06:10 CEST] <cone-658> ffmpeg 03Luca Barbato 07master:8a26ae5f94e6: mpjpeg: Check stream allocation [14:06:11 CEST] <cone-658> ffmpeg 03Michael Niedermayer 07master:a57ee6cca10d: Merge commit '8a26ae5f94e613dbf7eb8e3c60462e966e409cdc' [14:39:04 CEST] <cone-658> ffmpeg 03Rostislav Pehlivanov 07master:02dbed6e71e7: aacpsy: Add energy spread for each band [14:51:44 CEST] <cone-658> ffmpeg 03Andreas Cadhalpun 07master:9a345802edf7: huffyuvdec: validate image size [15:04:12 CEST] <Daemon404> asfdec_kostya.c [15:06:09 CEST] <wm4> michaelni: the asf merge is just bullshit [15:07:22 CEST] <Daemon404> 'carl says it's better; [15:07:28 CEST] <nevcairiel> if anything there should be a discussion and disclosure of tests performed that yielded that decision, not a dumb "carl said" [15:09:03 CEST] <nevcairiel> I'm sure I could try to be as biased as I wanted if i performed some closed-door test, and find overwhelming evidence that the new one is 100 times better =p [15:09:38 CEST] <Daemon404> oh so youve worked for the broadcast industry? [15:11:01 CEST] <Daemon404> by all reports, ive heard the new asf demuxer fixes many files [15:12:15 CEST] <wm4> carl doesn't agree [15:12:19 CEST] <wm4> (apparently) [15:12:49 CEST] <Daemon404> by accounts i do not mean 'libav' fwiw [15:12:56 CEST] <Daemon404> but users on doom9, and ffms2 bug tracker [15:13:10 CEST] <nevcairiel> he is worse than some of the libav people, he will just dislike and distrust it because it comes from libav [15:17:05 CEST] Action: Daemon404 is childish and laughs at 'PNS' [15:17:55 CEST] <nevcairiel> ... [15:19:20 CEST] Action: Daemon404 throws nevcairiel's ellipsis back at him [15:19:28 CEST] <nevcairiel> speaking about PNS, fancy prints of spectral results are nice and all, but I do wonder how the audible effect is [15:23:53 CEST] <nevcairiel> although i should probably stop caring and just move towards using opus [15:24:16 CEST] <Daemon404> right now opus-in-mp4 not being final is what stops me [15:24:32 CEST] <nevcairiel> i could probably even use mkv [15:24:49 CEST] <nevcairiel> although ts muxing support in ffmpeg would be nice [15:25:53 CEST] <michaelni> which tickets are fixed by using the rewritten asf demuxer ? [15:30:25 CEST] <cone-658> ffmpeg 03Michael Niedermayer 07master:96e73fa64941: avcodec/jpeg2000dec: Merge i_stepsize/2 decoder special case into dequant [16:13:58 CEST] <atomnuker> nevcairiel: there's a big difference at low bitrates (below 100ish kbps) [16:34:35 CEST] <cone-658> ffmpeg 03Paul B Mahol 07master:f76d7d4b01e8: avfilter: add showvolume filter [16:50:45 CEST] <cone-658> ffmpeg 03Rodger Combs 07master:7d0a19757e0b: lavc/adpcm: THP: set approx_nb_samples correctly [17:08:14 CEST] <cone-658> ffmpeg 03Andreas Cadhalpun 07master:2a4700a4f032: wmavoice: limit wmavoice_decode_packet return value to packet size [17:37:14 CEST] <Daemon404> i image C-E yelling out loud [17:37:18 CEST] <Daemon404> "LIBAV CANT MELT STEEL BEAMS" [18:05:20 CEST] <RiCON> there's a lack of concrete samples from both sides of the argument though [18:06:47 CEST] <jamrial> do we really need TWO asf demuxers? [18:07:25 CEST] <wm4> RiCON: I've asked the Libav side, but no reply yet [18:09:00 CEST] <Daemon404> any g2m or mss2 sample should do [18:12:51 CEST] <BBB> hm... [18:12:56 CEST] <BBB> is c-e online? [18:13:43 CEST] <Daemon404> of course not [18:13:54 CEST] <Daemon404> hes one of those 'email only' sort of people [18:13:55 CEST] <Daemon404> usually. [18:14:26 CEST] <Daemon404> i havent read the thread yet, but i assume it's something like this: [18:14:34 CEST] <Daemon404> everyone: we should investigate more, be reasonable [18:14:45 CEST] <Daemon404> carl: LIBAV IS EVIL WE CANT TRUST THEM TO TEST, LIBAV DID 9/11 [18:22:34 CEST] <BBB> well Im not gonna flame any further [18:22:42 CEST] <BBB> but I recall working on our existing asf demuxer at some point [18:23:07 CEST] <BBB> and (michaelni please pay attention) its the most horrific piece of utter shit bullshit spaghetti nonsense code Ive ever had the misfortune of working with [18:23:20 CEST] <BBB> if we were to delete it, I would be extremely happy [18:23:25 CEST] <BBB> and I say that not having looked at the new code [18:24:39 CEST] <wm4> but it must be good because of the man-years of maintenance! [18:34:02 CEST] <durandal_1707> technically mplayer demuxer is better than both of them [18:35:01 CEST] <Daemon404> its gpl isnt it [18:36:03 CEST] <durandal_1707> yea [18:37:56 CEST] <nevcairiel> BBB: thats really one of the major reasons libav decided to re-write it entirely, the old one is horrible code based on years of hackery and reverse engineering, while the new one is a clean implementation based on the spec (which wasn't available for the first one) [18:39:29 CEST] <nevcairiel> I find it funny that carl basically admitted though that he didnt test anything, so the comment in the commit is just a lie [18:40:48 CEST] <Daemon404> people wouldnt lie on the internet, nevcairiel [18:43:50 CEST] <jamrial> he didn't test it yet he lobbied against merging it? [18:43:57 CEST] <nevcairiel> its from libav [18:44:16 CEST] <nevcairiel> and the old one must be good, its been maintained for years already [18:44:20 CEST] <nevcairiel> thats basically the two arguments [18:45:13 CEST] <nevcairiel> i should sit down and sift through a bunch of samples [18:47:41 CEST] <kierank> wm4, BBB: re my latest email - what do we do with API timestamps being different to ffmpeg.c [18:51:16 CEST] <nevcairiel> ffmpeg.c has boat loads of timestamp magic [18:51:30 CEST] <nevcairiel> i dont even know off-hand how many -async and -vsync options there are [18:54:32 CEST] <BBB> kierank: I dont know [18:55:35 CEST] <BBB> kierank: if the goal is to reuse the test code as a mini-ffmpeg.c (not MiNi; just small) between codecs, then its all fine with me [18:55:47 CEST] <kierank> yes [18:55:49 CEST] <BBB> I just found the naming convention kind of weird since theres nothing flach264 specific int he tests [18:56:01 CEST] <BBB> so if thats just a temp state, its all fine [18:56:19 CEST] <kierank> the goal is to make sure that at least from an output perspective the output of ffmpeg.c is the same as the api [18:56:24 CEST] <kierank> AND to test things ffmpeg.c doesnt test [18:56:25 CEST] <BBB> and yes the timestamp magic in ffmpeg.c and even embedded in the api has annoyed me forever [18:56:39 CEST] <wm4> BBB: yeah, the entire idea is that we don't just test a single API user [18:56:43 CEST] <BBB> I think api tests are good, but they still dont totally help us understand wtf is going on in utils.c [18:56:58 CEST] <BBB> (Im not saying you should fix that; the stated goal of api tests is fine) [18:57:07 CEST] <BBB> (Im just expressing a personal frustration :) ) [18:57:24 CEST] <kierank> What I don't want to see as a result of this is ffmpeg.c timestamp hacks going into utils.c [18:57:29 CEST] <kierank> because that's what some people want [18:57:36 CEST] <BBB> muhahahah you just wait :( [18:57:50 CEST] <nevcairiel> i think we're enough to veto that =p [18:57:53 CEST] <BBB> I mean I agree with you, basically [18:58:13 CEST] <nevcairiel> if they want they can shove that into its own api that people can use, but not automatically in the demuxer [18:58:30 CEST] <BBB> but my point was more orthogonal to that, I think Im fine with magic code living in utils.c, if theres particular ways to test that code, have it be documented, easily understood what it does, why it does it, etc. [18:58:40 CEST] <BBB> I think the timestamp code in utils.c is one of these busfactor=1 pieces of code [18:58:48 CEST] <nevcairiel> there is a flag to disable the magic, if you really wanted to [18:59:08 CEST] <wm4> which flag? [18:59:25 CEST] <BBB> wm4: I think he means a hypothetical flag that we could add in the future [18:59:28 CEST] <BBB> wm4: not one that exists today [18:59:34 CEST] <nevcairiel> it exists [19:00:00 CEST] <nevcairiel> AVFormatContext->flags set AVFMT_FLAG_NOFILLIN [19:00:02 CEST] <BBB> ohryl [19:00:16 CEST] <wm4> I wonder what happens if I use that with my code [19:00:20 CEST] <nevcairiel> then it will not run the magic info generation code [19:02:22 CEST] <Daemon404> [17:57] <+nevcairiel> i think we're enough to veto that =p <-- dir listing api says different [19:02:52 CEST] <nevcairiel> i think we might raise more of a stink if it impacts core APIs instead of just adding more noise [19:03:13 CEST] <Daemon404> time shall tell [19:37:41 CEST] <cone-658> ffmpeg 03Michael Niedermayer 07master:ecbd111768c7: avcodec/jpeg2000dec: Remove unused variables [21:32:28 CEST] <cone-658> ffmpeg 03Rostislav Pehlivanov 07master:305859c0d41d: aaccoder: add intensity stereo coding support for the trellis quantizer [21:32:29 CEST] <cone-658> ffmpeg 03Rostislav Pehlivanov 07master:55397b0e76a4: aaccoder: add intensity stereo support to encode_window_bands_info quantizer [23:22:35 CEST] <Jumbt> https://trac.ffmpeg.org/ticket/2439#no1 has any progress been made on this? would be very useful [23:23:45 CEST] <cone-658> ffmpeg 03Michael Niedermayer 07master:4b0a475846a1: avformat/asfdec: Reduce FRAME_HEADER_SIZE to 11 [23:23:46 CEST] <cone-658> ffmpeg 03Michael Niedermayer 07master:a5f617e9ec8e: avformat/asfdec: Print a warning if data is skiped due to less than a frame header being left [23:30:16 CEST] <durandal_1707> Aha, so asf_o is better for that sample [23:37:08 CEST] <cone-658> ffmpeg 03George Boyle 07master:c78b82f4da74: tests/Makefile: Fix missing trailing slash for API tests dir target [23:40:14 CEST] <Compnn> theres some other fun stuff in asf , binary codecs? truespeech maybe [23:48:57 CEST] <anoop_r> well i know its not the place to say but forgive me.with all your support i had succeded in my project [23:49:09 CEST] <anoop_r> if anybody have windows + .net 4.5 and kind heated enough to test my software please [23:49:16 CEST] <anoop_r> http://www.qfs.mobi/f2403835 [23:50:29 CEST] <anoop_r> you can provide feedback to keepitsimplepltd(a)yandex.com also if i violated anything inform me [00:00:00 CEST] --- Mon Jun 29 2015

1 0

ffmpeg.log.20150628
by burek 29 Jun '15

29 Jun '15

[08:44:05 CEST] <lillian> hey everyone [11:46:31 CEST] <nashgul> hi all [11:47:28 CEST] <nashgul> i'm recording with my webcam and the alsa input, but there is a delay of 2 seconds between the 2 inputs, how can i adjust that delay? [11:48:28 CEST] <nashgul> my command is: ffmpeg -f v4l2 -i /dev/video0 -i alsa hw:1,0 -f ogg -q:v 7 -q:a 7 -ar 44100 -ac 2 output.ogg [11:49:07 CEST] <nashgul> really the video is delayed [11:53:13 CEST] <nashgul> :-D, sorry, -itsoffset , forget my question [12:15:51 CEST] <Fjorgynn> yeah [13:39:35 CEST] <RobertNagy> I have an issue where I have two wav files, 0.wav and 1.wav which I can concat using the concat demuxer without issue. However, I want to concat in a browser (aac, mp3, aac). However, I have found no way to encode the segments to either these formats and still be able to concatenate them without distortion. [13:40:35 CEST] <RobertNagy> I have tried to make sure that 0.wav and 1.wav start and end on 1024/960 sample count boundaries. [13:41:04 CEST] <RobertNagy> But still I can't concat them seamlessly, even with cat (aac) or ffmpeg-concat (opus) [13:41:15 CEST] <RobertNagy> haven't even gotten as far as trying in a browser [13:41:26 CEST] <RobertNagy> is this even possible or am I aiming for a lost cause? [19:29:39 CEST] <edoceo> I'm merging two audios, into a overlay in a grid. But when I'm done there is no audio. I can see the audio stream in the media but it's silent [19:30:05 CEST] <edoceo> I'm using the amerge filter: [0:a][1:a]amerge=inputs=2[aout]' -map '[aout]' -ac 2 [19:36:00 CEST] <razzledazzle> Hey guys can I ask a question related to transcoding but not directly related to ffmpeg? [21:55:30 CEST] <edoceo> It seems that the -ss and -itsoffset values on inut are not used by the amix and amerge filters in a filter complex? [21:55:48 CEST] <edoceo> How can I get the audio to seek as well as the video in this grid merge? [22:05:01 CEST] <Mavrik> evenin [22:29:02 CEST] <edoceo> evenin? I'm not finding anything about that in man or google [23:39:30 CEST] <anoop_r> hi [23:40:12 CEST] <anoop_r> with all your support i had succesfully created a transcoder using ffmpeg+c# [23:40:55 CEST] <anoop_r> if anybody have windows + .net 4.5 and kind heated enough to test my software please [23:41:16 CEST] <anoop_r> http://www.qfs.mobi/f2403835 [23:41:26 CEST] <anoop_r> check here [23:42:17 CEST] <anoop_r> if any part of this violating any thing of ffmpeg please inform me i am happy to correct [23:42:54 CEST] <anoop_r> you can provide feedback to keepitsimplepltd(a)yandex.com [00:00:00 CEST] --- Mon Jun 29 2015

1 0

ffmpeg-devel.log.20150627
by burek 28 Jun '15

28 Jun '15

[00:08:39 CEST] <cone-666> ffmpeg 03Luca Barbato 07master:92fdc80cab2a: nvenc: Fix NV12 input [00:08:40 CEST] <cone-666> ffmpeg 03Michael Niedermayer 07master:1221b7df6817: Merge commit '92fdc80cab2acad9f171ba38a08aa89b392bdadd' [00:38:16 CEST] <cehoyos> jamrial: icc defines _GCC_ depending on the gcc version on your system... [00:38:44 CEST] <cehoyos> I would not expect that all icc versions support the builtin asm command: Is there a problem doing a configure check? [00:39:11 CEST] <jamrial> i'm not adding any asm [00:40:57 CEST] <jamrial> i'm adding the usage of a gcc builtin and putting it inside a __GNUC__ >= 5 guard [00:43:51 CEST] <cehoyos> This guard triffers for icc. [00:43:55 CEST] <jamrial> there are other builtins used in different headers, like this same one or attributes.h, that only check for __GNUC__ [00:43:59 CEST] <jamrial> i don't see how this one is any different [00:44:53 CEST] <jamrial> if icc defines __GNUC__, shouldn't it support everything said define entails? [00:44:59 CEST] <cehoyos> The difference - afaict - is that av_builtin_constant_p exists forever, no? [00:45:22 CEST] <cehoyos> It doesn't, that's why some of the _GCC_ chacks also check for ICC [00:46:13 CEST] <jamrial> av_builtin_constant_p if __GNUC__ is defined translates into __builtin_constant_p [00:46:23 CEST] <jamrial> meaning icc uses it [00:47:22 CEST] <cehoyos> What I meant was: The ifdef looks as if this builtin was already supported in gcc 2.95.3. And icc had ten years to also support it. [00:47:46 CEST] <cehoyos> If your builtin is only supported in newest gcc, I find it very unlikely that it also works in older icc. [00:47:56 CEST] <cehoyos> Why is a confiigure check a problem? [00:48:15 CEST] <jamrial> the builtin has supported for several versions now [00:48:22 CEST] <jamrial> and how gcc 2.95.3? [00:48:38 CEST] <cehoyos> __builtin_constant_p does not check for the gcc version, only for gcc. [00:48:46 CEST] <cehoyos> Meaning it is expected to work with gcc 2.95.3 [00:48:57 CEST] <cehoyos> (We had not supported anything older for a very, very long time) [00:50:24 CEST] <iive> i don't think that there are any projects that do. [00:51:07 CEST] <jamrial> sorry, but i fail to see how this is any different to other uses of AV_GCC_VERSION_AT_LEAST in the tree [00:52:03 CEST] <cehoyos> Which uses do you mean? [00:52:21 CEST] <cehoyos> __builtin_constant_p does not use AV_GCC_VERSION_AT_LEAST [00:52:53 CEST] <jamrial> the ones you can find in many headers [00:53:06 CEST] <jamrial> for bultins or otherwise [00:53:18 CEST] <kierank> michaelni: are both fields guaranteed to be finished in decode_postinit in h264? [00:53:42 CEST] <cehoyos> __builtin_ctz is also very old... [00:54:03 CEST] <jamrial> and this one was not added yesterday [00:55:03 CEST] <iive> jamrial: what is the minimum gcc that supports that builtin? [00:55:33 CEST] <iive> also... can you still compiled ffmpeg on gcc 2.95.3? [00:55:36 CEST] <cehoyos> I mean that it exists in gcc for a very long time giving Intel a very long time to also support it. This does not seem to be the case for your new builtin, or do I misunderstand? [00:55:46 CEST] <cehoyos> iive: No, they broke it. [00:56:08 CEST] <iive> they? the evil ones? [00:56:11 CEST] <jamrial> not sure, but i know that 4.8 supports it [00:56:18 CEST] <cehoyos> No, the thieves and liars. [00:56:29 CEST] <iive> yes, the evil ones :) [00:56:59 CEST] <cehoyos> (It is possible that Michael fixed it, he at least tried at one point.) [00:58:02 CEST] <jamrial> again, how's any other AV_GCC_VERSION_AT_LEAST usage different than this one? [00:58:18 CEST] <cehoyos> Which other usage? [00:58:21 CEST] <jamrial> i see checks for gcc 4.6 with it that don't bother checking for icc [00:58:33 CEST] <jamrial> tons, cehoyos. use grep [00:58:34 CEST] <cehoyos> Don't you write in your patch that only 5.1 supports it? I don't understand? [00:58:43 CEST] <cehoyos> Please tell me (only) one. [00:58:55 CEST] <jamrial> AV_GCC_VERSION_AT_LEAST(4,6) in attributes.h [00:59:03 CEST] <cehoyos> Which one? [01:00:13 CEST] <cehoyos> I mean: Which define in attributes.h (there are many iirc)? [01:00:45 CEST] <jamrial> i just said, the one that checks for 4.6 [01:00:53 CEST] <cehoyos> Which define is it? [01:01:17 CEST] <iive> hum, av_const needs 2.6.x [01:01:30 CEST] <cehoyos> This triggers a warning in icc because not all versions support it. [01:01:39 CEST] <jamrial> AV_NOWARN_DEPRECATED, how hard is it to check the file? [01:01:40 CEST] <cehoyos> We silence the warning in configure. [01:01:53 CEST] <cehoyos> This also triggers a warning on compilation. [01:02:09 CEST] <jamrial> and i wrote the patch to check for 5.1 not because older versions don't support it, but because the builtin was broken in older versions [01:02:10 CEST] <cehoyos> When compiling with icc [01:03:02 CEST] <jamrial> i'm completely lost in this discussion [01:03:03 CEST] <cehoyos> Doesn't this support my point? [01:03:36 CEST] <jamrial> i can't say because i don't understand what you're trying to say [01:03:49 CEST] <iive> jamrial: so, constant_p is broken in all gcc before 5.1 ? [01:04:28 CEST] <jamrial> when the second argument is a constant, yeah [01:04:56 CEST] <jamrial> iive: wait no [01:05:06 CEST] <cehoyos> jamrial: icc defines _GCC_ depending on your system compiler, not depending on the actual supported features. [01:05:24 CEST] <cehoyos> If a gcc feature is new, it is unlikely (imo) that icc also supports it. [01:05:29 CEST] <jamrial> constant_p is not broken [01:05:37 CEST] <jamrial> the bzhi bultin is [01:05:40 CEST] <cehoyos> If such a feature is just an attribute, it only triggers a warning and we don't care. [01:05:48 CEST] <jamrial> it's not new [01:05:52 CEST] <cehoyos> If the feature is a command, it may fail compilation. [01:06:02 CEST] <jamrial> it's been available for several versions [01:07:15 CEST] <cehoyos> The problem is that I only have some icc versions here to test (each takes 1G), how can I test if the intrinsic is defined? [01:07:46 CEST] <iive> cehoyos: why icc cares what system gcc is installed? does it use it as front/back end? [01:09:48 CEST] <cehoyos> iive: Sorry, I cannot answer, I am just remembering many issues I dealt with in the past. [01:09:54 CEST] <cehoyos> No, it does not use it as backend. [01:10:56 CEST] <iive> heh... [01:12:19 CEST] <jamrial> the bzhi builtin i'm adding is available since 4.7 [01:12:40 CEST] <cehoyos> jamrial: Do I understand correctly that the intrinsic is only used if the compiler defines __BMI2__? [01:12:43 CEST] <iive> jamrial: if you know that the builtin is broken, put a check and add comment that it is available, but broken of prior compilers [01:12:50 CEST] <jamrial> iive: i did as much [01:12:54 CEST] <cehoyos> Is __BMI2__ docuemented somewhere? [01:13:18 CEST] <iive> what is the issue at hand then? [01:13:32 CEST] <jamrial> iive: i really don't know [01:13:40 CEST] <cehoyos> That icc cannot be detected with __GCC__ < xyz [01:13:41 CEST] <jamrial> cehoyos: gcc's documentation, problably [01:14:22 CEST] <michaelni> kierank, if i understand the question correctly, yes unless the stream is damaged/constructed like if it has only top fields [01:14:39 CEST] <iive> cehoyos: icc uses same builtin syntax and defines gcc (depending on the system gcc). [01:15:22 CEST] <cehoyos> https://software.intel.com/sites/landingpage/IntrinsicsGuide/ [01:15:32 CEST] <kierank> michaelni: basically closed captions are per field and so I need to decide where is the correct place to reset the caption buffer [01:15:33 CEST] <jamrial> this is not an intrinsic, cehoyos [01:15:34 CEST] <cehoyos> I do see BMI2 there [01:15:41 CEST] <jamrial> also, just sent a new version that makes sure icc doesn't use this [01:15:42 CEST] <cehoyos> What is it then? [01:15:54 CEST] <jamrial> a gcc builtin [01:15:59 CEST] <iive> so... it is possible for icc to define gcc version that is supposed to support builtin, that is not actually supported bu icc? [01:16:31 CEST] <cehoyos> iive: That was the case in the past: All the attributes in attributes.h triggered warnings on icc because icc defines __GCC__ [01:17:37 CEST] <iive> how so? if icc doesn't support them, they are most likely fatal errors!? [01:18:25 CEST] <iive> or does it just ignore them? [01:19:28 CEST] <cehoyos> So if icc (either now or in the future) supports the builtin, we don't use it? [01:19:44 CEST] <iive> maybe there should be AV_ICC_VERSION_AT_LEAST :) [01:20:04 CEST] <jamrial> if anyone cares about icc by then, they are free to send a patch [01:21:29 CEST] <iive> could the icc defining _gcc_be related to system headers mismatch? [01:22:28 CEST] <cehoyos> jamrial: I try to test atm, but with gcc 4.9.1, I get "implicit declaration of function __builtin_ia32_bzhi_si" [01:22:44 CEST] <cehoyos> when compiling adpcm.c [01:23:10 CEST] <iive> anyway, i'm off [01:23:13 CEST] <iive> have fun. [01:23:19 CEST] <iive> and keep it simple :) [01:28:39 CEST] <jamrial> cehoyos: with my patch applied as is? [01:29:44 CEST] <jamrial> gcc 4.7 and newer all declare the builtin when you run gcc with -mbmi2 (or related -march values). otherwise you get that error [01:29:56 CEST] <cehoyos> No, removing all checks to test if the builtin is used. [01:30:00 CEST] <jamrial> that's why i guarded it with the __BMI2__ check [01:30:09 CEST] <jamrial> well of course it will fail in that case [01:30:12 CEST] <cehoyos> I didn't know about -mbmi2 [01:33:11 CEST] <cehoyos> icc -march=core-avx2 does not define __BMI2__ so I have no objections about the original patch. [01:33:15 CEST] <jamrial> i still find it pretty absurd that icc arbitrarily defines __GNUC__ while not even providing all doing so entails [01:33:43 CEST] <cehoyos> (And icc -mbmi2 does not work, march=core-avx2 is the suggested alternative in the icc manual) [01:34:06 CEST] <cehoyos> Me too, there are many patches about this issue in the mailing list archive. [01:35:01 CEST] <cehoyos> http://article.gmane.org/gmane.comp.video.ffmpeg.devel/75876/match=icc [01:36:15 CEST] <michaelni> kierank, iam not sure how exactly CC is stored (even less so in relation to MBAFF if its per field) but if its in SEIs then ff_h264_reset_sei() may be an option. It wont be called per field for MBAFF nor if the muxer packs2 fields per packet currently [01:36:39 CEST] <michaelni> though [01:36:49 CEST] <kierank> michaelni: there are CCs per picture [01:37:15 CEST] <kierank> so 1 set in mbaff [01:37:22 CEST] <kierank> and 1 per field in field codec pictures [01:37:29 CEST] <kierank> the latter need to be merged somewhere [01:44:43 CEST] <michaelni> decode_postinit() might work for the merge but not at the very top of the function, output_frame() could be used too (this one is at the output side reordering wise) there are many other possible places, dunno where is best [03:58:28 CEST] <cone-718> ffmpeg 03George Boyle 07master:3f18cb97e5a7: fate/api-tests: Add directory and Makefile for API tests [03:58:28 CEST] <cone-718> ffmpeg 03George Boyle 07master:7728d231a6cb: fate/api-tests: Move api-flac-test to API tests directory. [04:08:09 CEST] <cone-718> ffmpeg 03Philip Langdale 07master:b69bea3ab8f8: avcodec: Add flag for experimental HWAccels and use it for VDPAU/HEVC [11:30:52 CEST] <cone-389> ffmpeg 03Michael Niedermayer 07master:aab4dbe534e2: avcodec/utils: Replace ENOTSUP by AVERROR_PATCHWELCOME [11:33:40 CEST] <wm4> is there even anything that explicitly checks for AVERROR_PATCHWELCOME error codes [11:56:05 CEST] <kierank> Hmmm I need field coded h264 samples [11:56:11 CEST] <kierank> With captions [12:06:09 CEST] <cone-389> ffmpeg 03Michael Niedermayer 07master:8672c0a6436c: avcodec/jpeg2000dec: Limit poc end fields to whats available in jpeg2000_decode_packets() [12:12:47 CEST] <Compnn> sounds pretty evil kierank [12:27:28 CEST] <kierank> Yes very evil [12:48:45 CEST] <cone-389> ffmpeg 03Derek Buitenhuis 07master:f4be604f1c4d: configure: Check for x265_api_get [12:52:14 CEST] <cone-389> ffmpeg 03Derek Buitenhuis 07release/2.7:6ff54eb87bd0: configure: Check for x265_api_get [15:37:20 CEST] <cone-389> ffmpeg 03Rodger Combs 07master:68f00fb40bb3: lavf/brstm: if the file lies about the last block's size, correct it [15:37:21 CEST] <cone-389> ffmpeg 03Michael Niedermayer 07master:5cccbc3b74b2: avcodec/jpeg2000dec: Check for mismatching per expected termination [17:57:38 CEST] <cone-389> ffmpeg 03James Almer 07master:1e51e517be0a: avutil/x86/intmath: use bzhi gcc builtin in av_mod_uintp2() [18:21:09 CEST] <cone-389> ffmpeg 03Derek Buitenhuis 07master:b56b12cd12f3: configure: Check for x265_api_get [18:21:10 CEST] <cone-389> ffmpeg 03Michael Niedermayer 07master:1edc87a665bf: Merge commit 'b56b12cd12f32bf11efff05632600a7619584b98' [18:28:51 CEST] <cone-389> ffmpeg 03Niklesh 07master:cf9051deac49: movtextenc.c: Support for Bold, Italic and Underlined Styles [18:54:04 CEST] <cone-389> ffmpeg 03Anton Khirnov 07master:e49e0f58e273: h264: make sure the slices do not overlap during slice threading [18:54:05 CEST] <cone-389> ffmpeg 03Michael Niedermayer 07master:f620315fa5b1: Merge commit 'e49e0f58e273237f83b4486f3536931ed1943d18' [19:23:22 CEST] <cone-389> ffmpeg 03Andreas Cadhalpun 07master:6b9fdf7f4f07: nutdec: check maxpos in read_sm_data before returning success [19:49:02 CEST] <cone-389> ffmpeg 03James Almer 07master:93e7b7fb3464: avutil/x86/intmath: add missing check for inline assembly [20:15:30 CEST] <LigH> Hi. [20:16:23 CEST] <LigH> Does FourCC tagging for AVI (e.g. "-vtag XVID") only work during conversions, but not during simple re-multiplexing (with "-vcodec copy")? [20:25:41 CEST] <BBB> should be possible at least [20:25:53 CEST] <BBB> what xact commandline? (this might belong on #ffmpeg, not #ffmpeg-devel) [20:26:33 CEST] <LigH> No affirmative answers yet in #ffmpeg [20:26:36 CEST] <LigH> "E:\Programme\MeGUI\tools\ffmpeg\ffmpeg.exe" -y -i "*.m4v" -vcodec copy -vtag XVID -r 25 "*.avi" [20:26:47 CEST] <LigH> As used in MeGUI [20:28:22 CEST] <BBB> Id file a bug on trac, it seems to me it should work [20:29:52 CEST] <cone-389> ffmpeg 03Anton Khirnov 07master:b114f6d48a06: avconv: factor out flushing the filters [20:29:54 CEST] <cone-389> ffmpeg 03Michael Niedermayer 07master:a50b00822784: Merge commit 'b114f6d48a06a4dad6882bc83e07463905f004c4' [20:30:22 CEST] <LigH> OK, will recommend that to Zathor and Kurtnoise [20:30:47 CEST] <LigH> \o [20:41:31 CEST] <cone-389> ffmpeg 03Anton Khirnov 07master:9a5e4fbec870: avconv: do not stop processing the input packet on decoding error [20:41:32 CEST] <cone-389> ffmpeg 03Michael Niedermayer 07master:a5f409bcc091: Merge commit '9a5e4fbec870c7d466b7a0aec92c70778efc96b5' [20:54:00 CEST] <LigH> Hello again; reported https://trac.ffmpeg.org/ticket/4682 [20:54:11 CEST] <LigH> Aha. [20:54:55 CEST] <LigH> I'll try to build a fresh ffmpeg now and test again. [20:55:06 CEST] <LigH> But the one in MeGUI was not too old... [21:11:36 CEST] <cone-389> ffmpeg 03Anton Khirnov 07master:6d4d3fee63c4: h264: make sure the current picture is not made a long ref multiple times [21:11:37 CEST] <cone-389> ffmpeg 03Michael Niedermayer 07master:77412f0ecd23: Merge commit '6d4d3fee63c46d921c4870feab79269af94e84e1' [22:07:37 CEST] <wm4> kierank: so how do you use the closed captions crap? you somehow shuffle the side data coming out of the decoder to somewhere else? [22:07:48 CEST] <kierank> I turn it into analogue pulses (lol) [22:08:01 CEST] <kierank> and put them at the top of the frame [22:08:17 CEST] <wm4> analogue? huh. [22:08:20 CEST] <kierank> yup [22:08:43 CEST] <kierank> http://nootropicdesign.com/projectlab/wp-content/uploads/2011/03/line21.jpg [22:10:05 CEST] <wm4> as part of the broadcast signal? [22:10:24 CEST] <kierank> yes [22:10:26 CEST] <kierank> damn that patch is bugged [22:16:03 CEST] <cone-389> ffmpeg 03Anton Khirnov 07master:5ec0bdf2c524: h264: do not update the context fields copied between threads after finish_setup() [22:16:04 CEST] <cone-389> ffmpeg 03Michael Niedermayer 07master:bd5c6a519e9b: Merge commit '5ec0bdf2c524224f30ba4786f47324970aed4aaa' [22:52:55 CEST] <Compnn> lol line21 :D [22:53:10 CEST] <Compnn> aka those dancing white lines on the top of the vhs signal [22:58:15 CEST] <kierank> is there a preferred way to do overflow checks in ffmpeg? [23:05:20 CEST] <wm4> as long as it's correct, who cares [23:15:22 CEST] <Compnn> kierank : should we make a list of files that are "leading by example" ? [23:15:30 CEST] <Compnn> of programming practices to follow ? [23:17:04 CEST] <kierank> lol [23:48:25 CEST] <wm4> kierank: I guess in general we just write "if(a+b>INT_MAX)", and then either add casts to a wider type (int64_t), or rearrange the expression in a way the overflow check itself can't overflow ("if(b>INT_MAX-a)") [23:53:20 CEST] <cone-389> ffmpeg 03Michael Niedermayer 07master:9ea256814dc2: avcodec/h264_slice: assert that the first slice implies unfinished setup [00:00:00 CEST] --- Sun Jun 28 2015

1 0

ffmpeg.log.20150627
by burek 28 Jun '15

28 Jun '15

[00:00:08 CEST] <llogan> sorry, no. i don't use ffserver. you may get an answer on ffmpeg-user mailing list, but i wouldn't be too hopeful [00:01:48 CEST] <llogan> i assume you already saw the docs and wiki examples? [00:01:55 CEST] <llogan> https://trac.ffmpeg.org/wiki/Streaming%20media%20with%20ffserver [00:07:59 CEST] <Exagone313> yes, i used this example [00:08:07 CEST] <Exagone313> and it does not work for me [00:08:42 CEST] <Exagone313> I don't like the mailing lists, what forum can I use? [00:08:53 CEST] <Exagone313> stackoverflow/other? [00:09:34 CEST] <llogan> no. stack overflow is only for programming questions. you could try super user, or Gman or Nabble for a "forum" interface to the mailing list [00:09:43 CEST] <llogan> dir.gmane.org/gmane.comp.video.ffmpeg.user [00:09:48 CEST] <llogan> http://ffmpeg-users.933282.n4.nabble.com/ [00:10:31 CEST] <llogan> not many ffserver answerers out there [00:12:13 CEST] <Exagone313> ok i'll try this link [00:12:36 CEST] <llogan> ping me when you send the message and i'll approve it in moderation queue [00:13:03 CEST] <Exagone313> no encryption? sorry i don't [00:13:09 CEST] <llogan> or wait 12-24 hours for me to go though the daily rounds [00:13:19 CEST] <Exagone313> or ok i generate a password [00:13:42 CEST] <llogan> what i meant was messages sent from unsubscribed users to the mailing list must be manually approved. this includes messages from nabble. [00:16:00 CEST] <Exagone313> ok i see [00:17:28 CEST] <Exagone313> I add my cfg directly here or I use a paste? [00:17:51 CEST] <llogan> provide all info in the email [00:18:37 CEST] <llogan> make it easy for the readers, unlike your previous paste. [00:18:59 CEST] <llogan> provide the command then the complete output. not command1 command2 command3 output1 output2 output3 [00:19:13 CEST] <Exagone313> I have to subscribe to the mailing list? [00:19:19 CEST] <llogan> no [00:19:55 CEST] <llogan> but then you'll have to keep checking periodically to see if you get a reply (unless nabble tells you somehow; i don't really know what nabble does) [00:20:29 CEST] <Exagone313> i prefer this way [00:20:58 CEST] <Exagone313> i remember a post on launchpad for a bug on ubuntu, without a possibility to unsubscribe to the mailing list so i banned the email [00:21:19 CEST] <llogan> launchpad sucks. [00:22:02 CEST] <Exagone313> and i see lots of mailing list with emails... the bots say thanks to it [00:23:29 CEST] <llogan> i've never recieved any spam from the single use, random, unique accounts i've used to test any potential spam load from sending to the ML [00:24:23 CEST] <Exagone313> just, this command seems good? ffmpeg -i rtmp://127.0.0.1/live/stream http://127.0.0.1:8182/ogg.ffm|http://127.0.0.1:8182/webm.ffm [00:24:43 CEST] <llogan> i don't know [00:24:51 CEST] <Exagone313> ok [00:26:46 CEST] <Exagone313> sound not ... [00:26:47 CEST] <Exagone313> ./ffmpeg.sh: line 2: http://127.0.0.1:8182/webm.ffm: No such file or directory [00:28:03 CEST] <Exagone313> lol it worked for 2 seconds [00:28:05 CEST] <Exagone313> I think [00:30:53 CEST] <pzich> Exagone313: does it work if you add -re ? [00:31:50 CEST] <Exagone313> where? [00:34:12 CEST] <Exagone313> pzich: [00:34:29 CEST] <pzich> `ffmpeg -re -i rtmp://...` [00:34:46 CEST] <Exagone313> with | ? [00:39:18 CEST] <Exagone313> pzich: the | definitely not work so i tried ffmpeg -re -i rtmp://127.0.0.1/live/stream http://127.0.0.1:8182/ogg.ffm -i rtmp://127.0.0.1/live/stream http://127.0.0.1:8182/webm.ffm [00:39:26 CEST] <Exagone313> (with and without the -re) [00:39:57 CEST] <Exagone313> apparently, the -re make it works for 2-3 seconds, mroe often, but it stills fail with conneciton reste by peer [00:45:04 CEST] <Exagone313> llogan: http://ffmpeg-users.933282.n4.nabble.com/Local-RTMP-to-live-stream-http-web… [00:50:46 CEST] <llogan> approved, but you forgot to include the command, console output, and ffserver conf [00:51:32 CEST] <Exagone313> i did [00:52:25 CEST] <Exagone313> you don't see it? [00:52:50 CEST] <llogan> https://ffmpeg.org/pipermail/ffmpeg-user/2015-June/027299.html [00:52:59 CEST] <llogan> that's what my client also sees [00:53:59 CEST] <llogan> i wonder if the commands and outputs were not included in the text/plain component of your message [00:54:34 CEST] <nashgul> this line doesn't streams and doesn't record the audio: http://pastie.org/10260966 [00:55:27 CEST] <llogan> Exagone313: how did you compose the message? [00:55:48 CEST] <Exagone313> can you copy or screen what you see? [00:56:03 CEST] <Exagone313> what you said is in <raw> nodes [00:56:53 CEST] <nashgul> the comand: http://pastie.org/10260966 , and the output: http://pastie.org/10260968 [00:56:54 CEST] <llogan> see the link i provided [00:57:49 CEST] <llogan> nashgul: it's easier if the command and console output are in the same paste. not sure why the trend lately is to separate them [00:58:00 CEST] <nashgul> ok ok [01:00:19 CEST] <nashgul> http://pastie.org/10260977 [01:01:06 CEST] <llogan> anyone else getting popups at the nabble link? first popup ive seen in a very long time [01:03:06 CEST] <Exagone313> I have ad blockers [01:03:12 CEST] <Exagone313> and I did [01:03:22 CEST] <Exagone313> i get one [01:03:27 CEST] <llogan> annoying [01:03:52 CEST] <llogan> nashgul: why are you using -vsync 2? [01:04:23 CEST] <nashgul> llogan: i don't know, ffmpeg tells me i must use it [01:05:44 CEST] <nashgul> the problem is the split and the mapping, if i don't make split the webcam streams audio correctly [01:06:07 CEST] <nashgul> well, the webcam has not audio, i use alsa input [01:06:55 CEST] <llogan> you probably don't need split if you're not performing other filtering [01:07:23 CEST] <nashgul> llogan: when i split correctly i'm going to apply some filters [01:07:41 CEST] <nashgul> add a logo.png, scroll text, etc.. [01:10:07 CEST] <llogan> try the mailing list for that one. [01:14:38 CEST] <Exagone313> llogan: can you make a screen? if you don't see my cmd/out/cfg i repost without the <raw> [01:16:22 CEST] <llogan> Exagone313: the link i provided shows what i see [01:16:45 CEST] <Exagone313> ok i didnt see [01:16:47 CEST] <llogan> you never answered my question about how you make the message [01:16:59 CEST] <Exagone313> ok i reply [01:18:37 CEST] <Exagone313> can you validate my reply? [01:19:09 CEST] <Exagone313> llogan: on your link my email is shown ....... [01:19:22 CEST] <llogan> message approved [01:19:36 CEST] <Exagone313> thanks [01:20:07 CEST] <llogan> your "email" is shown, but it is obfuscated [01:20:20 CEST] <Exagone313> "at" is not obfuscated for me [01:21:14 CEST] <llogan> no bot processes that [01:22:17 CEST] <llogan> if they did i would have received messages to my test accounts [01:22:36 CEST] <llogan> but it's been years and nothing appeared [02:12:20 CEST] <Zamboney> hi, quick question. i wanted to try encoding using nvenc, since ffmpeg supports since version 2.6. however the windows build linked from the download section of ffmpeg.org returns error 'unknown codec nvenc' when i try to use. so the question, is the version on the download site 2.6 or higher and if not, where do you find it [02:17:47 CEST] <Zamboney> or do i need to build it myself to have this functionality [02:19:17 CEST] <c_14> You have to build it yourself. [02:19:21 CEST] <c_14> Because of license incompatibilities. [02:19:43 CEST] <Zamboney> yeah i was looking at the forum on the download site and was getting that impression. ok, thanks [10:04:54 CEST] <nashgul> hi, good morning, what's wrong with this: http://pastie.org/10261420 ? [10:06:36 CEST] <nashgul> i'm trying to shoot the filter every 10 seconds, and when it's launched: store 't' in var '1', and then uses '1' to substract from the X position of the text, to make a right-left scroll [10:07:35 CEST] <nashgul> umm i don't find clearly examples of 'using expressions evaluations' [12:37:43 CEST] <Fjorgynn> Hey I' [17:21:38 CEST] <sine0> why would you need to have a constant bitrate when encoding for example an audio clip [17:21:52 CEST] <sine0> whats wrong with vbr in mp3 [17:23:13 CEST] <RobotsOnDrugs> early hardware mp3 players had issues with vbr [17:23:23 CEST] <sine0> ok [17:23:38 CEST] <RobotsOnDrugs> there is no reason not to use vbr today [17:23:47 CEST] <sine0> ok nice, thats cleared that up for me then [17:24:04 CEST] <sine0> so its more of a legacy option [17:24:52 CEST] <RobotsOnDrugs> it's also useful for streaming if you have limited bandwidth [17:25:06 CEST] <RobotsOnDrugs> but generally, yes [17:36:14 CEST] <sine0> -codec:v none [17:36:17 CEST] <sine0> is that correct syntax [17:37:40 CEST] <sine0> well it works im badass [19:55:22 CEST] <LigH> Hello. [19:56:00 CEST] <LigH> Is enforcing a FourCC while multiplexing to AVI, using parameter -vtag, still supported at all? [19:56:34 CEST] <c_14> Not sure why it shouldn't be, why? [19:58:04 CEST] <LigH> Because after multiplexing from MPEG-4 ASP (*.m4v), created by xvid_encraw, into AVI, the AVI container uses the FourCC "FMP4", even though MeGUI tries to enforce FourCC "XVID". [19:58:26 CEST] <LigH> Sample call: "E:\Programme\MeGUI\tools\ffmpeg\ffmpeg.exe" -y -i "*.m4v" -vcodec copy -vtag XVID -r 25 "*.avi" [20:03:55 CEST] <c_14> Hmm, when you say "still supported" does that mean it used to work? [20:04:47 CEST] <LigH> People keep reporting this solution, as if it worked once. I don't belong to the people who kept using this over years... ;) [20:09:12 CEST] <LigH> It may work while converting, but apparently does not work while multiplexing only (-vcodec copy). [20:26:59 CEST] <c_14> LigH: I tried it out on my system, can't get it to work either. If it used to work on some version, it's a regression. Otherwise it's a feature request. [20:27:08 CEST] <c_14> Either way, you can open a corresponding issue on trac. [20:30:37 CEST] <LigH> Will do so, thanks [20:30:55 CEST] <LigH> \o [22:23:07 CEST] <microcolonel> I'm trying to build a y4m from a sequence of PNGs, but ffmpeg is pretending the input sequence is 25fps even though I specified 60fps [22:23:57 CEST] <microcolonel> ease up fella, I'm on it [22:24:07 CEST] <microcolonel> ffmpeg -f image2 -start_number 20 -i ./rf_final/%04d.png -r 60 -pix_fmt yuv422p rf.y4m [22:24:13 CEST] <microcolonel> is the command [22:24:28 CEST] <c_14> ffmpeg -f image2 -start_number 20 -framerate 60 -i ./rf_final/%04d.png [22:24:41 CEST] <c_14> You specified the output framerate, not the input framerate [22:25:45 CEST] <microcolonel> hmm [22:25:50 CEST] <microcolonel> so r works on output [22:26:00 CEST] <microcolonel> but framerate, when you put it before -i, works on input [22:26:03 CEST] <microcolonel> or are they equivalent? [22:26:20 CEST] <microcolonel> this is deeply confusing. [22:26:39 CEST] <microcolonel> also https://en.wikibooks.org/wiki/FFMPEG_An_Intermediate_Guide/image_sequence this is wrong I guess [22:26:41 CEST] <c_14> r works on both, but if you put it after the input filename it works on the output, if you put it before the input filename it works on the input. -framerate is an option specific to the image2 demuxer. When using the image2 muxer/demuxer always use -framerate [22:26:57 CEST] <c_14> yes, that is very wrong [22:27:07 CEST] <c_14> https://trac.ffmpeg.org/wiki/Create%20a%20video%20slideshow%20from%20images [22:27:46 CEST] <microcolonel> maybe the title should be changed to "create a video slideshow or convert a sequence to a video" [22:27:50 CEST] <microcolonel> I didn't want a slideshow :P [22:28:27 CEST] <microcolonel> Anyway, thanks, c_14 [22:28:53 CEST] <c_14> Well, technically every video is a slideshow. [22:28:55 CEST] <c_14> np [22:30:46 CEST] <techtopia> is there someway i can use -t or -ss to encode to an i-frame [22:31:00 CEST] <c_14> You want to extract an i-frame from a video? [22:31:09 CEST] <techtopia> like if i did -t 60 is there someway to make it encode to the closest i-frame to 60 secons [22:31:10 CEST] <techtopia> no [22:31:32 CEST] <c_14> You want the video to end on an i-frame? [22:31:37 CEST] <techtopia> i want to encode up to an iframe, then later encode from that iframe to another that is another 60 seconds ahead [22:31:40 CEST] <techtopia> yeah [22:32:01 CEST] <techtopia> like the source file is still being written as im encoding [22:32:18 CEST] <techtopia> so can't do it all in one go, and if i don't cut on i-frames i get sync errors [22:32:51 CEST] <c_14> ffmpeg -ss start -t end -no_accurate_seek -i filename out.mkv <- might do it [22:32:58 CEST] <microcolonel> techtopia: you could try pausing the encoder process as you go [22:33:03 CEST] <techtopia> thanks will run a test [22:33:08 CEST] <techtopia> you can pause it? [22:33:13 CEST] <techtopia> how? [22:33:18 CEST] <microcolonel> in your terminal you can press C-z [22:33:25 CEST] <microcolonel> should stop the process [22:33:30 CEST] <techtopia> nice, will try that too :) [22:33:31 CEST] <techtopia> thanks guys [22:33:47 CEST] <microcolonel> enjoy :) [22:56:30 CEST] <toriso> Hi, I'm coding a video, and I'm having a message past duration too large, anyone know what it means? [22:58:15 CEST] <techtopia> can you post a log [22:58:23 CEST] <techtopia> of your encoding line and the ffmpeg output [22:58:26 CEST] <techtopia> pastebin it [23:02:21 CEST] <toriso> techtopia, yes, http://pastebin.com/CgRHGiNE [23:09:08 CEST] <toriso> techtopia: any idea? [23:29:00 CEST] <toriso> I fixed it [00:00:00 CEST] --- Sun Jun 28 2015

1 0

ffmpeg-devel.log.20150626
by burek 27 Jun '15

27 Jun '15

[00:17:07 CEST] <feliwir> is pross_ here already? [01:01:12 CEST] <cone-287> ffmpeg 03Michael Niedermayer 07master:f1e173049ecc: avcodec/jpeg2000: Remove CBLK limit [01:53:34 CEST] <rcombs> \o/ successfully made bsf_h264_mp4toannexb's name somewhat incorrect [05:13:13 CEST] <cone-912> ffmpeg 03Philip Langdale 07master:6e5e139fe370: avcodec/vdpau: Support for VDPAU accelerated HEVC decoding [11:21:01 CEST] <ubitux> WARNING: Option --enable-hwaccel=vdpau did not match anything [11:21:03 CEST] <ubitux> curious. [11:22:50 CEST] <ubitux> it seems it can only be auto-probed with --enable-hwaccels [11:23:04 CEST] <ubitux> if this is the desired behaviour, there is a bug in the "documentation" of the configure [11:23:11 CEST] <ubitux> but i have doubt that's expected [11:24:45 CEST] <nevcairiel> Its vdpau_h264 etc [11:25:01 CEST] <nevcairiel> Codec specific names [11:25:19 CEST] <nevcairiel> The general option is just --enable-vdpau [11:27:07 CEST] <ubitux> ah? mmh [11:37:14 CEST] <cehoyos> Hi! What is "-vsync drop" supposed to do? [11:37:52 CEST] <BtbN> I think it syncs the video by dropping frames. [11:38:01 CEST] <cehoyos> A user uploaded a hd cellphone recording that contains audio "holes". FFplay, MPlayer and WMP speed up video to keep A/V sync, QT fails to play the sample without desync. [11:38:17 CEST] <cehoyos> ffmpeg -async 1 -i ... inserts silent audio and keeps A/V sync. [11:38:35 CEST] <cehoyos> How is possible to drop frames to keep A/V sync? -vsync drop makes no difference here... [11:38:44 CEST] <cehoyos> (And I wonder what it does / how it works.) [11:39:05 CEST] <cehoyos> http://avmule.com/ffmpeg/source2.mp4 [11:39:17 CEST] <cehoyos> The most prominent drop is after 77 seconds, but there are several. [11:39:28 CEST] <BtbN> ah, drop does not drop frames. It drops timestamps from the frames, so the muxer generates fresh ones [11:39:42 CEST] <cehoyos> Ok, thank you, that explains it. [11:40:02 CEST] <cehoyos> So is it a feature request to drop frames if input audio contains holes to keep A/v sync? [11:40:16 CEST] <cehoyos> michaelni? [11:40:26 CEST] <BtbN> I can't imediately see a way to do that, so i'd say it is [11:41:14 CEST] <BtbN> cehoyos, try with vfr [11:41:17 CEST] <BtbN> -vsync vfr [11:41:57 CEST] <BtbN> But the default, auto, should do something similar already [11:42:15 CEST] <BtbN> I'm not sure if it means dropping timestamps or actual frames though. [11:42:28 CEST] <cehoyos> Well, if the default would not lead to A/v desync, there would be no issue... [11:42:56 CEST] <BtbN> Propably nobody every thought of actual holes in the audio stream [11:43:11 CEST] <cehoyos> I don't see a difference in the output file between -vsync vfr and -vsync drop and default [11:43:26 CEST] <cehoyos> That's why -async 1 exists, so the issue is known afaict. [11:45:31 CEST] <BtbN> My guess is that with actual holes in the audio stream, it never knows the video is out of sync [11:45:39 CEST] <BtbN> Because there are no timestamps to compare to [11:46:22 CEST] <cehoyos> It knows that the audio is missing (if not, -async 1 wouldn't help I guess): The FFplay status line shows the issue visibly. [11:46:50 CEST] <cehoyos> And ffplay has to do some visibly action to keep A/V sync: speed up video. [11:47:23 CEST] <BtbN> iirc -async 1 only affects the beginning of the stream, and does nothing if something happens in the middle [11:47:42 CEST] <cehoyos> That is not correct. [11:48:13 CEST] <cehoyos> But please feel free to test yourself, link is above. [11:48:24 CEST] <BtbN> "-async 1 is a special case where only the start of the audio stream is corrected without any later correction." At least that's what the doc says [11:48:40 CEST] <BtbN> i'm at work currently and can't listen to audio, so testing is kind of complicated [11:49:55 CEST] <cehoyos> docs can be wrong... [11:50:30 CEST] <cehoyos> (The reason is that -async one now maps to an audio filter that does not know anything about "beginning of streams". The doc may have been correct once upon a time) [11:53:11 CEST] <cehoyos> No, it seems the doc were never right, -async 1 also works for the sample before the audio filters were written. [12:17:19 CEST] <michaelni> cehoyos, i think its a feature request to drop the video during these holes of a different stream [12:25:32 CEST] <cone-666> ffmpeg 03Michael Niedermayer 07master:6c4a2f11ddde: avcodec/jpeg2000dec: Add coords to Jpeg2000Tile [12:25:33 CEST] <cone-666> ffmpeg 03Michael Niedermayer 07master:50b77e364f0a: avcodec/jpeg2000dec: iterate over positions with the special cases from jpeg2000 [12:25:34 CEST] <cone-666> ffmpeg 03Michael Niedermayer 07master:f5822ea3798a: avcodec/jpeg2000dec: Add missing \n to av_log() [12:37:40 CEST] <cehoyos> michaelni: Thank you, ticket created! [13:00:26 CEST] <cehoyos> michaelni: Concerning j2k, this may be the only remaining issue that users have reported before ami_stuff started to stress-test the decoder: http://samples.ffmpeg.org/ffmpeg-bugs/trac/ticket2586/ [13:32:56 CEST] <ashish_3805> hello, I am new to open-source. I am second year engineering student in india. I know c++, c and webstuffs . I want to get my hands in development by contributing to ffmpeg.. please help [13:34:12 CEST] <JEEBsv> start by building current git HEAD FFmpeg [13:34:28 CEST] <JEEBsv> getting used to what is used for configuration and building steps [13:34:35 CEST] <JEEBsv> then, well, find something that interests you [13:36:25 CEST] <Compn> ashish_3805 : there are many things to work on in ffmpeg :) [13:36:31 CEST] <Compn> we always need help [13:37:32 CEST] <cone-666> ffmpeg 03Michael Niedermayer 07master:29cc0a178e75: ffmpeg_opt: Fix sync_ist [13:37:53 CEST] <ashish_3805> thanks for ur help... [13:42:22 CEST] <cehoyos> Compn: I don't understand your comment in ticket 4661... [13:48:55 CEST] <Compn> cehoyos : which part is confusing ?> [13:49:45 CEST] <cehoyos> There is only one sentence, and I don't think it is understandable... [13:51:01 CEST] <Compn> feel free to edit/delete or make your own comment then ? [13:51:33 CEST] <cehoyos> Deleting is not welcome iirc... [13:51:50 CEST] <cehoyos> My comment would just be "wontfix" but I was hoping somebody else would close the ticket. [13:52:00 CEST] <Compn> why wontfix ? [13:52:21 CEST] <Compn> what if there are files that exist with some problem that work in fcp ? [13:52:32 CEST] <Compn> er some problem in ffmpeg, but those same files work in fcp ? [13:52:39 CEST] <J_Darnley> While it isn't the best sentence ever written, read it as: we need a sample, or a way to produce prores 422 with v210 colorspace, so we can fix this. [13:53:00 CEST] <kierank> cehoyos: it's not really a bug [13:53:06 CEST] <kierank> as such [13:54:26 CEST] <cehoyos> J_Darnley: What sample do we need? [13:54:43 CEST] <cehoyos> kierank: I did not claim it is a bug (I claimed it is an invalid ticket) [13:54:54 CEST] <kierank> I know [13:54:56 CEST] <Compn> i merely asked for more info .... [13:54:59 CEST] <kierank> I'm agreeing with you cehoyos [13:55:23 CEST] <Compn> i dont know the validity of the bug/request, nor do i care [13:55:59 CEST] <Compn> i do care about fostering a helpful attitude on the bug tracker, so that users reporting bugs arent scared away. [13:56:34 CEST] <Compn> because i dont mind using the bug tracker as user help [13:56:36 CEST] <cehoyos> I have a feeling that this is not a "user" in your sense [13:56:52 CEST] <Compn> but thats just my personal opinion and not speaking for anyone else or the project. [13:57:12 CEST] <cehoyos> But there are hundreds of tickets with status new that are probably valid that you could try to reproduce and fix the status... [13:57:18 CEST] <Compn> are you saying its someone who wants to create invalid prores files with ffmpeg ? [13:57:38 CEST] <cehoyos> (And I do know that what you write here is you personal opinion, at least I hope so ;-) ) [13:58:27 CEST] <cehoyos> What does the word "invalid" have to do with these two tickets? (And, while being there "fcp"?) [13:59:16 CEST] <Compn> fcp = final cut pro, one of the main tools used in dealing with prores files , i thought [13:59:30 CEST] <cehoyos> The user has a slightly unusual way of implementing his use-case and wants to avoid using FFmpeg decoders. I don't know why and I am not sure if we want to know it. [13:59:48 CEST] <cehoyos> I do know what "fcp" stands for, I don't understand how it is related to the ticket(s) in question. [14:00:32 CEST] <Compn> see, i thought he had a file that works in fcp but not ffmpeg [14:00:50 CEST] <cehoyos> There is no (not even remotely) indication for this in the ticket. [14:01:06 CEST] <cehoyos> But as I said: There are hundreds of tickets that need somebody to reproduce them... [14:01:53 CEST] <Compn> so encourage the devs who come in this channel to take up that job , like ashish_3805 earlier asked for what he should work on [14:02:17 CEST] <cone-666> ffmpeg 03Michael Niedermayer 07master:7ca0cd583125: avcodec/jpeg2000dec: iterate in tile sample space for CPRL & RPCL [14:02:31 CEST] <cehoyos> Unfortunately, you need some experience to reproduce tickets;-( [14:02:45 CEST] <Compn> :) [14:03:11 CEST] <cehoyos> (Or to say it differently: If you fail to undertand a relatively trivial report, how should a user with zero experience succeed?) [14:03:35 CEST] <Compn> in essence, that is what you are saying i am doing, failing to understand a relatively trivial report :) [14:03:44 CEST] <Compn> ehe [14:04:17 CEST] <Compn> er i'm reading things wrong today :P [14:04:23 CEST] <cehoyos> michaelni: Is "Fix sync_ist" related to my earlier question? [14:04:39 CEST] <Compn> maybe new user will have better reading comprehension [14:05:13 CEST] <cehoyos> compn: No, my original questoin was "Why do you request a sample that can be produced with "ffmpeg -f lavi -i testsrc -vcodec v210 out.mov". Only when you wrote that you thought that he has a sample that doesn't play, I realized that there is a misunderstanding. [14:05:18 CEST] <Compn> cehoyos : want to make a new custom query for this page , https://trac.ffmpeg.org/report , that lists only "new" tickets ? [14:06:05 CEST] <cehoyos> No, I don't really like the reports... [14:06:12 CEST] <Compn> or "new" tickets without comments. [14:06:12 CEST] <Compn> doh [14:06:18 CEST] <michaelni> cehoyos, i looked at sync ist code because of your question and saw the issue, i was wondering if it could be implemented using sync_ist ... [14:06:41 CEST] <cehoyos> (I prefer tantalizing the database) [14:06:49 CEST] <cehoyos> Thank you! [14:07:04 CEST] <cehoyos> And is it? [14:07:05 CEST] <Compn> cehoyos : so is that user using libavformat with quicktime binary codec ? [14:07:46 CEST] <Compn> also, i have to go work now. cya [14:07:48 CEST] <cehoyos> No, he wants us to move a decoder to libswscale which has (among other things) the issue that no all encoders follow the spec, so the decoder already contains some hacks that are probably not easy to port to libswscale. [14:08:21 CEST] <cehoyos> Or to say it differently: He wants us to add packed yuv422p10 [14:09:05 CEST] <cehoyos> But there are so many feature requests for which no solution exists currently (he only wants another solution). [14:09:22 CEST] <cehoyos> ... another solution than the one already in FFmpeg [14:09:34 CEST] <Compn> i understand [14:09:44 CEST] <Compn> i just didnt get that upon first reading the bug [14:09:50 CEST] <cehoyos> And packed yuv444p16 [14:10:27 CEST] <Compn> instead of a wontfix , i'd put 'requires bounty' :P [14:10:32 CEST] <cehoyos> Given that several people (developers) asked here repeatedly "why are there so many colour spaces?", we should probably not add new ones. [14:10:41 CEST] <cehoyos> I disagree. [14:15:21 CEST] <durandal_1707> there are so many pixel formats because there are different usecase [14:18:35 CEST] <cehoyos> durandal_1707: I thought there are so many pixel formats because hardware supports so many... [14:18:47 CEST] <cehoyos> Both as input and output pix_fmts. [14:19:05 CEST] <cehoyos> And then came the formats with 8<bpp<=16 [14:19:27 CEST] <cehoyos> And somebody explained that x264 would be slower if we didn't have yuv420p10 [14:22:17 CEST] <wm4> a 4:4:4 packed format would actually be a good idea [14:22:30 CEST] <wm4> because it's such a simple format [14:22:48 CEST] <durandal_1707> Y416 is defined by MSDN and used in hardware [14:25:44 CEST] <wm4> V210 on the other hand is rather complicated, so it's rightfully rejected [14:27:51 CEST] <durandal_1707> tiff works with packed yuvs and current code is big hack [14:28:29 CEST] <cone-666> ffmpeg 03Zhang Rui 07master:d38bc6361df6: avutil/log: modify AV_LOG_MAX_OFFSET for AV_LOG_TRACE [14:38:07 CEST] <cehoyos> wm4: Did you undertand that he wants an output-only colour-space? (I had originally not) [14:38:25 CEST] <cone-666> ffmpeg 03schenk michael 07master:b9161ef05231: avformat/hls: do not iterate to next sequence number if interruption is requested [14:49:15 CEST] <cone-666> ffmpeg 03Peter Ross 07master:ea8fec20573a: fate: test ea vp6 with alpha stream [16:01:01 CEST] <wm4> philipl: one of my users tried it, result: https://0x0.st/6h.jpg [16:01:40 CEST] <wm4> philipl: with 349.16 on a 960 [16:02:34 CEST] <Daemon404> nice watermark [16:03:36 CEST] <wm4> so you can feel like it's 2005 when it's 2015 [16:04:27 CEST] <Daemon404> in 2005 we were already on xvid [16:33:37 CEST] <philipl> wm4: Is that meant to be correct or incorrect? [16:33:54 CEST] <wm4> philipl: I think this interlacing effect is incorrect [16:34:04 CEST] <wm4> didn't you have similar issues? [16:34:08 CEST] <philipl> Yes. it's incorrect. [16:34:12 CEST] <philipl> This is what I'd expect to see. [16:34:21 CEST] <philipl> I did say it wasn't usable. [16:34:35 CEST] <philipl> Yeah. that's the nvidia bug [16:34:43 CEST] <wm4> ok... then why was it applied to master? anything that supports the new ffmpeg/vdpau API will use it by default [16:35:27 CEST] <philipl> Such as? I thought it would only opt-in. [16:36:08 CEST] <wm4> vlc and mpv [16:36:27 CEST] <wm4> (I'm not sure if vlc enables it by default - the player still can whitelist or blacklist on its own) [16:38:32 CEST] <philipl> Sorry. I did all my testing with the command line and mplayer, both of which are opt-in. [16:38:43 CEST] <philipl> And mplayer needs separate code changes to even use it. [16:38:50 CEST] <philipl> Didn't think it would automatically trigger anywhere. [16:39:21 CEST] <philipl> Is there a way to de-propritize it in the codec metadata? [16:43:59 CEST] <wm4> the "new" way of enabling hw decoding is by returning the vdpau pixel format in get_format unconditionally, and using av_vdpau_bind_context() [16:44:20 CEST] <wm4> the API user doesn't even need to know about any codecs [16:44:40 CEST] <wm4> so when AVHWAccel.init doesn't return an error, it'll be used [16:45:24 CEST] <wm4> I suppose we could add a flag for experimental codecs or so, and let default init fail if it's not set [16:45:42 CEST] <philipl> There are hwaccel capabilities but no flags are defined today. [16:46:05 CEST] <philipl> But I could add an experimental capability and repeat the equivalent logic from the codec level check [16:46:15 CEST] <wm4> av_vdpau_bind_context() takes flags which affect decoding in various ways [16:47:06 CEST] <wm4> e.g. AV_HWACCEL_FLAG_ALLOW_HIGH_DEPTH lies pretty much on the same code path as you'd need [16:47:16 CEST] <philipl> K [16:47:37 CEST] <wm4> (don't know if this is the best solution) [16:47:58 CEST] <philipl> Well, it's the best solution that doesn't involve me explicitly disabling it in some way. [16:48:59 CEST] <wm4> hm maybe we could actually use the global experimental codec flag? [16:49:08 CEST] <wm4> something about -strict [16:49:28 CEST] <philipl> Yes. That's what I started looking at. [16:49:38 CEST] <philipl> We can check that in bind_context [16:49:43 CEST] <wm4> AVCodecContext.strict_std_compliance and FF_COMPLIANCE_EXPERIMENTAL [16:52:20 CEST] <rcombs> https://ffmpeg.org/pipermail/ffmpeg-devel/2015-June/174642.html <-- who knows HEVC better than me and wants to poke at this [17:10:31 CEST] <BBB> rcombs: Daemon404 knows it pretty well, he might be able to [17:14:50 CEST] Action: rcombs turns to face Daemon404 [17:15:14 CEST] <rcombs> Daemon404: lemme guess: you don't like the pile of repetitive conditionals leading up to the alloc_and_copy calls? [17:15:49 CEST] <Daemon404> no i dont like the bsd's impl in the first place [17:15:53 CEST] <Daemon404> pre-hevc [17:16:05 CEST] <Daemon404> im not entirely sure about mp4's hevc crap. [17:16:17 CEST] <rcombs> I was trying to guess what you don't like about the existing implementation [17:17:15 CEST] <Daemon404> the insertion code [17:17:24 CEST] <Daemon404> trying to fudge it [17:33:22 CEST] <philipl> Eh, this is harder than it looks. The AVHWAccel isn't guaranteed to be set in vdpau_bind_context [17:33:40 CEST] <philipl> At least it's not set in the ffmpeg cli tool. [17:36:32 CEST] <durandal_1707> michaelni: where is midterm evaluation reached? [17:37:32 CEST] <michaelni> durandal_1707, it seems the page is not accessible before 16UTC [17:38:30 CEST] <wm4> sigh, all the time users complain that seeking in mpeg-ts or similar shit is broken [17:38:54 CEST] <Daemon404> because mpeg-ts isnt meant to seek [17:39:09 CEST] <nevcairiel> it usually works reasonably well though [17:39:10 CEST] <Daemon404> certainly not 188byte [17:39:14 CEST] <Daemon404> with no index [17:45:23 CEST] <philipl> wm4: review sent [19:13:10 CEST] <cone-666> ffmpeg 03Andreas Cadhalpun 07master:151dbe457960: mpegaudiodec: copy AVFloatDSPContext from first context to all contexts [19:20:51 CEST] <cone-666> ffmpeg 03Andreas Cadhalpun 07master:1f1e0a2971b2: vc1dec: use get_bits_long and limit the read bits to 32 [20:05:28 CEST] <[-T-]> t [20:25:05 CEST] <llogan> ding ding ding. we have another FFmpeg version SVN-r23418 user [20:25:42 CEST] <llogan> not a record though [20:40:56 CEST] <kierank> lglinskih: what have you been up to? [20:51:07 CEST] <lglinskih> kierank: I've sent few versions of my patch with H264 test, know I'm preparing fix of my FLAC test. [20:51:20 CEST] <kierank> ok [20:51:29 CEST] <kierank> what is the next thing you should work on? [20:51:36 CEST] <kierank> any preference? [20:51:40 CEST] <kierank> wm4, any ideas? [20:52:04 CEST] <wm4> for API tests? [20:52:06 CEST] <kierank> yes [20:52:25 CEST] <kierank> I guess it should be lossy audio [20:52:39 CEST] <wm4> here's a touch one: check whether seeking + decoding outputs the same results when going back and forth more than once [20:52:42 CEST] <wm4> *tough [20:53:00 CEST] <wm4> since ffmpeg.c apparently seeks only once, this almost never gets tested [20:53:28 CEST] <kierank> I wanted draw_horiz_band tested but that's rather niche [21:02:23 CEST] <cone-666> ffmpeg 03Michael Niedermayer 07master:2ec0ba1e2203: avcodec/jpeg2000dec: Parse POCs [21:02:24 CEST] <cone-666> ffmpeg 03Michael Niedermayer 07master:c72a83193118: avcodec/jpeg2000dec: Support progression order changes [21:05:32 CEST] <rcombs> https://bugs.webkit.org/show_bug.cgi?id=146346 0.o an Apple employee just CC'd himself on my BPG ticket in WebKit [21:12:57 CEST] <Daemon404> rcombs, be ashamed [21:13:02 CEST] <Daemon404> for opening that ticket [21:14:33 CEST] <rcombs> elaborate [21:21:34 CEST] <cehoyos> michaelni: You also enabled p0_03.j2k from the reference files with your latest j2k patch, it looks quite broken though... [21:26:47 CEST] <lglinskih_> kierank: Explain me, please, what is draw_horiz_band used for? [21:27:57 CEST] <wm4> good question [21:27:59 CEST] <wm4> MPlayer uses it [21:28:19 CEST] <Daemon404> its so you can draw a picture as it decodes [21:28:24 CEST] <Daemon404> in raster order [21:28:26 CEST] <Daemon404> i believe. [21:28:32 CEST] <Daemon404> i.e. you get one band at a time [21:28:34 CEST] <Daemon404> rather than one pic [21:28:43 CEST] <Daemon404> (i think) [21:29:10 CEST] <nevcairiel> does anyone really still use that [21:29:12 CEST] <wm4> I bet 100 flamewars on it doesn't work when using multithreaded [21:29:15 CEST] <nevcairiel> it seems like a silly thing [21:31:16 CEST] <Daemon404> i dont know why you would use it [21:31:24 CEST] <Daemon404> seems like it would end up with more overhead. [21:32:20 CEST] <kierank> If you need to process the data for hardware (e.g. v210 while it is still in cache). Also for lower latency [21:33:08 CEST] <Daemon404> it depends how you define latency [21:33:13 CEST] <Daemon404> it may actually be slower to get the entire pic [21:41:58 CEST] <cone-666> ffmpeg 03James Almer 07release/2.7:459090181fe5: library.mak: Workaround SDL redefining main and breaking fate tests on mingw [21:41:59 CEST] <cone-666> ffmpeg 03James Almer 07release/2.7:aebb9410c546: swscale/x86/rgb2rgb_template: add missing xmm clobbers [21:42:00 CEST] <cone-666> ffmpeg 03James Almer 07release/2.7:7f2ab5e50ff7: swscale/x86/rgb2rgb_template: don't call emms on sse2/avx functions [21:44:11 CEST] <cone-666> ffmpeg 03James Almer 07master:5abd4a932337: libvpx: disable unused function prototypes [22:08:05 CEST] <cone-666> ffmpeg 03Michael Niedermayer 07master:c56ba5c27064: avcodec/jpeg2000dec: Print what is found in place of EPH if EPH is not found [22:08:06 CEST] <cone-666> ffmpeg 03Michael Niedermayer 07master:b75c0a72ed3b: avcodec/jpeg2000dec: Fix tp_index for POC [22:10:14 CEST] <lglinskih_> oh, thanks, it became more clear [22:10:53 CEST] <lglinskih_> kierank: I think I'll start with seek test =) [22:23:25 CEST] <philipl> Question for everyone: Regarding my last patch - would hwaccel capability flags be considered part of the public API? Do I need a version bump to define one (the first one)? [22:26:39 CEST] <wm4> they're definitely public API [22:26:51 CEST] <wm4> the field is above the "No fields below this line are part of the public API." comment [22:27:09 CEST] <wm4> a micro bump is probably fine [22:27:46 CEST] <philipl> Ok. That's what I suspected. [22:28:38 CEST] <philipl> wm4: I'll change it to micro and then push. [22:28:45 CEST] <philipl> Thanks [22:28:58 CEST] <wm4> I wonder if this stuff is documented somewhere [22:29:03 CEST] <wm4> but if so, then nobody can find it [22:30:32 CEST] <philipl> Indeed. [23:34:59 CEST] <cone-666> ffmpeg 03Kieran Kunhya 07master:22291c372fa7: avcodec: Add support for per-frame AFD output in h264 [23:36:19 CEST] <cone-666> ffmpeg 03Andreas Cadhalpun 07master:04dfbc9441be: s302m: fix arithmetic exception [23:37:30 CEST] <jamrial> whoa, aacenc patchset [23:37:45 CEST] <jamrial> did the epic ticket finally come to an end? [23:39:50 CEST] <JEEBsv> wow [23:40:05 CEST] <atomnuker> I'm afraid not, that'll happen when klaussfreire sends his patches through [23:44:40 CEST] <jamrial> ah ok [23:44:44 CEST] <jamrial> glad to see some of the work finally making it to the tree [23:56:42 CEST] <cone-666> ffmpeg 03Andreas Cadhalpun 07master:072756cdd2f9: vc1dec: use get_bits_long and limit the read bits to 32 [23:56:43 CEST] <cone-666> ffmpeg 03Michael Niedermayer 07master:a4d76faf45cf: Merge commit '072756cdd2f949462520041e357f52f15d8c274d' [00:00:00 CEST] --- Sat Jun 27 2015

1 0

ffmpeg.log.20150626
by burek 27 Jun '15

27 Jun '15

[00:16:47 CEST] <feliwir> hey, what could be the reason for platform specifc bugs? [00:26:34 CEST] <feliwir> avformat_open_input fails on windows, but not on linux with the exact file with exact same version [00:29:42 CEST] <feliwir> how is that possible? [00:41:22 CEST] <feliwir> someone has latest ffmpeg (from github) on windows who's willing to test sth for me [00:48:33 CEST] <Exagone313> Hello, How to use multiple output? I tried "first_output|second_output" as described here: https://trac.ffmpeg.org/wiki/Creating%20multiple%20outputs but it does not work, it try one output [00:49:33 CEST] <Exagone313> ffmpeg -i rtmp://127.0.0.1/live/stream "http://127.0.0.1:8182/ogg.ffm|http://127.0.0.1:8182/webm.ffm" [00:50:28 CEST] <Exagone313> ok, it works without the quotes [02:35:40 CEST] <DHE> Exagone313: it says you must use "-f tee" as the output format [07:49:38 CEST] <gurpartap1> I want to log pts_time along with clock time respective to that frame and also segment name for that part..I am making segments of 20 seconds. [07:50:11 CEST] <gurpartap1> Here is commandline : ffmpeg -rtsp_transport tcp -re -i rtsp://192.168.0.127 -itsoffset 00:00:1.2 -i stream.sdp -async 1 -tune zerolatency -v verbose -vstats_file log27.txt -vcodec libx264 -acodec aac -strict -2 -s 600x480 -t 00:01:00 -f segment -segment_time 20 -segment_format mp4 -y "live_capture_%03d.avi" [07:56:10 CEST] <gurpartap1> Here is code snippet from ffmpeg.c http://pastebin.com/ZVgZ8Dcf [07:59:24 CEST] <gurpartap1> This is progress till now http://pastebin.com/Na1SBHgU [09:46:41 CEST] <feliwir> someone has the latest ffmpeg version (master branch) on windows? [10:35:57 CEST] <feliwir> noone? [10:37:46 CEST] <BtbN> https://github.com/FFmpeg/FFmpeg/archive/master.zip [11:29:52 CEST] <gurpartap1> BtbN: I am using segment filter to create chunks of 10 seconds.But it is creating chunks of wrong size? [11:30:36 CEST] <gurpartap1> BtbN: Here is commandline: ffmpeg -i rtsp://192.168.0.127 -map 0 -an -v "debug" -flags -global_header -vcodec libx264 -f segment -segment_time 10 out%03d.mp4 [12:24:24 CEST] <Abhijit> how can i make sure that a given mp4 video has its metadata at the begining and not at the end? [12:40:32 CEST] <hrw> is there a way to add language tags to audio streams? [12:44:12 CEST] <chungy> -metadata:s:a:0 language=eng [12:45:34 CEST] <hrw> chungy: thanks [12:46:10 CEST] <circ-user-NSwOK_> /msg nickserv help register [12:56:03 CEST] <Abhijit> circ-user-NSwOK_, without space before "/" [12:56:25 CEST] <Abhijit> circ-user-NSwOK_, and do it in server info tab. not here in any channel [12:56:54 CEST] <circ-user-NSwOK_> thanx@abhijit [13:20:17 CEST] <ashish_3805> hello, I am new to open-source. I am second year engineering student in india. I know c++, c and webstuffs . I want to get my hands in development by contributing to ffmpeg.. please help [13:30:24 CEST] <durandal_1707> ashish_3805 : better ask on -devel channel and mailing list [13:32:01 CEST] <ashish_3805> thanxx [13:43:41 CEST] <Abhijit> i created 8 seconds segments of a given mp4 video. how can i make html video tag to play all of them a single video? [14:56:34 CEST] <feliwir> can i set the looping behaviour of ffplay somehow? it should not try to seek but rather restart [14:56:44 CEST] <feliwir> because the file is not seekable [14:58:40 CEST] <sine0> in windows if i add a path variable i could use ffmpeg from any directory using cmd correct ? [15:04:37 CEST] <durandal_1707> feliwir: what container? [15:05:23 CEST] <feliwir> durandal_1707: ea [15:06:30 CEST] <durandal_1707> you could try -loop for ffplay [15:06:43 CEST] <feliwir> durandal_1707: i used that [15:06:48 CEST] <feliwir> it tried to seek and fails [15:07:41 CEST] <feliwir> i get this in console: ./data/movies/LoadingRing.vp6: error while seeking [15:07:48 CEST] <feliwir> (after it played once succesfully) [15:08:03 CEST] <durandal_1707> than either implement seeking in ea or do not use ffplay [15:08:41 CEST] <feliwir> durandal_1707: this file isn't meant to be seeked. No player can seek it [15:08:48 CEST] <feliwir> but i know that it is loopable [15:09:50 CEST] <sine0> if I am going to do a batch convert of 8 files from m4a to wav, do i have to create a variable [15:12:53 CEST] <sine0> ffmpeg -i *.m4a *.wav [15:13:09 CEST] <sine0> (conserve filenames) [15:13:36 CEST] <sine0> has to be a $.wav or something! [15:17:19 CEST] <DelphiWorld> yo all [15:17:30 CEST] <DelphiWorld> i added a cover art to a mp3 file to be uploaded to youtube [15:17:53 CEST] <DelphiWorld> ffmpeg -i mp3file.mp3 -i cover.jpg -loop 0 -acodec copy out.mp4 [15:17:58 CEST] <DelphiWorld> but the player dont play it [15:19:38 CEST] <c_14> you probably want ffmpeg -i mp3file.mp3 -loop 1 -i cover.jpg -shortest -c:a copy out.mp4 [15:20:10 CEST] <DelphiWorld> c_14: ok, i'lle try [15:20:24 CEST] <DelphiWorld> how to add ffmpeg to my path on windows... [15:20:33 CEST] <DelphiWorld> i dont see where i could add to the path [15:21:00 CEST] <c_14> You have to go to the system settings dialog and then click your way through n+1 different dialogs until you get to the PATH setting. Can't remember exactly where it is though. [15:21:06 CEST] <c_14> That or put the binary somewhere already in your PATH [15:22:56 CEST] <DelphiWorld> trying the cmd:P [15:29:13 CEST] <DelphiWorld> ok c_14 so the cover becaume video [15:29:16 CEST] <DelphiWorld> from 5MB to 11MB:P [15:29:51 CEST] <c_14> You can try playing around with lowering the framerate, but I'm not sure if YouTube needs a minimum framerate. [15:30:09 CEST] <c_14> You can also set the quality settings for the video encoder (in this case probably libx264). [15:30:10 CEST] <sine0> c_14: done it [15:30:56 CEST] <DelphiWorld> will use mpeg4 i think:P [15:31:27 CEST] <c_14> DelphiWorld: If your build has libx264 support .mp4 defaults to libx264. It defaults to mpeg4 if you don't though. [16:28:18 CEST] <RobertNagy> I'm trying to convert a m4v to mp4 but I keep getting the error 'pts has no value': 'ffmpeg -f h264 -i 1.h264 -c copy 1.mp4'. [16:28:40 CEST] <RobertNagy> The only way I've found around it is the following: 'ffmpeg -f h264 -i 1.h264 -c copy -f avi - | ffmpeg -i - -flags +global_header -c copy 1.mp4' [16:28:47 CEST] <RobertNagy> which is rather silly... is there a better way? [16:30:01 CEST] <Eduardo_1> it should be ffmpeg -i 1.h264 -c copy -f mp4 1.mp4 [16:30:09 CEST] <DHE> is it an error? when I try it it's just a warning and I still get my video [16:30:28 CEST] <RobertNagy> it's just a warning, however, some players can't play it correctly [16:30:32 CEST] <RobertNagy> e.g. chrome [16:30:55 CEST] <RobertNagy> @Eduardo_1: No difference. [16:30:56 CEST] <DHE> my cmd: fmpeg -r 60 -i input.264 -codec copy output.mp4 [16:31:01 CEST] <DHE> chrome plays it for me... [16:31:11 CEST] <RobertNagy> yes, it assumes it's 60 fps [16:31:17 CEST] <RobertNagy> but if your source material is 24 fps [16:31:19 CEST] <RobertNagy> it will play it [16:31:26 CEST] <RobertNagy> but it will be very unsmooth [16:31:46 CEST] <RobertNagy> and it's Chrome + MSE [16:31:47 CEST] <DHE> my commandline specifies the framerate. specify your own [16:32:07 CEST] <RobertNagy> you specify 60, which chrome seems to assume if it doesn't have the pts [16:32:54 CEST] <RobertNagy> either way, the resulting mp4 is not fully compliant, and I can't get ffmpeg to generate the pts without piping it through avi first. [16:34:41 CEST] <mpearrow> Are there any recommended tools for debugging corrupted/broken h.264 ? [16:35:01 CEST] <DHE> there is an option, -fflags genpts which MIGHT help... [16:35:08 CEST] <DHE> just skimming the ffmpeg help [16:35:22 CEST] <RobertNagy> yes, but that only works when inputting non-raw streams [16:35:42 CEST] <RobertNagy> which is why, if you look at my workaround, I first have to convert to avi and then pipe it again to ffmpeg [16:35:51 CEST] <RobertNagy> ooh, sorry, wrong workaround string [16:36:08 CEST] <RobertNagy> this works: ffmpeg -f h264 -i 1.h264 -c copy -f avi - | ffmpeg -fflags +genpts -i - -flags +global_header -c copy 1.mp4 -y [16:36:44 CEST] <RobertNagy> this doesn't: ffmpeg -fflags +genpts -f h264 -i 1.h264 -flags +global_header -c copy 1.mp4 -y [16:53:30 CEST] <Filarius> hello, what software I can use to make graphic for crf/qp/bitrate for video file with h264 video ? Target platform is Win [16:56:13 CEST] <BtbN> a graphic? [16:57:06 CEST] <mccoy_slack> hello, i'm using libavformat from ffmpeg to demux the output of the video card with gives me an mpeg transport with mpeg2video and mp2 audio, everything is but ffmpeg is always logging "first_dts not matching first dts in queue", what does that log mean? [16:57:12 CEST] <jarr0dsz3> hi everyone, im trying to download a rtmp stream as mp4 [16:57:21 CEST] <jarr0dsz3> although it fails, any experts could advice on the Cannot read RTMP handshake response error? [16:58:58 CEST] <jarr0dsz3> using something like ffmpeg -re -i rtmp://127.0.0.1/stream/test -c:a mp3 -c:v copy -copyts test.mp4 but it fails though [16:59:26 CEST] <jarr0dsz3> also i see a lot of ffmpeg output but no writing doing a ctrl + c to exit ffmpeg getting the handshake errors, tried debugging this for a full +1 hour, to no avail [17:01:11 CEST] <BtbN> i'd recommend just doing a full stream copy to flv [17:07:37 CEST] <Mista_D> anyway to change forced x264 level in the file's header without transcoding? [17:08:35 CEST] <c_14> Your friendly hex editor? [17:08:59 CEST] <klaxa> i think someone did that and then it was suggested to write a bitstreamfilter [17:09:05 CEST] <klaxa> not sure if anything came of that though [17:09:39 CEST] <c_14> Ye, if you want to do it with ffmpeg, you'd need a bitstream filter. Don't think one exists that does that though. [17:50:47 CEST] <nashgul> hi, good evening [17:52:03 CEST] <nashgul> i've a doubt with this expression: ffmpeg -i input.avi -i longo.png -filter_complex "[0:v][1:0] overlay=x=25:y=25" output.mp4 [17:52:27 CEST] <nashgul> what does it make ' [0:v][1:v] ' ? [17:53:05 CEST] <nashgul> i understand the rest of the expression [17:53:26 CEST] <nashgul> but not that pair of brackets expressions [17:54:59 CEST] <klaxa> overlay takes two inputs, [0:v][1:v] specifies those inputs [17:55:18 CEST] <klaxa> [0:v] means from the first input (0) take the video stream (v) [17:55:25 CEST] <klaxa> [1:v] means from the second input (1) take the video stream (v) [17:55:55 CEST] <klaxa> your first input (0) in this case is input.avi and the second one (1) is longo.png [17:56:36 CEST] <nashgul> klaxa: but if i don't put that pair of brackets, logo.png appears alike [17:56:55 CEST] <nashgul> i think is not neccesary, not? [17:57:10 CEST] <klaxa> i'm guessing the default behavior is to take the video streams of the first two inputs [17:58:12 CEST] <klaxa> if you are building more complex filters, referencing like [0:v] makes it easier to keep track of things and is probably necessary even depending on the kind of filtering you do [17:59:08 CEST] <nashgul> klaxa: umm, ok, i think i've understood now [17:59:22 CEST] <nashgul> klaxa: thanks :-D [17:59:34 CEST] <klaxa> glad i could help :) [18:09:29 CEST] <Mista_D> is there a Linux version of "direct264" bistream editor tool adapted to a more recent FFmpeg verison (need ISMV support)? [19:19:54 CEST] <tomM___> so i'm looking for a way to create video from still images, but with time durations specified for each image [19:19:57 CEST] <tomM___> can ffmpeg do that? [19:20:27 CEST] <tomM___> (i.e. a "slideshow" but without each image being on the screen for the same amount of time) [19:21:12 CEST] <relaxed> ffmpeg and scripting can do it [19:21:48 CEST] <pzich> hmm, there's probably a way to with some complex filters, or you can duplicate/symlink the images and assemble that as a movie at whatever framerate you want [19:22:03 CEST] <tomM___> relaxed: i'm comfortable scripting -- how would i go about it? [19:22:58 CEST] <relaxed> I think the easy way would be to pipe rawvideo from mulitple ffmpeg's to one ffmpeg that's encoding the video [19:24:34 CEST] <tomM___> relaxed: so one ffmpeg for each image in the video? [19:24:47 CEST] <tomM___> pzich: hm, that's an interesting idea. would that make it very slow? [19:25:00 CEST] <pzich> only one way to find out ;) [19:25:02 CEST] <pzich> (I'm not sure) [19:25:36 CEST] <relaxed> how many different images are there? [19:26:46 CEST] <tomM___> relaxed: well, i was hoping to be able to get as long as i want [19:26:53 CEST] <tomM___> e.g. up to 45 mins [19:27:00 CEST] <tomM___> with maybe an image every 5 secs [19:28:10 CEST] <nashgul> hi again, i'm using -filter_complex with some expressions, for a drawtext scroll i need the frame or the time which -filter_complex is launched, 'n' and 't' returns the the frame and the total time [19:29:21 CEST] <nashgul> exactly i'm doing this: drawtext=text='string1':x=w-(n*5):y=(h/PHI):enable=between(t,30,50) [19:30:01 CEST] <nashgul> this line works if i launch the filter at the beginning, but if the filter starts at 30' doesn't works [19:30:58 CEST] <pzich> so subtract that from n? [19:31:58 CEST] <nashgul> ummm, n grows with the time, the solution would be A=n, x=w-A-(n*5) [19:32:47 CEST] <nashgul> can i use variables in ffmpeg? [19:36:01 CEST] <nashgul> sorry, would be: x=w+A-(n*5) [19:40:05 CEST] <relaxed> how are you quoting the filter? you're example isn't working here [19:41:00 CEST] <relaxed> your* [19:42:01 CEST] <nashgul> ffmpeg -i input.avi -filter_complex "drawtext=text='string':x=w-(5*n):y=(h/PHI)+th:enable='between(t,20,30)'" -strict -2 output.mp4 [19:48:10 CEST] <nashgul> i can calculate an offset: framerate=25 -> A=25*(10 secs)=250 => x=w+250-(n*5) and put in the script but i doesn't exact [19:48:20 CEST] <nashgul> i can calculate an offset: framerate=25 -> A=25*(10 secs)=250 => x=w+250-(n*5) and put in the script but it is not exact [19:48:24 CEST] <tomM___> ok, i'm running some timing tests with the symlinks now. if it's way too slow, can anyone recommend any other tools than ffmpeg (though i know i'm on #ffmpeg :) ) [19:50:45 CEST] <relaxed> nashgul: it works here if I add setpts=PTS-STARTPTS in the beginning of the filter chain. [19:51:45 CEST] <relaxed> wait, it works without it too :) [19:52:13 CEST] <pzich> tomM___: you may be better off using image2pipe then you can read in the image once in your script and replay it to ffmpeg as many times as you want [19:53:15 CEST] <relaxed> nashgul: which version are you using? [19:53:32 CEST] <nashgul> relaxed: 2.7.1 [19:54:01 CEST] <nashgul> relaxed: setpts doesn't works [19:54:34 CEST] <relaxed> so nothing happens? [19:54:49 CEST] <nashgul> the strings does not appears at 10' [19:54:53 CEST] <nashgul> the string does not appears at 10' [19:56:23 CEST] <nashgul> relaxed: is there any way to set n=0 when the filter is launched? [20:10:39 CEST] <nashgul> bah, i'm not very clever, if i use t instead n i can do the scroll easier [20:10:42 CEST] <nashgul> :-D [20:13:09 CEST] <nashgul> enable='between(t,10,20)':x=w+(10*45)-(45*t) [20:21:51 CEST] <samons> hi [20:22:34 CEST] <samons> unrecognized option c:v [20:24:04 CEST] <samons> http://pastie.org/10260630 [20:24:31 CEST] <llogan> FFmpeg version SVN-r23418, Copyright (c) 2000-2010 [20:24:34 CEST] <llogan> absolutely ancient [20:25:41 CEST] <samons> i don't know how i get that version? i've k-lite media codec pack. I just grab latest version of FFMPEG [20:25:56 CEST] <llogan> what is your OS? [20:26:00 CEST] <samons> XP [20:26:08 CEST] <llogan> http://ffmpeg.zeranoe.com/builds/ [20:26:23 CEST] <samons> 20150626 version i've [20:26:23 CEST] <llogan> hopefully it will work on your old OS [20:27:14 CEST] <samons> which date to look? [20:27:35 CEST] <samons> 2013 or 2014? [20:27:39 CEST] <llogan> none. just use the latest available [20:27:52 CEST] <samons> i'm using latest :) [20:28:04 CEST] <samons> maybe need to reset ffmpeg path [20:28:16 CEST] <llogan> you're using something from 2010 [20:28:43 CEST] <samons> yeah, that what i figured to use c:v i need newer version [20:29:16 CEST] <tomM___> soo... i tried pzich's suggesion of using symlinks to create a photo slideshow, and setting e.g. framerate to 24fps -- and what happens is strange. i tried alternating 2 images and switching between them, and the timing is not regular [20:29:47 CEST] <tomM___> so like it plays back a flickering video that doesn't flicker evenly [20:29:55 CEST] <chungy> Hmm. Too bad it doesn't seem any of the stable versions are available for windows [20:30:09 CEST] <tomM___> ( pzich relaxed ^^) [20:30:13 CEST] <llogan> current git master is considered stable [20:30:21 CEST] <llogan> releases are for distros [20:31:03 CEST] <llogan> and Zeranow also provides most releases if you need that for whatever reason [20:31:04 CEST] <tomM___> the command i'm using is: "ffmpeg -framerate 24 -i /tmp/foo/img%06d.jpg -c:v libx264 -r 30 -pix_fmt yuv420p out.mp4" [20:32:02 CEST] <tomM___> it's also saying it's got 126 frames, and i'm using 100 images at 24fps... which seems very wrong [20:32:25 CEST] <tomM___> (in the progress bar it says "frame= 126") [20:36:43 CEST] <samons> i use abspath of where ffmpeg is extracted it seems works [20:37:29 CEST] <pzich> tomM___: you're specifying that the input is 24fps and encoding at 30fps, which is probably why your frames aren't even [20:38:00 CEST] <samons> thanks :) [20:38:41 CEST] <samons> it show N-73165 2000-2015 [20:38:49 CEST] <llogan> that's better [20:39:27 CEST] <samons> i don't know how ffmpeg is installed older version need to google it :) [20:43:26 CEST] <samons> thanks for help :) [20:54:06 CEST] <pzich> tomM___: so I wrote this script which repeats images based on the filename, seems to be working in my tests: http://pastebin.com/q5cnq889 [20:58:26 CEST] <tomM___> pzich: thanks [21:05:24 CEST] <tomM___> so pzich : $C in your script represents the number of frames to play the image for? [21:11:32 CEST] <tomM___> pzich: hm, it also seems not to work -- i just get a 00:00 video [23:06:21 CEST] <Exagone313> Hello, I try to convert a RTMP (h.264 & mp3) to ogg + webm for live streaming. But it does not work and I get one of these errors: "av_interleaved_write_frame(): Connection reset by peer" "Missing video stream which is required by this ffm". ffmpeg version 2.7.1-static http://johnvansickle.com/ffmpeg/ [23:06:31 CEST] <Exagone313> What can I do? I can link my cfg [23:07:51 CEST] <pzich> I don't know the answer, but they're going to want to see a paste with the whole command and its output [23:09:18 CEST] <Exagone313> for the second error - is there a command to assume that if there is no video stream, use a black frame while waiting for the video? [23:09:33 CEST] <Exagone313> or the last frame [23:14:23 CEST] <Exagone313> http://pastie.org/private/14l6bgaksot3inrsvyng [23:19:57 CEST] <nashgul> hi again, ffmpeg -i input.avi -f tee "[f=ogg]salida.ogg|[strict:2]salida.mp4" =>> output file #0 does not contain any stream [23:20:07 CEST] <nashgul> what's the problem with that line? [23:20:26 CEST] <llogan> Exagone313: you trimmed the console output [23:20:40 CEST] <nashgul> ok, sorry [23:20:58 CEST] <Exagone313> llogan: you need the headers? [23:21:03 CEST] <llogan> of course [23:21:07 CEST] <Exagone313> the footers said the message is repeated [23:21:24 CEST] <llogan> yes, you can trim the multiple repeating lines [23:21:32 CEST] <Exagone313> http://pastie.org/10260867 [23:23:06 CEST] <nashgul> this is my script: http://termbin.com/keg2 [23:23:30 CEST] <nashgul> this is the output of ffmpeg: http://termbin.com/b0pc [23:24:20 CEST] <llogan> please provide provide an unscripted command [23:24:34 CEST] <nashgul> :-D, ok ok, sorry again [23:26:27 CEST] <nashgul> this is my line and the output: http://pastie.org/10260871 [23:48:14 CEST] <llogan> nashgul: tee muxer is used to output the same data to various formats, without needed to perform multiple encoding [23:49:05 CEST] <llogan> what video and audio formats are you trying to place in both ogg and mp4? [23:49:51 CEST] <nashgul> llogan: ok, i'm reading about that now, i'm beggining to think i'm using tee in wrong way [23:50:05 CEST] <nashgul> llogan: i'm trying with 'split' [23:50:54 CEST] <llogan> that will be a better option [23:51:02 CEST] <nashgul> llogan: ok [23:53:33 CEST] <llogan> nashgul: for video only, it could be like: split[v0][v1]" -map "[v0]" -c:v libx264 out.mp4 -map "[v1]" -c:v libtheora out.ogg [23:54:00 CEST] <nashgul> llogan: http://pastie.org/10260904 [23:54:14 CEST] <nashgul> but that line does not streams audio :\ [23:54:32 CEST] <nashgul> i want to stream video+audio [23:55:45 CEST] <llogan> you need to map the audio too. [23:56:00 CEST] <nashgul> llogan: i don't know how i do that [23:56:06 CEST] <llogan> add "-map 0:a" for each output [23:56:20 CEST] <nashgul> umm [23:57:28 CEST] <nashgul> llogan: yeah! thanks a lot! :-D [23:59:28 CEST] <Exagone313> llogan: do you know something about my errors? [00:00:00 CEST] --- Sat Jun 27 2015

1 0

ffmpeg-devel.log.20150625
by burek 26 Jun '15

26 Jun '15

[00:11:03 CEST] <cone-218> ffmpeg 03James Zern 07master:469398f8b7a8: libvpxdec: report profile for vp9 [00:11:04 CEST] <cone-218> ffmpeg 03James Zern 07master:766a8346cfb4: libvpxdec: report colorspace [00:18:31 CEST] <cone-218> ffmpeg 03Reynaldo H. Verdejo Pinochet 07master:89234deaa2a4: ffserver: simplify URLContext cleanup [00:18:32 CEST] <cone-218> ffmpeg 03Reynaldo H. Verdejo Pinochet 07master:1714fe2990d5: ffserver: factor out stream params printing [00:18:33 CEST] <cone-218> ffmpeg 03Reynaldo H. Verdejo Pinochet 07master:469c335c5567: ffserver: unify comment formating & drop unneeded braces [00:18:34 CEST] <cone-218> ffmpeg 03Reynaldo H. Verdejo Pinochet 07master:36a617c1625d: ffserver: make statement block unconditional by exiting early [00:18:35 CEST] <cone-218> ffmpeg 03Reynaldo H. Verdejo Pinochet 07master:08724da976e9: ffserver: make statement block unconditional by exiting early [00:26:33 CEST] <cone-218> ffmpeg 03Michael Niedermayer 07master:2d8bf3d126a3: avcodec/jpeg2000dec: Check PPx / PPy values [00:26:34 CEST] <cone-218> ffmpeg 03Michael Niedermayer 07master:67fc89824373: avcodec/jpeg2000: Improve right bottom precinct positions [01:16:45 CEST] <cone-218> ffmpeg 03Michael Niedermayer 07master:9ff5e7974a8f: avcodec/jpeg2000dec: Limit x/y step to 1<<30 [04:28:34 CEST] <cone-765> ffmpeg 03Ludmila Glinskih 07master:b04c630a3dbe: libavformat: Add H264 API test [04:57:55 CEST] <cone-765> ffmpeg 03Ronald S. Bultje 07master:d5a36edda437: vp9: fix profile check for intraonly frames. [11:34:47 CEST] <feliwir> who is peter ross in here? [11:36:13 CEST] <feliwir> pross_: how to get the alpha mask with your vp6 patch? [11:49:23 CEST] <wm4> I'd assume it'll output a YUVA pixfmt, if it doesn't discard the alpha part [11:52:16 CEST] <durandal_1707> michaelni: does gsoc student need to post something on ml before 26.6? [11:52:57 CEST] <michaelni> its good if they post but not strictly mandatory [12:19:43 CEST] <durandal_1707> nobody likes new wip filter :( [12:43:21 CEST] <cone-287> ffmpeg 03rogerdpack 07master:1bcdac2da3ca: ffmpeg: log error message when shutting down from too many signals [13:15:53 CEST] <cone-287> ffmpeg 03Peter Ross 07master:a2517fca1e7d: electronicarts: move video stream properties into dedicated structure [13:15:54 CEST] <cone-287> ffmpeg 03Peter Ross 07master:803bdc546942: electronicarts: demux alpha stream [15:12:22 CEST] <[-T-]> BtbN: did you look at the quicksync communty package ? [15:12:57 CEST] <BtbN> no [15:14:41 CEST] <[-T-]> ok [15:15:24 CEST] <[-T-]> I am not sure where we are on this subject, is there a chance we could see a proper support (including linux) being implemented ? [15:21:20 CEST] <BtbN> I think qsv encoding works fine? [15:21:22 CEST] <BtbN> On both OS? [15:21:46 CEST] <[-T-]> yes but not decoding and picture processing [15:22:07 CEST] <[-T-]> a full hw transcoding would be great ... [15:22:19 CEST] <BtbN> Don't see a point in implementing qsv based decoding [15:22:35 CEST] <BtbN> there's dxva on windows and vaapi/vdpau on linux. [15:23:05 CEST] <[-T-]> i don't think it's possible to decode with vaapi in ffmpeg in a transcoding context ? [15:23:15 CEST] <BtbN> Use the vdpau wrapper. [15:23:24 CEST] <[-T-]> ok that works... [15:23:30 CEST] <[-T-]> i mean [15:23:39 CEST] <BtbN> libav* supports vaapi just fine, but the cli tool doesn't. [15:23:44 CEST] <[-T-]> vdpau works with a nvidia [15:23:59 CEST] <[-T-]> what do you mean by vdpau wrapper ? [15:23:59 CEST] <BtbN> vdpau works for basicaly everything [15:24:12 CEST] <BtbN> Use the vdpau wrapper to decode on your intel card. [15:24:15 CEST] <[-T-]> oh [15:24:23 CEST] <[-T-]> if that's possible, it's great :) [15:24:31 CEST] <[-T-]> can you point me where to look ? [15:24:50 CEST] <BtbN> x11-libs/libvdpau-va-gl is the name of it on gentoo. [15:25:34 CEST] <[-T-]> interesting [15:25:43 CEST] <[-T-]> i will do some researches thx [15:26:07 CEST] <[-T-]> i have ported ffmpeg_vdpau.c to be used into tvheadend [15:26:18 CEST] <[-T-]> that works and i gain about 15%cpu out of 35 [15:26:29 CEST] <[-T-]> but it wakes up my nvidia card and I don't like that [15:26:29 CEST] <BtbN> ffmpeg_vdpau.c is part of the ffmpeg cli tool [15:26:33 CEST] <[-T-]> yes [15:27:05 CEST] <BtbN> Adding vdpau code to tvh seems rather out of place in it [15:27:07 CEST] <[-T-]> if there a better way to use vpdau to decode a frame in a transcoding context ? [15:27:09 CEST] <BtbN> juse use the cli tool [15:28:10 CEST] <[-T-]> I see some files such as: ./libavcodec/vdpau_h264.c [15:28:19 CEST] <[-T-]> i guess they cannot be used to get back the frame right [15:28:27 CEST] <[-T-]> it's only rendering onscreen ? [15:29:57 CEST] <BtbN> There is no accelerated rendering in ffmpeg [15:30:07 CEST] <BtbN> The only renderer i know of is the sdl thing [15:30:24 CEST] <BtbN> And it's not meant to be used in applications, more of a test tool for the cli [15:31:00 CEST] <[-T-]> ok [15:31:20 CEST] <[-T-]> if I use -hwaccel vdpau in command line, it's going to end up in ./libavcodec/vdpau_h264.c right ? [15:31:34 CEST] <BtbN> IT will use ffmpeg_vdpau.c [15:32:49 CEST] <[-T-]> ok that's it [16:11:29 CEST] <J_Darnley> Have you people noticed that on the fate page the results table gets silently cut off if the browser window is too narrow? [16:15:19 CEST] <J_Darnley> Okay, how do I go about submitting a patch for the CSS on the fate page? Is it in the website repo or the fate repo? [16:32:47 CEST] <ubitux> so, we can't know the pixel format until we decode the first frame [16:33:13 CEST] <ubitux> and so after every frame we need to eventually reconfigure the filtergraph [16:37:03 CEST] <ubitux> i'm also kind of confused to have a AVCodecContext.pix_fmt wrong with hwaccel [17:06:49 CEST] <rcombs> http://puu.sh/iCjd9/917bcdd90d.mkv <-- this sample's audio stream seems to anger Chrome, and mediainfo returns some strangeness on it; anyone mind taking a look to see if there's anything incorrect in lavc's handling of the HE-AACv2? [17:08:28 CEST] <[-T-]> BtbN, I tried https://github.com/i-rinat/libvdpau-va-gl [17:08:33 CEST] <[-T-]> it runs but very slowly [17:10:27 CEST] <[-T-]> and vdpau, using -hwaccel vdpau, ends up using ./libavcodec/vdpau_h264.c.. as I said [17:12:49 CEST] <BtbN> Of course it uses that, but it's not a self contained h264 decoder [17:12:58 CEST] <[-T-]> yes yes [17:13:02 CEST] <BtbN> It's merely the vdpau interop for h264 [17:13:06 CEST] <[-T-]> ok [17:14:39 CEST] <[-T-]> is G. Beauchesne on the chan ? [17:15:23 CEST] <av500> [-T-]: __gb__ [17:15:45 CEST] <[-T-]> og :) [17:15:56 CEST] <[-T-]> that all makes sense now haha [18:47:26 CEST] <cone-287> ffmpeg 03rogerdpack 07master:87961eff23ab: modify exit message when signaled [18:47:27 CEST] <cone-287> ffmpeg 03Shawn Rainey 07master:2c77ca459e9c: configure: Fix ranlib failure with older cross toolchains [20:15:59 CEST] <cone-287> ffmpeg 03Rodger Combs 07master:0643b4bf74a4: lavf/brstm: expose the loop point when present [20:16:00 CEST] <cone-287> ffmpeg 03Rodger Combs 07master:388a6f581ab6: lavc+doc: adjust names to reflect ADPCM THP not being GameCube-only [22:07:30 CEST] <cone-287> ffmpeg 03Paul B Mahol 07master:cdcf637d56c6: avcodec/evrc: make it possible to disable postfilter [23:52:49 CEST] <cone-287> ffmpeg 03Michael Niedermayer 07master:8c22143e7e00: ffmpeg: Use 2 instead of STDERR_FILENO [00:00:00 CEST] --- Fri Jun 26 2015

1 0

ffmpeg.log.20150625
by burek 26 Jun '15

26 Jun '15

[00:36:09 CEST] <you_all_suck> hey you dregs, I have an ffmpeg issue [00:36:09 CEST] <you_all_suck> http://pastebin.com/ZX1jxxWc [00:36:24 CEST] <you_all_suck> I'm trying to convert a series of bmp images to a video file [00:36:41 CEST] <you_all_suck> repeated attempts have only served to frustrate my goal of doing something that NOT only should be very simple [00:36:47 CEST] <Anoia> calling us dregs and saying we all suck is a great way to get help. [00:36:57 CEST] <you_all_suck> but should be the minimum expectation of any video editing tool [00:37:20 CEST] <you_all_suck> Anoia: I call em as I see em [00:37:22 CEST] <you_all_suck> now help me [00:38:26 CEST] <Anoia> fuck off [00:38:45 CEST] <Anoia> (not official channel stance) [00:39:08 CEST] <you_all_suck> I get that alot [00:39:16 CEST] <you_all_suck> the question [00:39:24 CEST] <you_all_suck> *the question's still open to anyone who can answer it [00:39:37 CEST] <you_all_suck> otherwise I'm fixing to tweet about my negative experience using this tool [00:41:37 CEST] <absent2> lol [00:48:39 CEST] <Anoia> don't forget the #ffmpegsuxballs hashtag. [00:49:14 CEST] <Anoia> and maybe #peoplewonthelpmeafteriinsultthem [00:49:30 CEST] <absent2> FFMPEG fails using the slackbuild because it can't find caca, but I have libcaca installed. Works fine on my laptop but not the server. Weird. [00:49:32 CEST] <Anoia> #nfiwhy [01:10:28 CEST] <Anoia> no twitter post about it yet [01:10:34 CEST] Action: Anoia goes to sleep [01:14:10 CEST] <absent2> looks like alien bob's package installed fine, something weird with the slack package I guess [01:20:50 CEST] <you_all_suck> well I waited here and even had enough time to take a crap [01:20:56 CEST] <you_all_suck> and I still don't see a solution [01:21:08 CEST] <you_all_suck> talentless hacks, all of you [01:21:53 CEST] <you_all_suck> I won't recommend this tool to anyone else that I know and I'll shortly update my blog with my unpleasant experience with it and with the irc support channel [01:22:11 CEST] <you_all_suck> thanks for nothing [01:34:02 CEST] <Anoia> lol [01:34:20 CEST] <Anoia> still not on twitter [01:36:55 CEST] <absent2> they see me trollin [01:36:57 CEST] <absent2> they hatin [01:37:00 CEST] <absent2> ... [03:01:01 CEST] <bertieb> In case there's someone here who knows now: is there a way to specify an output codec_time_base (tbc) when encoding? [04:52:15 CEST] <troy_s> Anoia: We will miss him. I am certain that his blog and Tweets will have a tremendously adverse impact on the FFMPEG project. [04:56:17 CEST] <ac_slater_> hey all. [04:56:50 CEST] <ac_slater_> Im muxing some mpegts. What if I say a stream's timebase is 1/15 but I insert into the context (mux) more that 15 times a second? [09:43:28 CEST] <feliwir> how to resolve undefined reference to `DtsProcOutputNoCopy' when linking with ffmpeg? [09:43:46 CEST] <feliwir> i am not sure from which library that function is [09:47:48 CEST] <feliwir> help is appreciated [10:23:54 CEST] <chrisss123456> Hi, I have a typical nooby question: I'm trying to convert an mp4 file into an mpg file, without loss. I used the standard ffmpeg -i input.mp4 output.mpg but while the video was still in 1080p, the quality was shitty. Any help? [10:31:01 CEST] <cbsrobot_> chrisss123456: try: ffmpeg -i input.mp4 -t 1 -q:v 0 output.mpg [10:31:19 CEST] <cbsrobot_> and if you like the sample output, remove the -t 1 [10:31:20 CEST] <chrisss123456> cbsrobot_: thanks! what do those flags mean? [10:31:52 CEST] <cbsrobot_> if the file is tooo heavy, you could rise the -q value [10:36:28 CEST] <hrw> can someone point me into some howto to convert multiaudiolanguage/dvd into multiaudiolanguage/h264 something? [10:42:06 CEST] <feliwir> does someone know if the ffmpeg vp6 decoder does actually fill the frame with ALPHA_DATA? [10:44:22 CEST] <cbsrobot_> feliwir: I think there is commit on the ml [10:44:48 CEST] <cbsrobot_> hrw: check for the -map option - I think there's something on the wiki [10:45:02 CEST] <hrw> cbsrobot_: thanks [10:45:05 CEST] <cbsrobot_> hrw: https://trac.ffmpeg.org/wiki/How%20to%20use%20-map%20option [10:45:19 CEST] <feliwir> cbsrobot_: ml? [10:45:24 CEST] <cbsrobot_> Mailinglist [10:45:26 CEST] <cbsrobot_> ffmpeg-devel [10:45:54 CEST] <feliwir> cbsrobot_: i saw that one [10:46:20 CEST] <cbsrobot_> and ? [11:10:01 CEST] <feliwir> how do i apply a change from the ml? [11:14:24 CEST] <durandal_1707> git apply patch [11:16:02 CEST] <feliwir> durandal_1707: and how to get the patch out of the mail? [11:16:22 CEST] <rcombs> if you've got the message as a .mbox then `git am` can help [11:17:29 CEST] <feliwir> i got it as a mail. And the mail has the patch in it like this: http://pastebin.com/mjgBiXwm [11:17:36 CEST] <feliwir> is this already applyable? [11:21:27 CEST] <chungy> yeah [11:21:54 CEST] <chungy> looks like you cut off the headers. [11:22:14 CEST] <feliwir> i applied this now [11:22:30 CEST] <chungy> Your email client probably has a "Save as mbox" option, use that :) [11:22:38 CEST] <feliwir> will this change already make AVFrame being filled with rgba data when decoding vp6 that has alpha? [11:22:44 CEST] <chungy> otherwise git won't take the commit message or author information as intended [11:24:51 CEST] <feliwir> is the patch author in here? peter ross? [11:32:16 CEST] <JEEBsv> I think he's on the -devel channel [11:34:56 CEST] <feliwir> ah, never saw that channel before [11:39:43 CEST] <hrw> when 'ffmpeg -i' lists several audio tracks is there a way to get language info? [11:40:04 CEST] <Mavrik> doesn't ffprobe list it? [11:41:58 CEST] <hrw> Stream #0:2: Audio: ac3, 48000 Hz, 5.1(side), fltp, 384 kb/s [11:41:58 CEST] <hrw> Stream #0:3: Audio: ac3, 48000 Hz, 5.1(side), fltp, 384 kb/s [11:41:58 CEST] <hrw> Stream #0:4: Audio: ac3, 48000 Hz, 5.1(side), fltp, 384 kb/s [11:42:56 CEST] <Mavrik> That usually means it can't recognise the language :/ [11:43:03 CEST] <hrw> ok [11:43:31 CEST] <hrw> used mplayer to find out en/cs/pl/en-with-comments [11:45:06 CEST] <hrw> started "ffmpeg -f concat -i /tmp/del/l -c:v copy -c:a:0 copy -c:a:2 copy -c:a:3 copy out.vob" and then can reencode video if this works [12:02:54 CEST] <hrw> http://pastebin.com/kZiyVRK8 - what I am doing wrong? I want to copy audio streams without reencoding them and ffmpeg wants to reencode [12:07:07 CEST] <hrw> numbers were wrong ;d [12:09:48 CEST] <Mavrik> I think just "-codec copy" works as well :) [12:10:20 CEST] <hrw> :D [12:13:09 CEST] <hrw> now only make audio stream #2 -> #1 and #1 -> #2 so it will start with proper language [12:25:38 CEST] <nick123> Anyone know what this mean "overread end of atom" [12:26:26 CEST] <feliwir> does ffmpeg have some functionality to mask one AVFrame on top of another? [13:03:05 CEST] <bertieb> nick123: Seems to be a warning log message (see https://ffmpeg.org/pipermail/ffmpeg-cvslog/2013-November/070705.html) [13:05:58 CEST] <nami> hey. no way to play media with ffmpeg static on linux? [13:07:06 CEST] <nami> there is only: ffmpeg, ffprobe, ffserver and qt-faststart in it [13:10:24 CEST] <BtbN> If you don't build it with SDL support, no. [13:13:00 CEST] <nami> http://johnvansickle.com/ffmpeg/ [14:54:01 CEST] <hrw> is there a way to add language tags to audio streams? [15:13:15 CEST] <feliwir> how frequently zeranoe builds are created? [15:15:09 CEST] <feliwir> damn there was just a build that doesn't have the changes i needed -.- [15:20:23 CEST] <feliwir> can anyone actually confirm this works: https://www.ffmpeg.org/platform.html#Native-Windows-compilation-using-MinGW… [15:23:36 CEST] <feliwir> because i keep getting this:Unknown OS 'msys_nt-6.3'. [15:29:14 CEST] <feliwir> can anyone please help me out with this problem? [15:32:34 CEST] <feliwir> it also tells me to use enable cross compile (which doesn't change a thing) [15:33:30 CEST] <feliwir> http://pastebin.com/FS0x9GT6 [15:33:35 CEST] <feliwir> (yes cl is in path) [15:51:14 CEST] <feliwir> did noone here ever compile ffmpeg that way? [15:55:16 CEST] <bertieb> feliwir: If no-one answers that means probably not [15:55:47 CEST] <bertieb> But a quick search turns up https://trac.ffmpeg.org/ticket/3203 and https://ffmpeg.org/pipermail/ffmpeg-user/2012-December/011861.html which you may already have sene [15:55:56 CEST] <bertieb> s/sene/seen/ [15:58:26 CEST] <bertieb> In fact, a search for "cl is unable to create an executable file" returns a number of results from zeranoe forums itself, which may be of use [15:58:30 CEST] <feliwir> bertieb, it didn' really get answers [16:01:20 CEST] <feliwir> ah i think i got it [16:01:23 CEST] <bertieb> feliwir: Indeed, but it suggests starting points for getting futher information which are not included in your pastebin link. If you have searched and read those trac issues/ml threads/forums posts it is worth saying what you have tried [16:01:55 CEST] <bertieb> feliwir: Since that will let people know you have done your own research before coming here [16:02:23 CEST] <feliwir> it was using the wrong link.exe [16:02:37 CEST] <feliwir> maybe you should mention that on the mainpage since every msys2 installation has that [16:03:09 CEST] <bertieb> If you are referring to me, I'm just here to get help with a problem, like you [16:03:25 CEST] <feliwir> ah okay. What's yours? [16:03:31 CEST] <feliwir> i was talking to ffmpeg staff :D [16:04:02 CEST] <bertieb> An email to them with a well-phrased suggestion stands more chance of being read if that's you intention [16:04:41 CEST] <bertieb> Thanks for asking, I've posted my problem to the ml, but here it is: https://ffmpeg.org/pipermail/ffmpeg-user/2015-June/027255.html [16:05:18 CEST] <bertieb> (There are also a couple of posts on SuperUser relating to it) [16:07:01 CEST] <davidpeach> hi there [16:07:36 CEST] <bertieb> But to summarise- Small question: can I set output tbc in some way? Much longer question: can I match codec parameters to concat demux video files together [16:07:53 CEST] <davidpeach> i was in need of some help with streaming some ffmpeg output please? I have it in place and working - i just needed some advice if possible? [16:07:58 CEST] <feliwir> bertieb, sorry no idea [16:08:10 CEST] <bertieb> Since I've not yet had a reply on SU or in here I suspect it is a fairly niche question [16:08:19 CEST] <bertieb> But I'm relatively patient :) [16:08:32 CEST] <bertieb> feliwir: Appreciate you asking anyway [16:08:50 CEST] <bertieb> davidpeach: Ask (don't ask to ask) [16:11:16 CEST] <davidpeach> oki. So i am using nodejs and am running the ffmpeg command (this time just creating a new video from current - no filters) and when I pipe it out in the response stream, the result is very choppy. I think the rate at which it is creating the new video is much slower than playback speed. Any advice on speeding up the creation process / better way to approach the problem? [16:11:38 CEST] <BtbN> pur more cores into the host [16:11:59 CEST] <davidpeach> My end goal is to use the complex filter, inserting images over the base video to create new video and piping it straight out to user. [16:12:16 CEST] <davidpeach> currently all videos need to be made in advance (there's thousands of them) [16:12:39 CEST] <davidpeach> BtbN : how do I do that please? [16:12:43 CEST] <feliwir> hm. Make install is faling now :'D [16:12:52 CEST] <BtbN> By taking out the old, slow, CPU, and putting in a new one. [16:13:09 CEST] <BtbN> If your server is overwhelmed with a single live-transcode, i think you are out of luck with that approach for a productive public setup. [16:13:20 CEST] <feliwir> ideas: http://pastebin.com/W02HNgA8 ? [16:14:34 CEST] <bertieb> feliwir: Are you cross-compiling? [16:20:19 CEST] <feliwir> bertieb, i am compiling on windows for windows [16:20:24 CEST] <feliwir> so i guess no [16:20:40 CEST] <bertieb> feliwir: Have you disabled cross-compilation related options? [16:20:56 CEST] <feliwir> bertieb, i am only doing: ./configure --toolchain=msvc --enable-shared [16:21:17 CEST] <feliwir> like the guide on the homepage told me to do [16:22:44 CEST] <BtbN> of course make install fails on windows [16:22:48 CEST] <BtbN> if make succeeded, you are done [16:23:24 CEST] <bertieb> feliwir: Can you run what has been built? [16:23:51 CEST] <feliwir> BtbN, why the homepage says to do make install then? [16:24:17 CEST] <BtbN> It does? [16:24:23 CEST] <feliwir> bertieb, i have no idea where it is located [16:24:31 CEST] <feliwir> BtbN, https://www.ffmpeg.org/platform.html#Microsoft-Visual-C_002b_002b-or-Intel-… [16:24:34 CEST] <bertieb> feliwir: Which homepage? I'e searched for "make install" on http://ffmpeg.org and get no results [16:24:55 CEST] <BtbN> that make install seems strange there [16:24:59 CEST] <BtbN> just don't run it [16:25:32 CEST] <feliwir> BtbN, i can't find any .lib or .dll files in my ffmpeg folder [16:25:45 CEST] <BtbN> But make succeeded without errors? [16:26:09 CEST] <feliwir> yes, only a couple warnings [16:26:28 CEST] <BtbN> You should have the ffmpeg binaries in your build dir then [16:26:30 CEST] <feliwir> wait, nevermind [16:26:46 CEST] <feliwir> http://pastebin.com/2rw0vJVy [16:27:18 CEST] <bertieb> feliwir: that's included in your make install output [16:27:31 CEST] <bertieb> Did the previous 'make' command also generate that? [16:27:36 CEST] <feliwir> bertieb, also occurs when running make [16:27:37 CEST] <feliwir> yes [16:28:09 CEST] <bertieb> Hmm. Do you need the shared libraries option? You could try a static build [16:29:54 CEST] <feliwir> bertieb, i tried that before [16:29:59 CEST] <feliwir> also no files generated [16:30:07 CEST] <bertieb> Same error? [16:30:19 CEST] <feliwir> i gonna try "make" now instead of "make -j4" [16:30:27 CEST] <feliwir> maybe it doesn't like that [16:30:45 CEST] <davidpeach> Thanks anyway [16:33:08 CEST] <Eduardo_1> hello [16:39:50 CEST] <Eduardo_1> I just downloaded, compiled, and installed ffmpeg 2.7.1, to use nvenc. It works with no problem at all, but i'm unable to use 2 pass encoding, the log file that ffmpeg generates always has size 0. [16:39:50 CEST] <Eduardo_1> the command is: ffmpeg -i input.mp4 -y -vcodec nvenc -2pass 1 -preset llhq -b:v 1000k -profile:v main -pass 1 -an -f mp4 test.mp4 [16:43:00 CEST] <Eduardo_1> how should i do it? [16:52:47 CEST] <feliwir> bertieb, that was the error :D [16:53:02 CEST] <feliwir> someone knows what are the possible reasons of avformat_open_input failure? [16:53:30 CEST] <bertieb> feliwir: Presumably some race condition in the makefile then, glad you got the compile tow ork tho [16:53:48 CEST] <feliwir> thanks for the help [16:55:06 CEST] <feliwir> the return value is 0xbebbb1b7 [17:00:10 CEST] <feliwir> what does it mean? [17:00:17 CEST] <feliwir> the file specified does exist 100% [17:07:22 CEST] <feliwir> the error is this: Invalid data found when processing input [17:07:33 CEST] <feliwir> this morning it worked on linux -.- [17:07:39 CEST] <feliwir> now on windows i get this [17:55:45 CEST] <feliwir> can someone tell me where that error comes from? [18:24:56 CEST] <delarita_> Do someone know something about how to influence pic_init_qs_minus26 value in a ffmpeg commnand ? [20:16:50 CEST] <Odyssey> Hi everyone. I'm using GNU/Linux since a few months and I want my system to be as free as possible. FFmpeg is a free sotware package, but why is it in the "non-free" repository on debian ? (I'm very new in the GNU/Linux world, so I'm not an expert ^_^) [20:19:30 CEST] <relaxed> Odyssey: "apt-cache show ffmpeg" lists Filename: pool/main/f/ffmpeg/ffmpeg_2.7.1-1_amd64.deb [20:21:17 CEST] <bertieb> Odyssey: Also, some parts have (had?) non-free licenses [20:22:41 CEST] <relaxed> so, pool/main isn't pool/non-free [20:25:18 CEST] <Odyssey> So there's two repository ? One with non free package and one without them ? [20:26:27 CEST] <relaxed> Where are you getting that it's non-free? [20:28:02 CEST] <Odyssey> Because I have to install Ffmpeg but I have to add this line in the source.list : "deb http://www.deb-multimedia.org jessie main non-free", that's why I'm asking if it's free or not ^_^ [20:28:23 CEST] <relaxed> which debian version? [20:28:37 CEST] <relaxed> ffmpeg should be in jessie and unstable [20:29:21 CEST] <relaxed> er, it's not in Jessie [20:29:40 CEST] <Odyssey> (Yes, I'm on Jessie) [20:30:41 CEST] <relaxed> it's in unstable not, but I don't recommend using external repos for anything. If you want a gpl'd version you can use my static build. http://johnvansickle.com/ffmpeg/ [20:31:04 CEST] <relaxed> I'm meant "it's in unstable now" :) [20:32:16 CEST] <Odyssey> Thank you :) [20:32:45 CEST] <relaxed> Debian will include it in the next release [20:57:41 CEST] <postmodern> what is the precision of dvd chapter start-times? [20:57:48 CEST] <postmodern> they seem to always end in 0 or 500 ms [22:26:37 CEST] <Eduardo_1> hello, anybody knows how to make 2 pass encoding using nvenc ? [22:48:00 CEST] <scottcall> Hello! I have a quick Is this possible question for you all. I have a video that I need to replace about 1 minute of in the middle. I know I can take the original, cut it into two pieces and then concat it with the new middle part but I was hoping to do it all in one swoop with select filters and a concat. While the first segment (from file A) gets put into the final file just fine, but when the 2nd segment (from the new file) is append [22:48:01 CEST] <scottcall> I get buffer overflow errors. My questions is, is this possible or should I just go ahead the split the file up before running a second command to stitch them back together? [22:51:42 CEST] <scottcall> Okay will do. I just trimmed my command down to remove some of my complications I added in debugging Ill put the pastebin link up in a sec [22:54:44 CEST] <scottcall> http://pastebin.com/C81d0XVj [23:01:44 CEST] <llogan> does the output not play correctly? [23:05:22 CEST] <scottcall> llogan- the output plays out, but only contains the first segement, So the first 2773 frames are copied from input 0 to the output but where input 1 should come in is when the right before the overflow errors start. Ive let it sit for 5-10 minutes of errors to see if it picks back up but it doesnt. Ive even tried just using two different select filters from the same input and concating them but they have the same issue, buffer queue overflow [23:05:22 CEST] <scottcall> soon as the first select segment is complete. Ive used select for a while to split a file into pieces and multiple selects with multiple outputs work just fine, just seems trying to skip the step of writing them out causes problems. [23:09:32 CEST] <scottcall> I did eventually get [output stream 0:0 @ 0x1d296a0] 100 buffers queued in output stream 0:0, something may be wrong. after 10 minutes or so and then had to kill -9 the process because it wouldnt respond to the q command [23:10:39 CEST] <xdexter> Hello, Hello, I installed ffmpeg through rpmfusion, when I check the version appears to min "ffmpeg version 12:10:15", however I see that the site shows version 2.7, is that correct? [23:11:04 CEST] <llogan> scottcall: i had to do something similar lately for a client's prores files. i was just lazy and stream copied the segments to individual output files, then used concat demuxer, but i re-encoded the final output instead of stream copied. [23:11:48 CEST] <scottcall> Thats pretty much where I was at, I figured if I could streamline the process great, but I realize Im asking a lot of it ;-) [23:12:05 CEST] <llogan> xdexter: version 12:10:15 means nothing in FFmpegese. run "ffmpeg" in console and see what version it's using [23:12:56 CEST] <xdexter> llogan, http://pastebin.com/uksDZEFG [23:13:03 CEST] <llogan> scottcall: another idea is to try -ss and -t as input options for each segment instead of (a)trim with concat filter. i've never tried that though. [23:14:10 CEST] <llogan> xdexter: 0.10.15. that's an old branch. 2.7 is most recent release, but general users are recommended to use a build from current git master. releases are basically for distros and other distributors. [23:14:21 CEST] <llogan> http://johnvansickle.com/ffmpeg/ [23:14:24 CEST] <scottcall> -ss isnt frame accurate (at least in my experience) and since Im subbing in a 1:1 match (basically the same video with different localized credits) I need that precision. [23:14:49 CEST] <llogan> i was hoping you didn't need such accuracy [23:15:17 CEST] <scottcall> Its okay, Ill just live with temp files and tell the department requesting the feature they need to allow a little more time. I really do apreciate you taking a look tho [23:15:30 CEST] <xdexter> llogan, hmm, ok [23:15:42 CEST] <llogan> or https://trac.ffmpeg.org/wiki/CompilationGuide/Centos [23:16:41 CEST] <xdexter> llogan, Through some rpm package I can not use the latest version? [23:16:50 CEST] <llogan> not that i know of [23:17:48 CEST] <xdexter> llogan, Ok, thankyou [00:00:00 CEST] --- Fri Jun 26 2015

1 0