Ffmpeg-devel-irc

Download

ffmpeg-devel-irc@ffmpeg.org

July 2016

1 participants
62 discussions

ffmpeg-devel.log.20160725
by burek 26 Jul '16

26 Jul '16

[00:01:51 CEST] <JEEB> jamrial: < thresh> I'm going to find some courage and upgrade Albiero (which hosts git, mail, dns) to jessie tonight. [00:01:56 CEST] <JEEB> from #videolan [00:02:20 CEST] <JEEB> and then just a few moments ago < thresh> rebooting [00:02:32 CEST] <JEEB> so I would expect there to be a sitrep soon on how it either succeeded or failed [09:03:13 CEST] <bp0> michaelni, I have a patch to fix #5727 in af_hdcd, but it is kindof big as the functions had to be remade to process both channels together. [09:06:53 CEST] <bp0> I submitted an unrelated set yesterday, but they are not yet applied. Should I submit the new patch against master or after those patches? [10:33:01 CEST] <ubitux> OT but vlc doesn't support ffmpeg>2.8? [10:33:40 CEST] <funman> ubitux: why not? [10:33:57 CEST] <ubitux> dunno, it seems the only package that requires ffmpeg 2.8 on my machine is vlc [10:34:08 CEST] <ubitux> maybe some packager madness but that's curious [10:34:46 CEST] <funman> maybe vlc stable wasn't updated to some new API then [10:34:49 CEST] <funman> git it from git :) [10:34:51 CEST] <funman> get* [10:35:10 CEST] <ubitux> (http://sprunge.us/BBJA) [10:35:44 CEST] <ubitux> funman: i don't actually use vlc, it was pulled when i tried some phonon qt4 stuff [10:50:40 CEST] <ubitux> server is down? [10:50:45 CEST] <ubitux> ssh: connect to host source.ffmpeg.org port 22: Network is unreachable [10:51:21 CEST] <ubitux> (git pull) [10:52:45 CEST] <funman> yes [10:53:54 CEST] <ubitux> :( [10:54:13 CEST] <ubitux> Raz-: ping? [10:54:51 CEST] <nevcairiel> the videolan server is down where git is hosted, not one of ours [10:55:11 CEST] <nevcairiel> unless git moved and noone told me [10:56:38 CEST] <ubitux> ah, my bad then [10:56:58 CEST] <nevcairiel> apparently it went through upgrades yesterday, presumably follow-up maintenance [10:57:01 CEST] <nevcairiel> or everything died [10:57:03 CEST] <nevcairiel> who knows! [10:57:12 CEST] <andrey_turkin> who cares, it's git ) [11:26:26 CEST] <JEEB> nevcairiel: yeah it failed :/ [11:26:33 CEST] <JEEB> they're trying to fix it as we speak [14:26:28 CEST] <funman> kierank: what do you think of extending swresample to take fractional samplerate, then adding speexdsp to libswresample (like is done for soxr) [14:26:46 CEST] <kierank> don't see the point [14:26:57 CEST] <kierank> I'd rather use speexdsp directly [14:27:23 CEST] <funman> less upipe code to write ^_^ [14:27:34 CEST] <kierank> well you'd have to write the drift code anyway [14:27:39 CEST] <funman> true [16:22:21 CEST] <Raz-> ubitux, pong [16:22:32 CEST] <Raz-> big delays today :-) [16:22:43 CEST] <ubitux> Raz-: i thought it was an ffmpeg.org issue, but it was actually videolan [16:22:47 CEST] <ubitux> sorry for the hilight :p [16:50:46 CEST] <ubitux> videolan seems up again [16:59:01 CEST] <jamrial> yeah [17:06:27 CEST] <BBB> hey guys, is there still a reason that allocating a new AVFrame zeroes the contents? [17:06:30 CEST] <BBB> can we get rid of that? [17:20:13 CEST] <durandal_1707> BBB: nope [17:20:30 CEST] <BBB> nope, no reason? or nope, we cannot get rid of it? [17:25:14 CEST] <jamrial> BBB: av_frame_alloc() sets default values after zeroing the whole thing. isn't that the correct thing to do? [17:26:59 CEST] <BBB> isnt it the job of the decoder to fill frame contents instead of av_frame_alloc? [17:27:56 CEST] <nevcairiel> it gives you one with every field properly initialized, which isnt necessarily zero for all fields, and not all decoders write every single field [17:28:02 CEST] <nevcairiel> not sure whats wrong with that [17:32:06 CEST] <cone-568> ffmpeg 03James Almer 07master:26cb7232c343: Revert "lavc: always build dnxhddata" [17:32:06 CEST] <cone-568> ffmpeg 03James Almer 07master:0c75bd8e3cd7: avcodec/dnxhddata: move avpriv_dnxhd_parse_header_prefix to a header [17:33:10 CEST] <jamrial> oh nice, gitweb was updated [17:33:35 CEST] <jamrial> this version hightlights changes to each line [17:38:46 CEST] <nevcairiel> it was probably running on the version that shipped with debian for ever [18:09:13 CEST] <michaelni> jamrial, speaking of gitweb versions, is the version on our box (https://git.ffmpeg.org/ffmpeg.git) new enough ? iam asking as mplayer&rtmpdump and any other stuff have probably only that [18:10:11 CEST] <ubitux> we have word highlight apparently so i guess that's fine [18:10:42 CEST] <ubitux> i wonder if they handle word highlights with splits (multiple matches per line) [18:44:11 CEST] <jamrial> michaelni: as ubitux it has word highlight so it's new enough [18:44:30 CEST] <jamrial> michaelni: you may want to disable the snapshot functionality if bandwidth is a concern, though [18:44:50 CEST] <jamrial> it's disabled in git.videloan.org [18:48:36 CEST] <michaelni> i dont think its a problem, it wasnt one till now [18:49:58 CEST] <jamrial> alright [18:52:41 CEST] <cone-568> ffmpeg 03James Almer 07release/2.5:5141a0ceafd3: avformat/oggparsevp8: fix pts calculation on pages ending with an invisible frame [18:52:42 CEST] <cone-568> ffmpeg 03James Almer 07release/2.6:ebd4d7189c2a: avformat/oggparsevp8: fix pts calculation on pages ending with an invisible frame [18:52:43 CEST] <cone-568> ffmpeg 03James Almer 07release/2.7:470bdd397f2a: avformat/oggparsevp8: fix pts calculation on pages ending with an invisible frame [18:52:44 CEST] <cone-568> ffmpeg 03James Almer 07release/2.8:e9b7ae08d621: avformat/oggparsevp8: fix pts calculation on pages ending with an invisible frame [18:52:45 CEST] <cone-568> ffmpeg 03James Almer 07release/3.0:21c36d83f850: avformat/oggparsevp8: fix pts calculation on pages ending with an invisible frame [18:52:46 CEST] <cone-568> ffmpeg 03James Almer 07release/3.1:7cab4142c541: avformat/oggparsevp8: fix pts calculation on pages ending with an invisible frame [18:53:09 CEST] <jamrial> mmh, just got a bunch of "remote: fatal: unknown change class '|' in --diff-filter=A|C|M|R" error when pushing that [18:53:38 CEST] <jamrial> four per branch [18:57:42 CEST] <jamrial> something may have gone wrong when updating git and/or gitweb in videolan's box it seems [18:58:09 CEST] <jamrial> look at the history of almost every file in the tree. it will show every single merge commit, even if they didn't touch the file in question [18:58:42 CEST] <JEEB> jamrial: I recommend poking thresh @ #videolan [19:00:21 CEST] <jamrial> although the same seems to be happening in https://git.ffmpeg.org/ffmpeg.git, so not sure [19:23:12 CEST] <BBB> nevcairiel: Id like the data to be uninitialized (the buffer contents, as in, the pixel data) [19:23:23 CEST] <BBB> nevcairiel: since the decoder will decoder pixels [19:23:34 CEST] <BBB> nevcairiel: its about 0.2% of runtime so its a slight speedup [19:23:43 CEST] <BBB> nevcairiel: (for all decoders) [19:24:30 CEST] <nevcairiel> av_frame_alloc creates an empty frame, it doesnt allocate buffers or init them [19:25:03 CEST] <nevcairiel> it just fills AVFrame, the struct [19:26:27 CEST] <nevcairiel> if that takes 0.2% of all time then i would be surprised, but its also not something we should change, a freshly allocated struct should not contain uninitialized data hoping someone fills it [19:27:47 CEST] <BBB> something in vp9.c is calling av_buffer_allocz [19:27:53 CEST] <BBB> maybe its the segmentation buffer? [19:28:30 CEST] <nevcairiel> well buffer allocation is something else [19:28:52 CEST] <nevcairiel> normal frame buffers are not zeroed from what i can tell [19:42:39 CEST] <cone-568> ffmpeg 03Kacper MichajBow 07master:e947b75b1c76: libavformat/rtpdec_asf: zero initialize the AVIOContext struct [19:42:40 CEST] <cone-568> ffmpeg 03Josh de Kock 07master:384251daffb9: lavd/libdc1394: distinguish between enumeration errors and no cameras found [19:45:21 CEST] <jamrial> michaelni: did you get the errors i mentioned above when pushing that? [19:45:41 CEST] <michaelni> jamrial, yes [19:45:50 CEST] <jamrial> ok, thanks [19:45:53 CEST] <michaelni> "remote: fatal: unknown change class '|' in --diff-filter=A|C|M|R" [19:46:08 CEST] <jamrial> already mentioned it in #videolan, but thresh seems to be away [19:49:08 CEST] <omerjerk> I'm adding a new muxer [19:49:46 CEST] <omerjerk> I added a new entry in allformats.c [19:49:54 CEST] <omerjerk> but getting error - error: use of undeclared identifier 'CONFIG_ALS_MUXER' [19:50:05 CEST] <omerjerk> where do I need to define it? [19:50:39 CEST] <jamrial> you need to rerun configure after adding the entry in allformats.c [19:51:38 CEST] <omerjerk> oh okay [20:07:11 CEST] <bp0> michaelni, if I have a new patch to submit to the ML, but already have some un-applied patches waiting there, should I make the new patch against master or after the other ones? [20:08:24 CEST] <Chloe> bp0: if the new patch doesn't require the other patches then do it against master imo [20:08:36 CEST] <michaelni> bp0 is there a overlap between the patches ? if not then it wouldnt matter [20:09:01 CEST] <bp0> there is an overlap [20:09:32 CEST] <bp0> they are all for the same file, af_hdcd.c [20:09:54 CEST] <nevcairiel> well not everything touches the same things in those files [20:15:23 CEST] <bp0> ok, hmm [20:15:50 CEST] <bp0> well, I will submit and see, I can always resubmit [21:27:47 CEST] <BBB> michaelni: weve had this discussion about ME for encoding and true motion before, you know Im not a dummy and I fundamentally disagree [21:28:03 CEST] <BBB> michaelni: I dont think we (ffmpeg, as a project) should have multiple ME stacks in our tree [21:28:24 CEST] <BBB> its one thing to play around for academics but that code should never be merged as-is [21:28:40 CEST] <BBB> michaelni: Im just trying to set my expectations so that you dont go down a rabbit hole and end up wandering why I objected [21:29:21 CEST] <ubitux> (down a Big Buck Bunny hole) [21:29:23 CEST] <ubitux> (sorry) [21:30:55 CEST] <BBB> :) [21:31:08 CEST] <michaelni> BBB, ill resign as mentor and leave this to you [21:31:17 CEST] <michaelni> in fact iam not mentor of this one [21:31:22 CEST] <BBB> who is mentor? [21:36:28 CEST] <michaelni> paul is [21:37:01 CEST] <michaelni> BBB, also i think you should not fight this fight on the stuents back [21:37:28 CEST] <BBB> possibly [21:37:56 CEST] <BBB> can we at least agree to not make ffmpeg.git a dumping ground for all code ever written by people competent or not? [21:42:35 CEST] <durandal_1707> what's issue? [21:47:22 CEST] <iive> clean design [21:48:07 CEST] <durandal_1707> everything is local to lavfi [21:53:05 CEST] <durandal_1707> so I don't get it [22:40:13 CEST] <cone-568> ffmpeg 03James Almer 07master:fb91850fe11f: avutil/frame: access avframe fields directly in get_frame_defaults() [00:00:00 CEST] --- Tue Jul 26 2016

1 0

ffmpeg.log.20160725
by burek 26 Jul '16

26 Jul '16

[00:32:25 CEST] <torula> does anyone know how to use timeout with ffmpeg? [00:32:29 CEST] <torula> it just hangs for me. [00:32:36 CEST] <torula> $ timeout 5s ffmpeg .... [00:33:03 CEST] <torula> sometimes my ffmpeg command to generate a poster hangs and won't exit. so i just want to timeout and kill the process if that happens. [11:20:05 CEST] <k_sze[work]> source.ffmpeg.org (albiero.videolan.org) is down? [11:25:57 CEST] <JEEB> k_sze[work]: yes [11:26:03 CEST] <JEEB> thresh from videolan was upgrading it last night [11:26:06 CEST] <JEEB> it... failed :P [11:26:14 CEST] <k_sze[work]> oh noes [11:26:55 CEST] <JEEB> well, they're working on fixing it [11:27:02 CEST] <JEEB> use the github mirror if you require it [12:22:42 CEST] <sagax> hi all! [12:22:53 CEST] <sagax> how to change output device for ffplay? [13:48:00 CEST] <osiris> hi all [13:48:42 CEST] <osiris> is there anywhere a solution for never kernel then 3.14.5 with intel quick sync? they have to be kidding [14:12:45 CEST] <Guest_83746> Allah is doing [14:12:53 CEST] <Guest_83746> sun is not doing Allah is doing [14:12:54 CEST] <__jack__> again ?! [14:16:12 CEST] <iive> should i ban him? [14:16:24 CEST] <sagax> how to change output device for ffplay? [14:17:49 CEST] <__jack__> iive: you can [14:18:01 CEST] <__jack__> that crap robot spammed us on #btrfs [14:18:26 CEST] <__jack__> saying wild stuff, like "dollars are alive, they are allah or whatever" [14:19:02 CEST] <iive> __jack__: and recommendations for the ban regex? aka, should I ban all web gateways? [14:20:02 CEST] <__jack__> they said it is using dynamic IP, already tried to ban him, didn't work, I don't really know how is the right way [14:20:59 CEST] <durandal_1707> just kick? [14:22:39 CEST] <iive> the IP seems to be from pakistan, I guess i'll just block web irc for the whole range [14:22:55 CEST] <bencoh> :/ [14:23:33 CEST] <osiris> do u think its smart to announce how u do it even if he is still able to read it? [14:24:44 CEST] <furq> do you not think it would be a good idea to kick him as well [14:24:54 CEST] <iive> i don't mind blocking all gateways... [14:32:38 CEST] <samb1> hi I've got the latest checkout of the ffmpeg git, I ran ./configure --enable-libmp3lame --enable-avresample --enable-shared and got ERROR: libmp3lame >= 3.98.3 not found , however I have 3.99 installed (am on ubuntu precise) [14:33:08 CEST] <osiris> also the -dev package? [14:33:14 CEST] <samb1> ahh [14:33:58 CEST] <samb1> thanks :) that fixed it [14:34:17 CEST] <osiris> no worries [14:35:25 CEST] <osiris> samb1: did u run apt-get build-dep ffmpeg? [14:35:42 CEST] <osiris> that should have installed everything as needed [14:35:57 CEST] <furq> on debian/ubuntu that'll build the libav*-dev packages, which you probably don't want [14:36:04 CEST] <furq> s/build/install/ [14:36:30 CEST] <furq> as well as all the external libraries that the debian ffmpeg package uses, which you probably also don't want [14:37:06 CEST] <osiris> so u build every dependency from source? [14:37:37 CEST] <furq> you definitely don't want system libav*-dev packages installed if you're building with --enable-shared because it could cause linker issues [14:37:37 CEST] <samb1> osiris: no I checked out the ffmpeg git repo with the aim of building the whole thing from source because a few things are out of date on ubuntu precise [14:37:53 CEST] <furq> for external libraries i just install the dev packages i want from apt [14:37:58 CEST] <samb1> just confused a bit between which things were deps and which things I was building I think [14:38:29 CEST] <osiris> yeah, i stumbled over that old shit as well, but i dont get quick sync working anyways with a needed centos 7.1 or kernel < 3.15 [14:39:34 CEST] <osiris> dont get why intel is lacking so much support on that gr8 feature [14:39:49 CEST] <samb1> I wanted a lot of the newer libraries because I'm working with this essentia thing (see http://essentia.upf.edu/ ) and it has certain version requirements for theextractors [14:40:18 CEST] <samb1> so eh building the stuff fixed a lot of things ... maybe a lot of what am currently doing is superfluous tho [14:41:32 CEST] <osiris> samb1: looks like most is not needed, if your just going for audio ;) [14:41:49 CEST] <osiris> there is really a reason, why i mostly use arch in the meanwhile [14:42:07 CEST] <samb1> yeh those tags I put in the configure should get me what I want but sure it makes a whole load of other stuff too, the build takes ages [14:42:20 CEST] <furq> it'll take longer to go through and figure out what you don't need than it will to just compile it all [14:42:28 CEST] <samb1> yeh probs [14:42:33 CEST] <furq> also don't forget to run make -j [14:42:49 CEST] <osiris> yeah, -j is the key [14:43:17 CEST] <osiris> only make uses just one core [14:43:37 CEST] <samb1> not used the j option... I must be missing out [14:44:12 CEST] <osiris> furq: how u get a complete list of the denpendcies to install them manually? i was just getting one dependency after the other [14:44:17 CEST] <osiris> what took me ages [14:44:50 CEST] <osiris> samb1: then cancel and do make clean && make -j 8 [14:44:59 CEST] <furq> you don't need to make clean [14:45:04 CEST] <furq> make -j8 will pick up where the previous make left off [14:45:13 CEST] <furq> or -j4 or whatever [14:45:50 CEST] <furq> i've never had a hard time finding deps but i usually don't build with many [14:46:22 CEST] <furq> except for mingw, and i have a script which does all that for me [14:48:03 CEST] <samb1> it had compelted but yeh that's faster [14:48:08 CEST] <samb1> completed [15:56:48 CEST] <Kadigan_KSB> Hey all. I was wondering -- I have a spherical video clip (square video frame containing spherical content). Can I use ffmpeg to convert it to equirectangular video? [15:58:32 CEST] <Kadigan_KSB> (this would happen, I assume, by unwrapping and then stretching the video) [15:59:12 CEST] <Kadigan_KSB> When I process the video in the designated app for it (360fly Director), I input 1504x1504 video files and receive 2560x1280 files. However, the app in question cannot do (or I do not know how to do) batch encodes. [16:13:38 CEST] <bencoh> Kadigan_KSB: lenscorrection filter, maybe? [16:17:11 CEST] <durandal_1707> Kadigan_KSB: see panorama filter on ml archives [16:18:45 CEST] <Kadigan_KSB> Okay, I'll check it out. [22:18:35 CEST] <yuken> Is there any way to use a graphics card (more specifically, an nVidia GPU - 750Ti) to accelerate compression/encoding with libx264 or libx265? Under Windows. [22:19:53 CEST] <durandal_1707> no [22:21:00 CEST] <iive> libx264 support look-ahead pass using opencl, so any recent card could use it. [22:21:42 CEST] <iive> it doesn't provide significant speed up and it kind of lowers quality a bit. [22:22:30 CEST] <flux> does it even reduce CPU load? [22:23:50 CEST] <iive> I do suspect that you might be interested in hardware encoder present in new cards, this however does not involve libx264/5 at all. it's done enitrely by the driver/hardware. [22:24:26 CEST] <yuken> iive, do you mean NVENC and VCE? [22:24:50 CEST] <iive> yeh... probably :) [22:25:06 CEST] <yuken> Videos are way too large with the former, can't use the latter as my card is nVidia. [22:27:58 CEST] <furq> there's also qsv on any recent intel cpu, but it suffers from the same problem as nvenc [22:28:08 CEST] <furq> which is the problem of being a rubbish h.264 encoder [22:36:08 CEST] <JEEB> well the use cases are somewhat different [22:37:07 CEST] <JEEB> software transcoders tend to aim for high compression as well as try to optimize themselves as much as possible for CPUs, while the pre-built encoding chips are aimed for low-latency and keeping at least realtime if not much more with... some sort of compression [22:37:29 CEST] <JEEB> for example nvidia's private APIs for screen capture are quite useful (too bad they are private) [22:46:37 CEST] <Mavrik> *sigh* [22:46:44 CEST] Action: Mavrik found himself in need of a fast JPEG2000 decoder. [22:47:55 CEST] <JEEB> rip [22:48:04 CEST] <durandal_1707> ffmpeg one is fast if you have bunch of CPUs [22:48:28 CEST] <JEEB> I had fun when transcoding the tears of steel master on a quad core [22:48:36 CEST] <JEEB> (actual 4K) [22:53:08 CEST] <cbsrobot> Mavrik: did you try http://grokimagecompression.github.io/grok/ ? [22:53:46 CEST] <Mavrik> durandal_1707, yeah, I have mobile CPUs :P [22:53:51 CEST] <Mavrik> cbsrobot, nop, never heard of it [22:54:07 CEST] <Mavrik> Oh, AGPL :D [23:01:01 CEST] <Ryfi> anyone know why my ffmpeg instance will continuously use more and more cpu until it's restarted? http://pastebin.com/LjspfUbC [23:03:58 CEST] <Ryfi> chttp://puu.sh/qeg9W/8529b8a5b5.png [23:08:06 CEST] <durandal_1707> Ryfi: does it happen without loudnorm? [23:11:43 CEST] <Ryfi> ill let you know [23:21:17 CEST] <Ryfi> seems that loudnorm is the culprit, which is unfortunate since that's the whole reason why im using ffmpeg .. ugh lol [23:25:29 CEST] <durandal_1707> Ryfi: could you open bug report? [23:26:04 CEST] <Ryfi> actually i think im wrong [23:26:18 CEST] <Ryfi> its still increasing slowly.. its really slow so its hard to tell unless i wait [23:27:41 CEST] <Ryfi> http://puu.sh/qehQw/9f1eaa706b.png [23:51:53 CEST] <Ryfi> submitted: https://trac.ffmpeg.org/ticket/5735 [00:00:00 CEST] --- Tue Jul 26 2016

1 0

ffmpeg-devel.log.20160724
by burek 25 Jul '16

25 Jul '16

[00:44:30 CEST] <cone-078> ffmpeg 03Kacper MichajBow 07master:69630f4d304a: libavutil/opt: Small bugfix in example. [03:56:01 CEST] <cone-778> ffmpeg 03Michael Niedermayer 07master:783a2568b28d: tests/fate/filter-audio: fate-filter-chorus: Randomly change parameters to some values which pass on arm/mips/x86-32/64 [05:21:23 CEST] <bp0> is there way a fetch the original message raw from this? http://ffmpeg.org/pipermail/ffmpeg-devel/2016-July/196654.html [05:21:40 CEST] <bp0> something like wget ...?raw or similar [05:56:37 CEST] <Compn> you can download the entire month archive in tgz [05:56:51 CEST] <Compn> or possibly pull raw from gmane ffmpeg-devel group [05:56:59 CEST] <Compn> but i dont think pipermail provides what you want, bp0 [05:57:26 CEST] <bp0> alright [05:57:28 CEST] <bp0> thanks [05:59:22 CEST] <Compn> if you want to test patches , theres probably some project that does kinda what you want [06:00:03 CEST] <Compn> or ask fellow devels what setup they use for easily taking mail patches and applying them to git trees [06:00:12 CEST] <Compn> probably just a keystroke in certain mail clients [06:00:51 CEST] <Compn> or if you just want mail headers, pretty sure gmane can do that [06:00:52 CEST] Action: Compn afk [06:06:04 CEST] <bp0> wget -O fatepatch.txt http://article.gmane.org/gmane.comp.video.ffmpeg.devel/216794/raw [06:06:11 CEST] <bp0> git am <fatepatch.txt [06:06:16 CEST] <bp0> worked, thanks Compn [11:48:22 CEST] <durandal_1707> Compn: when you gonna add 3 codecs to mplayer? [12:20:53 CEST] <cone-905> ffmpeg 03Clément BSsch 07master:5c14018fc4f6: lavfi/curves: do not automatically insert points at x=0 and x=1 [12:20:53 CEST] <cone-905> ffmpeg 03Clément BSsch 07master:4a8f5f1fd83d: lavfi/curves: add plot option [12:20:53 CEST] <cone-905> ffmpeg 03Clément BSsch 07master:51a873d44114: lavfi/curves: remove pointless logging since the addition of plot option [12:20:53 CEST] <cone-905> ffmpeg 03Clément BSsch 07master:4eee06ae8730: lavfi/curves: add various const where it makes sense [12:20:53 CEST] <cone-905> ffmpeg 03Clément BSsch 07master:050f790594ee: lavfi/curves: dynamically allocate LUTs [12:20:54 CEST] <cone-905> ffmpeg 03Clément BSsch 07master:f19f5b906d70: lavfi/curves: move alloc and init of LUTs inside config_input() [12:20:54 CEST] <cone-905> ffmpeg 03Clément BSsch 07master:e30cdac14b55: lavfi/curves: add 16-bit support [12:20:55 CEST] <cone-905> ffmpeg 03Clément BSsch 07master:39c6d4a8c592: lavfi/curves: reindent after previous commit [12:20:55 CEST] <cone-905> ffmpeg 03Clément BSsch 07master:62a31aecf607: lavfi/curves: pass log ctx as void* instead of AVFilterContext* [12:20:56 CEST] <cone-905> ffmpeg 03Clément BSsch 07master:b470d81f4e8e: lavfi/curves: prefix init and uninit function names [12:22:46 CEST] <cone-905> ffmpeg 03Clément BSsch 07master:6e971f7d66f9: lavfi: bump minor after recent curves filter additions [12:34:34 CEST] <cone-905> ffmpeg 03Mark Reid 07master:6108cb2ce3ee: tests/fate: add dnxhr encoding tests [14:16:28 CEST] <Compn> durandal_1707 : i need a new build of mplayer. or maybe i just have to boot my linux box... [14:16:31 CEST] <Compn> been busy [14:22:45 CEST] <cone-905> ffmpeg 03Clément BSsch 07master:b8aaedcd0147: lavfi/selectivecolor: fix picking black as neutral when alpha is present [15:34:21 CEST] <cone-905> ffmpeg 03Clément BSsch 07master:308f9b1c4944: lavfi/selectivecolor: add 16-bit support [15:49:25 CEST] <jimi> hi ! [15:49:29 CEST] <jimi> hola! [15:50:26 CEST] <ubitux> Compn: i was refering to the virus one we already talked about; and they're mails from the provider, with a custom form communication only i can access; not going to forward anything, it's a pita, pb is "solved" anyway [15:50:47 CEST] <jimi> do yo know how to set an specific lenght when you convert a file? Thanks [15:56:59 CEST] <jimi> its solved thanks [16:21:16 CEST] <Compn> ubitux : oh, the rtmpdump binaries false positive ? thats annoying. [16:21:31 CEST] <Compn> ubitux : sorry didnt mean to annoy you further about it :D [18:50:49 CEST] <cone-722> ffmpeg 03Ivan Uskov 07master:b4054100f675: Revert "Merge commit '3c53627ac17fc6bdea5029be57da1e03b32d265d'" [20:22:06 CEST] <cone-722> ffmpeg 03Carl Eugen Hoyos 07master:9bd35a76fc94: lavc/h264_ps: Be more verbose when truncating likely oversized PPS. [22:51:22 CEST] <mateo`_> michaelni: can I directly include libavcodec/dnxhdata.o from libavformat/Makefile ? [22:51:51 CEST] <michaelni> no [22:52:51 CEST] <michaelni> mateo`_, i think you patch is ok as it is [22:53:12 CEST] <mateo`_> ok [22:53:13 CEST] <nevcairiel> this is why avpriv stuff is evil and should never have been created in the first place :p [22:54:03 CEST] <nevcairiel> also can we all stop responding to Ivan until he fixes his mail client? i literally cannot read his mails without re-formating them first [22:55:01 CEST] <JEEB> yea, that type of responding is not exactly readable [22:57:22 CEST] <michaelni> agree about avpriv, about ivan he is qsv maintainer, ignoring him due to email formating is childish, also i dont think he is the only one who prefixed quotes with initials [22:57:43 CEST] <nevcairiel> he is the only one i have ever seen [22:57:54 CEST] <nevcairiel> and he was asked to change it several times already [22:58:58 CEST] <nevcairiel> (by multiple people [22:58:59 CEST] <nevcairiel> ) [22:59:08 CEST] <cone-722> ffmpeg 03Matthieu Bouron 07master:2adbea4e216c: lavc: always build dnxhddata [23:06:44 CEST] <michaelni> the initials before lines of reply is one of the 3 attribution stiles listed at : https://en.wikipedia.org/wiki/Posting_style [23:08:19 CEST] <michaelni> if this is unwanted on the mailing list then it should be documented where we docuemnt about top posting [23:08:33 CEST] <jhu> trying to compile latest ffmpeg using ./configure --prefix=$HOME --disable-yasm --extra-ldflags='-L/users/jhu/lib' --enable-libmp3lame [23:08:59 CEST] <jhu> but says cannot find lib3lame even though I pointed directly to it using the extra-ldflags above. [23:09:20 CEST] <jhu> I just compiled the latest libmp3lame [23:09:23 CEST] <nevcairiel> michaelni: none of my mail clients properly folds comments on that style, and I'm not the only one to say that, so yes, it should be avoided, especially when its only one person doing it [23:13:42 CEST] <jamrial> jhu: you probably also need to specify the header include path with --extra-cflags [23:13:54 CEST] <jamrial> in any case, this kind of question is for the #ffmpeg channel [23:14:13 CEST] <jhu> ok. will do. sorry for the noise. [23:24:49 CEST] <mateo`_> jamrial: isn't the avpriv_dnxhd function built all the time since it's avpriv ? [23:28:09 CEST] <jamrial> the avpriv functions have been there for months and even years, and the file has never been compiled unconditionally [23:30:20 CEST] <jamrial> it has always been compiled only if either the dnxhd de/encoder, dnxhd demuxer or mxf muxer were enabled [23:34:13 CEST] <mateo`_> jamrial: but there were no dependency from lavformat on it [23:34:35 CEST] <jamrial> as i said, mxf muxer and and dnxhd demuxer [23:34:51 CEST] <jamrial> you'd simply be adding mox muxer to those [23:34:56 CEST] <jamrial> mov [23:35:56 CEST] <mateo`_> jamrial: ok, sorry, I didn't notice that those entries were in *libavcodec*/Makefile [23:38:06 CEST] <mateo`_> I'll revert my commit and submit a new patch [23:44:57 CEST] <jamrial> why was that function added to dnxhddata.c anyway? couldn't it have been added to dnxhddata.h instead to avoid all the avpriv nonsense? [23:51:49 CEST] <cone-722> ffmpeg 03Carl Eugen Hoyos 07master:71167f7f8434: lavc/Makefile: Fix standalone compilation of the svq3 decoder. [23:54:41 CEST] <jamrial> huh, i'm getting 403 errors when trying to check git web from https://git.videolan.org [23:56:23 CEST] <mateo`_> me too [23:57:04 CEST] <cone-722> ffmpeg 03Carl Eugen Hoyos 07release/3.1:2e1be2271506: lavc/Makefile: Fix standalone compilation of the svq3 decoder. [23:58:44 CEST] <mateo`_> jamrial: i'll soon be off, feel free to revert and apply the patch if you want to, otherwise i'llbe back in 10h~. [00:00:00 CEST] --- Mon Jul 25 2016

1 0

ffmpeg.log.20160724
by burek 25 Jul '16

25 Jul '16

[00:00:04 CEST] <KDDLB> I use OS X mainly [00:00:16 CEST] <KDDLB> and Linux in servers I have access to [00:00:38 CEST] <JEEB> win for gaming, lunix for work and servers, OS X when having to deal with OS X development (build servers f.ex.) etc [00:00:45 CEST] <JEEB> s/OS X/iOS/ [00:01:04 CEST] <furq> i use windows, debian and freebsd on a daily basis [00:01:09 CEST] <furq> it's as enjoyable as you'd imagine [00:01:34 CEST] <JEEB> it gets fun enough when you can start sharing some of your dotfiles on windows as well with cygwin or whatever else you like [00:07:37 CEST] <KDDLB> how about __MACOSX and/or .DS_Store [00:07:42 CEST] Action: KDDLB runs away [00:07:55 CEST] <JEEB> those tend to end up in archives from ISO/IEC members [00:08:00 CEST] <KDDLB> yeah [03:33:05 CEST] <mccc> Hello, I'm cross compiling ffmpeg for the raspberry pi from a Debian development VM. I was compiling alsa from source and while getting help earlier someone recommended using the libasound2-dev:armhf and libasound2-data:armhf packages to compile ffmpeg against, and use the libasound2:armhf and libasound2-data:armhf packages on the pi. [03:33:33 CEST] <mccc> I've tried to do that, but I don't think I have alsa support compiling in correctly. [03:33:42 CEST] <furq> the guy who recommended that sounds like a cool guy [03:35:59 CEST] <mccc> If it was you I appreciate it :) [03:36:02 CEST] <mccc> http://pastebin.com/MWDxW1L3 [03:36:21 CEST] <mccc> What I don't think I've been able to find is how to ./configure ffmpeg's build correctly. [03:36:39 CEST] <mccc> ./configure --help doesn't display anything about alsa, [03:37:07 CEST] <mccc> and the guide I was using at https://trac.ffmpeg.org/wiki/CompilationGuide/RaspberryPi#CompilingALSAlibr… [03:37:47 CEST] <mccc> seems to presume an alsa-lib compiled from source, and I'm still not sure what's enabling it (--extra-libs=-ldl? again, ./configure --help doesn't help..) [03:38:10 CEST] <furq> try --enable-indev=alsa --enable-outdev=alsa [03:38:34 CEST] <furq> iirc it's enabled by default but it might give a better error message if you explicitly enable it [03:38:59 CEST] <mccc> Okay, thank you, trying that now. [03:39:50 CEST] <mccc> Also, it looks like the libasound2-dev:armhf installed in /usr/include and /usr/lib/arm-linux-gnueabihf, [03:40:12 CEST] <mccc> is there anything I need to do to get those to ./configure? [03:41:39 CEST] <furq> /usr/include is the default include path [03:41:51 CEST] <furq> you might need --extra-ldflags="-L/usr/lib/arm-linux-gnueabihf" [03:42:03 CEST] <furq> you also don't need the toolchain on that wiki page if you're on debian testing or unstable [03:42:10 CEST] <furq> just inatall gcc-arm-linux-gnueabihf from apt [03:42:15 CEST] <furq> install [03:43:57 CEST] <mccc> Okay, thank you, but I'm on 8.5 (Jessie) [03:44:10 CEST] <furq> never mind then [03:44:17 CEST] <mccc> with --enable-indev and outdev, I do get: [03:44:43 CEST] <mccc> WARMING: /home/user/x-tools/arm-unknown-linux-gnueabihf/bin/arm-unknown-linux-gnueabihf-pkg-config not found, library detection may fail. [03:44:51 CEST] <mccc> WARNING: using libfdk without pkg-config [03:45:02 CEST] <mccc> but otherwise configure ends cleanly. Compiling and testing now. [03:45:16 CEST] <furq> check the list of enabled libs in configure before building [03:45:30 CEST] <furq> you can ignore those warnings but it might still be missing alsa [03:46:48 CEST] <furq> you can install pkg-config:armhf if you want to get rid of the pkg-config warning [03:47:11 CEST] <mccc> would it be listed as just "alsa"? It's not listed. [03:48:32 CEST] <furq> it should show up in "enabled indevs" [03:48:38 CEST] <furq> i just tried it and it works here without --extra-ldflags [03:48:49 CEST] <furq> but i'm using a different toolchain so maybe the lib paths are set up properly [03:49:44 CEST] <mccc> let me try --extra-ldflags="-L/usr/lib/arm-linux-gnueabihf".. [03:49:52 CEST] <furq> that's probably the culprit [03:50:14 CEST] <furq> it looks like it silently disables alsa if libasound2 isn't found [03:54:51 CEST] <mccc> Unfortunately it's still not listed. I do have /usr/lib/arm-linux-gnueabihf/libasound.so and libasound.so.2 and libasound.so.2.0.0. [03:58:57 CEST] <furq> https://wiki.debian.org/CrossToolchains#For_jessie_.28Debian_8.29 [03:58:59 CEST] <furq> maybe try with that toolchain [04:00:26 CEST] <mccc> Thank you. I'll try with that. Also, here's the output of my ./configure: [04:00:27 CEST] <mccc> http://pastebin.com/Hz72qddy [04:00:48 CEST] <mccc> The command I used: [04:02:44 CEST] <mccc> http://pastebin.com/mYZ3WzhJ [04:05:15 CEST] <furq> http://vpaste.net/aD9qk [04:08:31 CEST] <furq> you can install libfdk-aac-dev:armhf as well [04:09:55 CEST] <furq> http://vpaste.net/asYz5 [04:09:58 CEST] <furq> that should be all you need [04:10:20 CEST] <furq> which is surprising because i figured that would need --enable-nonfree [04:10:48 CEST] <mccc> I thought so too - and the license is why I didn't think it would be available from a package and that I'd need to compile it from source. [04:11:09 CEST] <furq> you'll need to enable non-free in apt.conf [04:11:18 CEST] <furq> https://packages.debian.org/jessie/libfdk-aac-dev [04:12:25 CEST] <mccc> Will I do this both in the development environment and in the pi, or will libfdk-aac-dev be compiled statically in to ffmpeg? [04:13:01 CEST] <furq> you'll probably need to install it on the pi as well [04:14:46 CEST] <furq> you can normally use --extra-ldflags="-static" but it looks like the fdk package doesn't provide a static lib [04:16:42 CEST] <mccc> Okay, giving this a run now. [04:16:48 CEST] <mccc> Thank you again. [05:50:03 CEST] <Mista-D> is there a search option for FFmpeg documentation on the website please? [05:57:48 CEST] <mrelcee> well - 24+ hours of encoding no crashes running from the git sources. 3.1.1. crashed every couple hours for me.. same --settings in configuration.. whatever works I'm a happy camper.. :) [09:49:05 CEST] <Guest_84847> Allah is doing [09:49:06 CEST] <Guest_84847> sun is not doing Allah is doing [09:49:10 CEST] <Guest_84847> moon is not doing Allah is doing [09:49:24 CEST] <Guest_84847> stars are not doing Allah is doing [09:49:35 CEST] <Guest_84847> planets are not doing Allah is doing [09:49:52 CEST] <Guest_84847> galaxies are not doing Allah is doing [09:57:31 CEST] <mccc> Hi furq, thank you, using the toolset and .configure command you recommended, I was able to compile and get this to run with libfdk and alsa on the pi. Sincerely appreciate the help :) [10:26:56 CEST] <Fjorgynn> varför? [17:08:18 CEST] <wuOz> hi, is there an option to stream second per second of a file? [18:16:43 CEST] <wuOz> wuOz: It is the -re option with ffmpeg [18:18:59 CEST] <JEEB> also a horrible hack that can break rather easily. so while you can use it and it might work for your use case - make sure you don't get too used to using it (as it can break under your feet for various reasons) [19:31:39 CEST] <livingBEEF> is there somewhere a list of 3.0 api changes? [19:32:45 CEST] <livingBEEF> The realease changelog only mentiones that they're there... [20:06:53 CEST] <JEEB> livingBEEF: there's a file in the repo either in root or under docs [20:07:00 CEST] <JEEB> apisomething [20:08:37 CEST] <livingBEEF> thanks [21:05:10 CEST] <DelphiWorld> YO ffmpegsters [21:24:51 CEST] <kelnoky> hey, I am at a slight loss. I tried to google, but got confusing results and I am under time contraints. I need to convert an m4b file into an mp3 file, but cant really find a way to do it [21:25:42 CEST] <kelnoky> ugh, might have solved it [21:25:44 CEST] <kelnoky> this always happens [21:33:13 CEST] <furq> m4b? [21:37:43 CEST] <kelnoky> yeah, that was my initial problem [21:37:45 CEST] <kelnoky> audiobook format [21:42:44 CEST] <__jack__> do you have a sample ? [21:47:11 CEST] <kelnoky> I did indeed solve it [21:47:26 CEST] <__jack__> ok! [21:47:31 CEST] <kelnoky> it was as simple as 'faad --stdio INPUT.m4b | lame --preset standard - OUTPUT.mp3' [21:47:44 CEST] <furq> it looks like it's just an m4a with chapters [21:47:50 CEST] <furq> ffmpeg opens it fine here [21:47:51 CEST] <kelnoky> the other google results I found just didnt talk about m4bs that's why I was confused [21:47:54 CEST] <kelnoky> yeah [21:48:00 CEST] <kelnoky> it is [22:28:27 CEST] <harrynoob> hi. when using ffmpeg (from Kodi, in a libreelec build; not sure which version, can find if necessary) to play a HLS stream, the probing seems to sort-of-recognize it as such, but then just borks out. here's a bit of logging related to it http://pastebin.com/VFsjZpG8 anyone have an idea what's going on? i see that the # of bytes read is not increasing [22:38:32 CEST] <oni> hello there. has anyone tried to build on osx with --enable-frei0r and managed it successfully? So far I have not [22:38:36 CEST] <oni> ERROR: frei0r.h header not found [22:38:42 CEST] <oni> I bet this is a common thing [22:39:12 CEST] <oni> (that header is currently in /usr/local/include) [23:00:31 CEST] <KDDLB> hmmm [23:01:07 CEST] <KDDLB> why are the default parameters for video encoder h264_videotoolbox (macOS) make it look horrible? :/ [23:04:18 CEST] <JEEB> I was going to answer to that facetiously, but then I thought again. I mean, it's a 3rd party library based thing which uses the same SW encoder as QT if not a hardware encoder (in case such is even supported non !iOS) [23:04:37 CEST] <JEEB> and QT's SW decoder and encoder were well known to be not exactly the best in town [23:04:59 CEST] <KDDLB> hmm [23:05:16 CEST] <furq> is it even appreciably faster than x264 on osx [23:05:52 CEST] <JEEB> so unless you have a licensing requirement that requires you to either get a payware license from x264 LLC or use something that isn't GPL (like x264 by default), I would keep away from the videotoolbox one [23:06:42 CEST] <KDDLB> so, no hardware-accelerated encoding? [23:07:00 CEST] <JEEB> I think on iOS you might get that, but NFI if any of that is on OS X | macOS [23:07:25 CEST] <JEEB> and even then, you don't use HW encoders for the quality [23:07:50 CEST] <JEEB> you use them because your CPU can't handle jack shit and/or because you cannot distribute an AVC decoder and/or encoder for some reason [23:08:04 CEST] <KDDLB> I have a Mac mini [23:08:33 CEST] <KDDLB> and I gave up, I just used -vcodec prores [23:08:37 CEST] <KDDLB> :D [23:08:44 CEST] <JEEB> so for which use is this anyways? [23:09:16 CEST] <KDDLB> testing yadif [23:09:18 CEST] <JEEB> you might want to look into ut video if you're planning on opening it in a video editor or something [23:09:53 CEST] <JEEB> because I see zero reason to use prores unless you specifically want lossy crapola out of your thing [23:10:07 CEST] <JEEB> in which case you might as well use -c:v libx264 -preset ultrafast -q:v 0 [23:10:09 CEST] <JEEB> which is lossless [23:10:49 CEST] <JEEB> of course lossless AVC isn't open'able in commercial editors [23:11:00 CEST] <JEEB> so if that's what you want, then you will want to use something like Ut Video [23:11:21 CEST] <KDDLB> it was mostly for testing [23:11:33 CEST] <furq> KDDLB: is h264_videotoolbox actually hitting the cpu [23:11:42 CEST] <KDDLB> not sure [23:11:46 CEST] <furq> googling suggests it might be using quicksync if you have a recent enough cpu [23:11:53 CEST] <furq> although i can't find anything concrete [23:11:54 CEST] <JEEB> if you want to test how much using yadif slows you down, then you use `-f null -` [23:12:08 CEST] <JEEB> which skips everything after input and filtering [23:12:22 CEST] <JEEB> just pushes shit to /dev/null [23:12:30 CEST] <furq> if that is the case then the answer is because quicksync is a bad h264 encoder [23:12:47 CEST] <JEEB> that way you can kind of check how much is the speed with just decoding + yadif [23:12:55 CEST] <JEEB> instead of having some sort of encoding in the mix [23:13:04 CEST] <jhu> trying to compile latest ffmpeg using ./configure --prefix=$HOME --disable-yasm --extra-ldflags='-L/users/jhu/lib' --enable-libmp3lame [23:13:10 CEST] <JEEB> (you still haven't defined WTF you want to actual achieve and thus it's really hard to recommend anything at all) [23:13:19 CEST] <KDDLB> dont worry [23:13:30 CEST] <jhu> but says cannot find lib3lame even though I pointed directly to it using the extra-ldflags above. [23:13:40 CEST] <JEEB> see what config.log says [23:13:44 CEST] <KDDLB> Im just experimenting, JEEB [23:13:53 CEST] <JEEB> and I'm trying to help you experiment [23:14:12 CEST] <furq> jhu: you need to pass the path to lame/lame.h [23:14:13 CEST] <JEEB> also if you are on x86 or x86_64 please explain why on earth would you not want the hand-built optimizations to not be built in [23:14:17 CEST] <furq> presumably with --extra-cflags [23:14:30 CEST] <JEEB> if he built LAME properly there should be a pkg-config file, no? [23:14:45 CEST] <furq> if there was a pkg-config file then presumably he wouldn't need extra-ldflags [23:14:46 CEST] <JEEB> and then if he sets PKG_CONFIG_PATH=/your/prefix/lib/pkgconfig it should JustWork [23:14:54 CEST] <furq> but yeah that is the actual correct way to solve this [23:14:54 CEST] <JEEB> furq: don't expect people to know about that stuff [23:15:17 CEST] <JEEB> I've seen enough people trying to run circles around issues where pkg-config would have helped them [23:15:22 CEST] <furq> sadly i don't expect people to know about pkg-config [23:15:30 CEST] <KDDLB> !! [23:15:38 CEST] <KDDLB> h264_videotoolbox is using the CPU actually [23:15:39 CEST] <JEEB> + some people don't even set a prefix and `make install` [23:15:44 CEST] <JEEB> KDDLB: then it's the QT one [23:15:46 CEST] <jhu> extra-cflags worked! [23:15:47 CEST] <JEEB> which is also shit [23:15:49 CEST] <furq> well the prefix should be /usr/local anyway [23:15:54 CEST] <jhu> thanks for the quick response. [23:16:08 CEST] <JEEB> jhu: do note my comment regarding --disable-yasm [23:16:13 CEST] <furq> i've never been so unhappy to be right [23:16:13 CEST] <JEEB> if you are building for intel arch [23:16:41 CEST] <JEEB> if you are building for !intel arch then it shouldn't require disable-asm to begin with, though [23:18:25 CEST] <jhu> I don't have root access. [23:18:46 CEST] <JEEB> so? [23:18:58 CEST] <JEEB> build yasm, add to your PATH [23:19:09 CEST] <JEEB> just like you're doing with whatever you're doing right now [23:19:19 CEST] <jhu> ok. let me see about that. [23:19:36 CEST] <JEEB> and this is if and only if yasm isn't packaged for your OS [23:22:08 CEST] <YLivay> hello, does anyone know why ffmpeg complains about invalid output format when I specify "-f png" but if I just slap a ".png" to the destination file it outputs png just fine? [23:23:11 CEST] <JEEB> because there is no such container called "png" in FFmpeg [23:23:26 CEST] <JEEB> see `ffmpeg -formats |grep "png"` [23:23:49 CEST] <YLivay> I saw that, but I still don't understand why it can't output it if it actually does S: [23:24:10 CEST] <JEEB> because it either uses the raw data muxer or the apng one most probably? [23:24:23 CEST] <YLivay> ffmpeg -i /path/to/input -vframes 1 /path/to/output.png works for example [23:24:48 CEST] <JEEB> yes, because there's a decoder and encoder for that *video* format (as opposed to a "container" format) [23:25:01 CEST] <JEEB> -f sets containers, -c sets video|audio|subtitle formats [23:25:18 CEST] <JEEB> see `ffmpeg -codecs |grep "png"` [23:25:19 CEST] <YLivay> Ahhh I see [23:25:39 CEST] <JEEB> you could see from the command line output which muxer the png encoding actually uses [23:26:39 CEST] <YLivay> Lavc56.26.100 png is that it? [23:27:16 CEST] <JEEB> you can just post the whole output on a pastebin if you don't know what to parse of it, and link here [23:28:26 CEST] <harrynoob> never mind, my issue seems to be non-ffmpeg related but instead entirely caused by Kodi, as running ffmpeg standalone works fine [23:30:28 CEST] <YLivay> http://pastebin.com/PsgUU44u [23:30:46 CEST] <JEEB> Output #0, image2, to 'out.png': [23:30:48 CEST] <YLivay> this doesnt make much sense to me, is encoder = muxer? [23:30:53 CEST] <JEEB> so it's using a muxer called image2 [23:31:03 CEST] <JEEB> and the video format is png [23:31:31 CEST] <YLivay> so -c png for the container and -f would be image2? [23:31:57 CEST] <JEEB> YLivay: muxer is short for multiplexer, you put streams (video|audio|subtitle) into it and get some container format with those stream(s) [23:32:06 CEST] <JEEB> encoder is generally the part that generates streams [23:32:19 CEST] <JEEB> usually from raw data that is provided by a decoder [23:32:30 CEST] <JEEB> demuxer=>decoder=>encoder=>muxer [23:32:34 CEST] <JEEB> is the usual flow [23:32:50 CEST] <YLivay> Oh, thanks a lot man! [23:33:37 CEST] <JEEB> and no, -c sets the decoder or encoder (decoder before the input (-i) and encoder after it) [23:33:54 CEST] <JEEB> and -f sets the demuxer or muxer (same thing regarding before/after input) [00:00:00 CEST] --- Mon Jul 25 2016

1 0

ffmpeg-devel.log.20160723
by burek 24 Jul '16

24 Jul '16

[00:46:25 CEST] <kode54> https://gist.github.com/kode54/5224b2446c871b5251945f990bdce507 [00:46:36 CEST] <kode54> that introduces two changes [00:46:45 CEST] <kode54> 1) read id3v2 "COMM" frames [00:47:05 CEST] <kode54> 2) parse "iTunSMPB" comment frames for file info [00:48:17 CEST] <atomnuker> mp3_parse_itunes_tag() doesn't look good at all [00:49:58 CEST] <atomnuker> do if (!s->metadata || !(de = av_dict_get(s->metadata, "iTunSMPB", NULL, 0))) return [00:50:13 CEST] <atomnuker> that'll get the indentation level down by 1 [00:51:20 CEST] <atomnuker> do the same for the sscanf [00:51:39 CEST] <kode54> okay, then [00:51:46 CEST] <pomaranc> does anyone know if someone is working on a mp4 dash muxer? [00:51:50 CEST] <atomnuker> sanitize the hell out of the string before sscanf too [00:55:57 CEST] <kode54> and how do I do that? [00:56:13 CEST] <kode54> this is supposed to be a machine generated string [00:56:42 CEST] <kode54> for loop and make sure it contains nothing but hex and whitespace? [00:57:46 CEST] <atomnuker> just do a basic sanity check on the length [01:05:55 CEST] <kode54> done [01:06:18 CEST] <kode54> it now checks if length is too short for the minimum size of the known number of fields, and too long for the same [01:06:26 CEST] <kode54> and the indenting should be significantly better [01:13:50 CEST] <atomnuker> make the i in for (i = 0; i < 8; ++i) { go i++ instead of ++i [01:32:46 CEST] <kode54> is it pleasing now? I also forgot something with calculating the duration [01:33:09 CEST] <kode54> aww [01:33:11 CEST] <kode54> I botched it [01:35:08 CEST] <kode54> there we go [01:37:59 CEST] <atomnuker> yeah, test it and post it on the ML [01:46:39 CEST] <kode54> verified working with this iTunes rip I originally designed it for, based on a published topic on the field format [01:50:10 CEST] <kode54> do I need to filter my email addresses when formatting the patches? [01:52:51 CEST] <atomnuker> filter? [01:56:47 CEST] <Chloe> kode54: use git format-patch -1 to generate the patch (note, dont use git diff, or diff), and then send it as an attachment or use git send-email to send it [01:57:03 CEST] <kode54> k [02:00:35 CEST] <kode54> ffmpeg-devel(a)ffmpeg.org ? [02:01:42 CEST] <Chloe> yes [02:04:26 CEST] <Chloe> kode54: I did `git config sendemail.to ffmpeg-devel(a)ffmpeg.org` to setup the repository's email so that I can just use `git send-email 0001-my-patch.patch` (after setting up the global smtp settings as well) [02:04:47 CEST] <kode54> k [02:05:50 CEST] <Chloe> kode54: this may also have more information: https://ffmpeg.org/developer.html#Submitting-patches-1 [02:06:26 CEST] <Chloe> oh yeah, remember to `make fate` [02:07:15 CEST] <JEEB> pomaranc: you probably missed libavformat/dashenc.c? [02:13:06 CEST] <pomaranc> JEEB: wow, I did [02:13:21 CEST] <pomaranc> JEEB: the documentations says there is only webm_dash_manifest [02:13:43 CEST] <pomaranc> I will try it out, thanks [02:14:23 CEST] <JEEB> yeah, many muxers don't have documentation :P you just see them on the -formats listing [02:14:35 CEST] <JEEB> and if you just specify output file name as .mpd [02:48:05 CEST] <Chloe> kode54: there seem to be a few whitespace issues with your first patch: http://sprunge.us/MZHQ your second patch applies cleanly though [02:48:14 CEST] <kode54> oh crap [02:48:26 CEST] <kode54> did I use tabs or something? [02:48:30 CEST] <kode54> dammit [02:50:17 CEST] <kode54> weird with the first one [02:50:30 CEST] <kode54> I even remade the patch [02:51:03 CEST] <Chloe> don't worry about it, just send a v2 with a comment 'fixed whitespace' or equivalent [02:59:30 CEST] <kode54> I hope that does it [02:59:54 CEST] <kode54> dammit, I set --subject [03:00:57 CEST] <kode54> clearly, git send-email is not as useful as I'd like it to be [03:04:56 CEST] <Chloe> you can use --annotate [03:05:03 CEST] <kode54> ah, oops [03:05:11 CEST] <Chloe> And then you can edit the subject directly [03:39:50 CEST] <Chloe> kode54: could a copy of the sample you're using to test? [03:40:33 CEST] <kode54> it's a full album ripped using iTunes [03:40:43 CEST] <kode54> should I package it up somewhere? [03:42:48 CEST] <Chloe> kode54: see PM. [03:53:33 CEST] <kode54> weird, the delay/padding isn't being subtracted as far as ffprobe is concerned [03:54:27 CEST] <Chloe> kode54: I haven't looked at any of the IDv2 stuff before today, but I'd say that putting it in mp3dec is probably the wrong approach [03:55:07 CEST] <kode54> is there a better place to put it so it's picked up by the stream data? [03:55:44 CEST] <cone-941> ffmpeg 03Mark Reid 07master:0a088dea3d33: ffmpeg.c: copy video profile when using stream_copy [03:57:57 CEST] <Chloe> Well the reason I think that is because iTunes' AAC also supports gapless (not just mp3) [04:26:06 CEST] <Chloe> kode54: I'm not entirely sure where else it could go, I was just wondering how it could be added to AAC as well with minimal code duplication [04:36:47 CEST] <kode54> Chloe, AAC/MP4 has a different iTunsMPB field format, though [04:36:51 CEST] <kode54> slightly different, at least [04:40:41 CEST] <Chloe> kode54: Oh, I didn't realise this. It should be fine to have an mp3 version then, the only issue is that there's no output in ffprobe I guess, and I think ffmpeg may already have aac gapless support. (ffprobe shows something with my sample). ugh, it's 3:40am, too late to be doing this. bye [04:41:21 CEST] <kode54> see you tomorrow, or something :] [04:41:34 CEST] <kode54> I need to fix this ID3v2 code too [04:41:44 CEST] <Chloe> good luck :) [04:53:10 CEST] <kode54> I compiled my ID3v2 patch and it definitely came out in the ffprobe, as far as the tags being displayed [04:53:18 CEST] <kode54> now to test the gapless thing [04:54:01 CEST] <kode54> comes out correct length [04:54:37 CEST] <kode54> are you sure you used the copy you freshly built (./ffprobe) to test the patch? [04:54:42 CEST] <kode54> you can recheck in the morning [09:24:34 CEST] <rcombs> holy fuck DTS CD is the most idiotic thing I've ever seen [09:24:51 CEST] <rcombs> the format itself, and the ridiculous hacks people have come up with to work with it [09:25:05 CEST] <rcombs> "let's pretend DTS is PCM and put it on CDDAs" [09:25:17 CEST] <rcombs> "wait won't that blow out people's speakers if they play it on CDDA players" [09:25:35 CEST] <rcombs> "hmm good point, let's add 2 padding bits for every 14 bits of audio data" [09:25:48 CEST] <rcombs> <later&> [09:26:03 CEST] <rcombs> "let's rip DTS CDs as WAVs as if they were regular PCM CDDAs" [09:26:36 CEST] <rcombs> "let's make players support reading that" [09:26:39 CEST] <rcombs> <later&> [09:27:30 CEST] <rcombs> "hmm I know how to add metadata tags to FLACs but not to WAVs& guess I'll pretend this WAV really does contain PCM and compress it with FLAC, then tag that" [12:16:21 CEST] <Chloe> kode54: ok, yes it does work. I'm not sure what happened. I think I was just missing a make clean [12:19:40 CEST] <cone-036> ffmpeg 03Chris Cunningham 07master:d59820f6fec3: libavformat/matroskadec: fix unsigned overflow to improve seeking [12:19:40 CEST] <cone-036> ffmpeg 03Nikolay Aleksandrov 07master:8ec599caa4b8: MAINTAINERS: add myself to the project server [14:06:38 CEST] <cone-036> ffmpeg 03Mark Reid 07master:e47981dab7fb: libavformat/movenc: add dnxhr compatibility for apple players [14:06:39 CEST] <cone-036> ffmpeg 03Mark Reid 07master:c2cd1a5ea7a2: tests/fate/dnxhd: add dnxhr prefix tests [15:30:55 CEST] <kasper93> Hi [15:31:20 CEST] <kasper93> Any reason that ffio_init_context doesn't initialize protocol_whitelist field? [15:31:35 CEST] <kasper93> (and protocol_blacklist for that matter) [15:32:57 CEST] <cone-036> ffmpeg 03Mark Reid 07master:cdca1902ab35: libavcodec/dnxhdenc: add support for variable mircoblock counts [15:32:58 CEST] <cone-036> ffmpeg 03Mark Reid 07master:ece12db4f1a5: libavcodec/dnxhdenc: add support for dnxhr encoding [15:33:23 CEST] <kasper93> Right now rtp/asf crashes because of that. [15:34:07 CEST] <JEEB> is that because of uninit data access or because nothing is whitelisted? [15:34:24 CEST] <JEEB> although I guess crash means the previous one [15:34:34 CEST] <JEEB> s/previous/former/ [15:34:44 CEST] <kasper93> uninit data access. [15:35:09 CEST] <kasper93> in utils.c:501 "if (!s->protocol_whitelist && s->pb && s->pb->protocol_whitelist) {" [15:35:26 CEST] <kasper93> s->pb->protocol_whitelist is uninit. [15:35:41 CEST] <kasper93> av_strdup crashes. [15:58:32 CEST] <michaelni> kasper93, is there a ticket or testcase about this ? [16:04:44 CEST] <kasper93> michaelni: I have regional locked stream that crashes. But looking at the code it will be reproducible with every asf stream. Take a look at rtpdec_asf.c:118. init_packetizer inits AVIOContext but the protocol_whitelist is not initialized. [16:05:18 CEST] <kasper93> I'm just not sure if it should be init in the rtpdec_asf.c or actually in ffio_init_context for everyone. [16:37:43 CEST] <michaelni> kasper93, is whitelist ever initialized before ffio_init_context() so that it would wipe it out ? [17:04:44 CEST] <michaelni> kasper93, you could post a patch that sets the lists in rtpdec_asf.c and point in the commit message out that you are unusure if a more generic solution would make sense [17:05:16 CEST] <michaelni> aka implement the simpler and see if everyone is happy with that [20:09:42 CEST] <cone-303> ffmpeg 03Jan Sebechlebsky 07master:75bd5d3e2dde: avformat/tee: Rescale ts using av_packet_rescale_ts [22:54:58 CEST] <Compn> ubitux : what kind of rtmpdump abuse mails did you geT? [22:55:28 CEST] <Compn> or if you can, forward them to me compn(a)ffmpeg.org [22:55:37 CEST] <Compn> if thats easy for you, otherwise no problem [00:00:00 CEST] --- Sun Jul 24 2016

1 0

ffmpeg.log.20160723
by burek 24 Jul '16

24 Jul '16

[00:11:48 CEST] <nichego> are there any multithreaded deinterlacers available with ffmpeg that match the quality of traditional yadif,mcdeint pair? [00:16:25 CEST] <furq> not without using avisynth/vapoursynth [00:19:45 CEST] <nichego> furq: can you point me to reading on how to use a vapoursynth filter and ffmpeg in the same pipeline? [00:23:10 CEST] <furq> http://vpaste.net/D9xj7 [00:23:14 CEST] <furq> that's what i use for deinterlacing dvds [00:23:57 CEST] <nichego> thank you. that's great hands-on help [00:24:33 CEST] <furq> you'll need a million different libs for qtgmc, but you can use other deinterlacers multithreaded with much less effort [00:24:56 CEST] <furq> i'm not sure if qtgmc works at all on *nix, i had to use a couple of precompiled binaries for windows [00:25:51 CEST] <furq> i should really have made a list when i was wading through forum posts piecing it all together [00:31:14 CEST] <nichego> it's explained here [00:31:16 CEST] <nichego> https://bbs.archlinux.org/viewtopic.php?id=190192 [00:31:29 CEST] <nichego> i just needed some pointers for keywords to search with [00:31:34 CEST] <nichego> your script did the job :) [00:32:24 CEST] <furq> it's a shame that vapoursynth seems to be a bit slower than avisynth [00:32:36 CEST] <furq> but it does have the advantage of not constantly segfaulting eight hours into an encode [00:33:03 CEST] <nichego> anything faster than 0.7 fps on an i7 is welcome [00:33:15 CEST] <nichego> that's the kind of thing i'm getting out of yadif,mcdeint [00:33:44 CEST] <furq> that is impressively slow [00:34:05 CEST] <furq> you can use nnedi in ffmpeg but it's singlethreaded so it's pretty much useless [00:35:13 CEST] <nichego> the slowness gives me time to figure how to use native qtgmc with vapoursynth, for future encodes [00:37:07 CEST] <furq> that looks much simpler than it does on windows [00:37:27 CEST] <furq> i ended up having to download a bunch of dlls from dropbox, which is a very secure thing to do [00:37:42 CEST] <furq> although on balance probably not that much worse than using the AUR [00:37:43 CEST] <nichego> yeah, vapoursynth really is a blessing in those regards [00:37:55 CEST] <nichego> at least the code can be reviewed [00:38:02 CEST] <nichego> rather than downloading random binaries [00:38:15 CEST] <furq> this was for vapoursynth [00:38:27 CEST] <furq> although it was only a couple of things that weren't on github, rather than literally everything like it was with avisynth [00:40:21 CEST] <nichego> it will get better over time [00:40:24 CEST] <nichego> with aur, too [00:44:06 CEST] <nichego> ls [00:44:13 CEST] <nichego> wrong window [00:48:34 CEST] <nichego> it is rather surprising that f3kdb needs python-pygments. why is that even relevant [00:50:22 CEST] <nichego> thanks again, furq [02:06:26 CEST] <mrelcee> i actually had a coredump today for the first time ever from ffmpeg. 99m into an encode.. went to encode it again and it encoded fine. i blame cosmic rays [02:11:52 CEST] <DHE> needs ECC memory. at least that's what I hear people way [02:12:07 CEST] <DHE> *say [02:13:52 CEST] <mrelcee> that, i do not have [02:15:20 CEST] <mrelcee> i am hoping it's not a 3.1.1 bug.. i got up the gumption to compile my own on freebsd with a little help from here to use gmake instead of make.. freebsd port is stuck at 2.8.7 [02:17:29 CEST] <DHE> if it is a software bug, would be a race condition or uninitialized memory usage most likely. both are a pain to identify [02:17:35 CEST] <DHE> if it is cosmic rays, well, all bets are off [02:27:17 CEST] <mrelcee> well i was. surprised when the file encoded fine after the error.. [04:48:24 CEST] <mrelcee> anudder core dump... i know the system is rock solid stable, cosmic rays seem lless likely now.. [09:11:27 CEST] <NoobGuy1> Hello. I currently use nginx to encode a stream from down to a 3.5 Mb stream for twitch. I have a couple spare identical servers, I was wondering 1, is there any sort of way to "farm" it out, or 2, would there be any advantage to doing the transcoding with two server in series, one right after another? [09:11:54 CEST] <NoobGuy1> Maybe the first server takes it down to 5 Mb, then the next to 3.5? Would there be any possible gain in quality? Thank you. [09:13:34 CEST] <NoobGuy1> I have a 25 mb stream (from NVENC H.264) sent to the server, that transcodes it to 3.5mb. Then sends it to the streaming provider. Trying to add some specifics. Thanks. [09:21:47 CEST] <Mavrik> NoobGuy1, video is stored in a lossy format. Every time you transcode it you permanently lose quality no matter the settings. [09:22:07 CEST] <Mavrik> Meaning - repeatedly transcoding the same video instead of only once is going to hurt you way more. [09:22:47 CEST] <Mavrik> "Farming" out for such videos makes little sense as well - a good Xeon will transcode that before the network transfer to another machine is properly complete. [09:49:12 CEST] <drastik> ffmpeg -i https://streamlink/playlist.m3u8 -c copy -bsf:a aac_adtstoasc -f mp4 filename.mp4 [09:49:19 CEST] <drastik> i am using this to record a live stream [09:49:29 CEST] <drastik> the command partially works, but [09:49:55 CEST] <drastik> http://pastebin.com/SKPDvQqC [09:49:55 CEST] <drastik> i get this in the console at some poin [09:49:56 CEST] <drastik> and ffmpeg never exits [09:50:24 CEST] <drastik> it still at this point [09:50:42 CEST] <drastik> noone streams now, so i know its not recording anything [09:55:19 CEST] <drastik> btw if i execute this on a small stream, i will work fine (it will exit and wont output anything) [10:02:59 CEST] <drastik> any help is greatly appriciated [13:00:02 CEST] <Vollstrecker> Hi folks, my mythtv is segfaulting on some mkv, and on some not. Using ffprobe the only difference I found (beside resolution etc.) is: yuv420p(tv, smpte170m/bt470bg/bt709) on the not working file and yuv420p(tv, bt709) on the working. I just can't find the right way to change that. Any hints? [13:02:47 CEST] <Vollstrecker> I tried with colorspace=bt709 and colormatrix=bt709:bt709 and colormatrix=bt709 but always get Unable to find a suitable output format and invalid argument [14:44:08 CEST] <Kadigan> Hello. I was wondering if someone could point me in the direction of documents that properly explain the -bufsize option? I don't understand what is meant by "enforce your requested 'average' across each Xk worth of video" [14:45:18 CEST] <JEEB> you have maxrate and bufsize, and those should always be used together [14:45:25 CEST] <Kadigan> I mean, the way I understand it, the "[n]k worth of video" is a value that says "this part of video should be encoded at [n]k to be visually identical to the original", and this confuses me [14:45:37 CEST] <JEEB> no [14:45:38 CEST] <Kadigan> JEEB: https://trac.ffmpeg.org/wiki/Encode/H.264 this page claims otherwise [14:45:56 CEST] <Kadigan> it proposes maxrate of 4000k and bufsize of 1835k [14:46:04 CEST] <Kadigan> and I do not understand how that value was arrived at. [14:46:16 CEST] <JEEB> that depends completely on your usage scenario [14:46:25 CEST] <JEEB> I mean, it's bandwidth limiting [14:46:35 CEST] <Kadigan> Okay. Better question: how should I arrive at -bufsize? [14:47:00 CEST] <JEEB> it means that you are telling the encoder that over bufsize amount of bits, never go over maxrate bits per second [14:47:18 CEST] <Kadigan> What is "bufsize amount of bits"? [14:47:39 CEST] <Kadigan> Why do I need a second limiter, when my bitrate is already limited by -maxrate? [14:47:57 CEST] <Kadigan> Is my top rate limited by -maxrate? [14:47:57 CEST] <JEEB> because maxrate by itself does jack shit as the algorithm requires two variables [14:48:17 CEST] <Kadigan> JEEB: okay, then I am obviously confused as to how the algorithm works. [14:48:20 CEST] <JEEB> if you just want to have an output of average bit rate over the whole video of X, then you don't touch maxrate|bufsize [14:48:42 CEST] <JEEB> if you want to bandwidth limit for f.ex. DVD|BD specification or for over-the-network stuff [14:48:46 CEST] <JEEB> then you need maxrate|bufsize [14:49:03 CEST] <JEEB> because you have your main rate control method, and then you are adding *additional* limitations on top of that [14:50:02 CEST] <Kadigan> Okay. Let me explain my thought process: I have a video encoded at a specific rate. My understanding is that compression is always lossy, but I don't have the patience to compare two videos frame-by-frame to determine the right settings. As such, I assumed that if I set the compressor to minrate (original of video) and then put the ceiling a little bit higher, that would account for situations where it would want to put more data into a frame than [14:50:02 CEST] <Kadigan> there was due to compression-specific stuff. [14:50:22 CEST] <JEEB> no, never use minrate [14:50:46 CEST] <JEEB> it most often doesn't do what you think it does and is not a native thing in many encoders so its implementation is crapola [14:50:57 CEST] <Kadigan> I understand that it's wasteful, but I've yet to sit through* a visual comparison that claims it's identical in terms of quality. [14:51:00 CEST] <JEEB> also your original bit rate doesn't matter jack shit when re-compressing anything [14:51:34 CEST] <JEEB> you have your input and then depending on your compression settings, format etc the requirement for some sort of visual output differs [14:51:55 CEST] <Kadigan> JEEB: okay. So let's throw everything I know about bitrates out the window. How should I approach setting my compression settings? [14:51:58 CEST] <JEEB> for libx264 the simplest way is to use CRF, which is the closest we have to "constant quality" [14:52:25 CEST] <JEEB> you start with -crf and encode like 1500 frames of stuff (cut a few parts of the content you're going to be recompressing) [14:52:29 CEST] <JEEB> *-crf 23 [14:52:33 CEST] <JEEB> (which is the default) [14:52:40 CEST] <Kadigan> Here's another question: I've noticed that ffmpeg got 'h264' and 'libx264'. Which one should I use for compatibility with devices like TVs? [14:52:54 CEST] <Kadigan> (possibly BR players that can play files) [14:52:56 CEST] <JEEB> h264 is just a decoder [14:53:15 CEST] <JEEB> you can I think use the name of the format for encoding as well, but it will just map to a random external encoder library, in most cases libx264 [14:53:19 CEST] <Kadigan> So if I specify -c:v h264, it assumes -c:v libx264? [14:53:30 CEST] <JEEB> most probably the auto-mapping will go like that [14:53:38 CEST] <JEEB> because there is no internal AVC encoder [14:53:50 CEST] <Kadigan> I see. So yes, I expect specifying libx264 is better than hoping "it will all go well". [14:53:57 CEST] <JEEB> yes [14:53:59 CEST] <Kadigan> Okay. Changed. [14:54:02 CEST] <Kadigan> Now, the bitrate? [14:54:10 CEST] <JEEB> scroll a few lines up [14:54:27 CEST] <JEEB> where I started writing "for libx264 the simplest way is to use..." [14:54:33 CEST] <Kadigan> The original video is .rmvb, and it's crap. I was also trying to go for smallest size of file, while not going beyond "barely watchable". [14:54:43 CEST] <Kadigan> Yeah, I saw the stuff about crf. [14:55:00 CEST] <JEEB> ok, so if the value looks crap then you go down and if it looks good you go up [14:55:12 CEST] <JEEB> then you will find the CRF value for yourself for that sort of content [14:55:27 CEST] <Kadigan> I usually encode either low-quality previews (where I use bitrate, maxrate minrate bufsize) for streaming on 2M lines, or for lossless storage/youtube. I don't really have a feeling for middle-ground encoding. [14:55:28 CEST] <JEEB> (you can roughly utilize the same or similar value with very similar input) [14:56:10 CEST] <Kadigan> Half the time I just -c:v prores_ks -profile:v yuv422p10le ^^ [14:56:12 CEST] <JEEB> so what I mean is that you cut/limit the encode to a few scenes of ~500 frames [14:56:27 CEST] <JEEB> and from those you can then quickly start iterating on the CRF value [14:56:31 CEST] <Kadigan> I see. So basically what I dread - video comparison. [14:56:57 CEST] <JEEB> there's no other way really :P and as I said, you can generally then use similar values for similar resolution and content [14:57:05 CEST] <Kadigan> Okay, so about CRF: if a frame needs less than what CRF would specify, will the encoder be wasteful? [14:57:19 CEST] <Kadigan> I assume not, but there may be details of the implementation I am not aware of. [14:57:22 CEST] <JEEB> no, because you picked the highest CRF value that still looks good enough for you [14:58:07 CEST] <Kadigan> The situation I mean is more like: "this scene could easily be visually-identically encoded at -crf 50, but user specified -crf 23" [14:58:19 CEST] <Kadigan> how does that work out?\ [14:58:38 CEST] <JEEB> if they are simple scenes they are easily coded with a lower quantizer as well [14:58:52 CEST] <Kadigan> I mean - will ffmpeg make that choice for me? [14:58:56 CEST] <JEEB> libx264 will [14:58:59 CEST] <JEEB> with crf [14:59:04 CEST] <Kadigan> ie. will I get the same output file size for crf 23 and crf 50 [14:59:09 CEST] <Kadigan> if the video fits perfectly into crf 50. [14:59:32 CEST] <Kadigan> (I'm generalizing here, please assume I would expect small variations) [14:59:37 CEST] <JEEB> if it's easily coded it won't be the same but the difference shouldn't be too big [15:00:08 CEST] <Kadigan> I see. So basically using -crf 23 is the way to go unless you need specific requirements met, or the video is of higher quality. [15:00:58 CEST] <JEEB> my recommendation is still to at least try it once and if it's "good enough", sure :P but generally you want to go through a few iterations of ~1500 frames or so (three 500 frame scenes or whatever) [15:01:03 CEST] <Kadigan> I'm wondering about all this because the target device is a Sony TV, and I have absolutely no idea how it'll react. I'm already limiting to Main/3.1, which I assume will be enough. [15:01:06 CEST] <JEEB> such test encodes should be quick enough [15:01:20 CEST] <Kadigan> Yeah, but I don't have the device. ^^ [15:01:21 CEST] <JEEB> main/3.1 seems like OK limits for a lot of hardware [15:01:45 CEST] <JEEB> Kadigan: but you can still check the output on your shit :P unless you have no screen whatsoever [15:02:04 CEST] <JEEB> and those test encodes should be rather quick to create since as I said you usually use 1500-2000 frames for them [15:02:04 CEST] <Kadigan> Yeah, but I'm running CCCP+SVP+madVR+MPC-HC, [15:02:11 CEST] <Kadigan> so I assume that it will play anything I throw at it :D [15:02:49 CEST] <Kadigan> (I've confirmed it will also play 4K and 8K video, which the TV in question would most certainly crap out at) [15:02:52 CEST] <JEEB> SVP will be your main headache [15:03:01 CEST] <JEEB> but that's off-topic on this channel [15:03:13 CEST] <Kadigan> No, I have a separate setup that bypasses SVP for this very reason. [15:03:33 CEST] <Kadigan> I assume that as long as I don't hit the HW bitrate limitations on the TV [15:03:43 CEST] <Kadigan> (I've found that a good ceiling value is 20M/s) [15:03:45 CEST] <Kadigan> I should be good. [15:04:22 CEST] <JEEB> blu-ray official limits are 40M maxrate and 30M bufsize, but even those players generally tend to be able to take in more [15:05:03 CEST] <Kadigan> I've noticed that some players will not play unless the file is strictly within spec, or within their own exotic spec... But maybe I'm just biased. [15:05:26 CEST] <JEEB> most modern stuff supports up until high/4.1 (profile/level) [15:05:30 CEST] <Kadigan> (as a rule of thumb, cheaper players will accept and play a wider variety of stuff, too...) [15:05:44 CEST] <JEEB> some sony stuff supports only main/4.1 [15:05:56 CEST] <JEEB> and then the PSP f.ex. only supported main/3.0 [15:05:57 CEST] <Kadigan> +below, or STRICTLY Main/4.1? [15:06:20 CEST] <JEEB> Kadigan: (constrained) baseline is a subset of main, main is a subset of high [15:06:34 CEST] <Kadigan> Ah. [15:06:39 CEST] <Kadigan> That makes sense, actually. [15:06:46 CEST] <JEEB> profiles = features [15:06:48 CEST] <Kadigan> So always backwards-compatible due to inclusion. [15:06:55 CEST] <JEEB> levels = amount of memory required by playback device [15:07:08 CEST] <Kadigan> Now that makes sense. Didn't know that. [15:07:56 CEST] <JEEB> so basically levels define stuff like resolution, maximum available pictures in reference frame buffer etc [15:08:05 CEST] <JEEB> and bit rate limits [15:08:46 CEST] <JEEB> so in theory if you want to be *really* pedantic you need to always stick to some sort of maxrate/bufsize but so much hardware is capable of over those limits that we generally tend to ignore them ;) [15:08:50 CEST] <Kadigan> Wow... So let me understand this: the Profile and Level now dictate the minimum set of "decoder features" and "framebuffer memory", stuff that I recall from the good old days of S3 Virge. [15:09:26 CEST] <Kadigan> As in, a handy way to describe hardware capabilities without using GL Extensions and chip inspection. [15:09:31 CEST] <JEEB> yup, although not limited to framebuffer memory [15:10:16 CEST] <JEEB> but for the decoder chip the required amount of memory and the required amount of memory bandwidth is what's being defined [15:10:44 CEST] <JEEB> I think they define the buffer sizes in macroblocks so you can then indirectly calculate the amount of reference frames based on your resolution padded to 16x16 blocks [15:10:47 CEST] <Kadigan> So a bit better metric, because it combines more information. [15:11:18 CEST] <Kadigan> Damn, video encoding (while still something to figure out) has gotten a lot easier over the years, hasn't it. [15:11:29 CEST] <JEEB> (this is what is done when you set -level 41 f.ex on the libx264 wrapper - it will limit the amount of used reference frames accordingly) [15:11:30 CEST] <Kadigan> (and I assume ffmpeg is in big part to thank for that) [15:12:02 CEST] <Kadigan> I see. [15:12:14 CEST] <JEEB> yeah, AVC was a major step in making the video decoding process bit-exact (yes, MPEG-1/2 Video and MPEG-4 Part 2 weren't bit-exact), and they had clear "tiers" for things like that [15:12:33 CEST] <JEEB> of course crappy vendors have created crappy hardware during the years, but mostly we are in a very happy land compared to before [15:13:16 CEST] <Kadigan> So, ffmpeg just finished encoding the file. I used aac at 128k and crf 23. The file is quite bigger than the original .rmvb (400 vs 550MB), but I assume it has to do with my audio settings (the original was MP3 at 86kbps) and file/encoding structure. [15:13:54 CEST] <Kadigan> It's not an issue - just something I'm musing on. [15:14:26 CEST] <Kadigan> Here's a good question: is there a video player for Windows that You know of that can play two video files side-by-side in sync? [15:14:44 CEST] <Kadigan> (speficically for the purposes of comparing two encodes of the same source file) [15:15:48 CEST] <JEEB> not that I know. I usually just open two mpvs and/or vapoursynth editor with two things interleaved [15:16:03 CEST] <JEEB> also as I noted, don't just encode the whole thing unless it's really short [15:16:37 CEST] <JEEB> you can -ss and -t (seek and time to encode) in ffmpeg cli if you want to use that for cutting a part out [15:17:35 CEST] <Kadigan> Yeah, I typically use -ss and -vframes [15:17:47 CEST] <Kadigan> And yes, it took me a while to figure out where to put -ss [15:18:47 CEST] <Kadigan> (I eventually got used to source-seeking for minutes and then output-seeking to frames) [15:20:08 CEST] <Kadigan> Also, another good use-case for ffmpeg: I quite often use it in conjuction with an XML parser to cut up output from an NLE - say, I have 20 videos which I did in a row (30-90s each) using the same effects/tools, and I don't want to export each and every one separately, because PITA. [15:21:14 CEST] <Kadigan> Anyway, thanks for the assist JEEB. [15:21:16 CEST] <JEEB> for stuff like editing I recommend taking a look at vapoursynth [15:21:27 CEST] <JEEB> and possibly vapoursynth editor for preview purposes [15:21:35 CEST] <JEEB> you write a script and that outputs you raw YCbCr or RGB [15:21:41 CEST] <Kadigan> I assume You mean - instead of using an NLE. [15:21:47 CEST] <JEEB> not necessarily [15:21:56 CEST] <Kadigan> Does it run on OS X? [15:21:58 CEST] <JEEB> yes [15:22:55 CEST] <Kadigan> Ah, so it's a cross between avisynth (which I remember being a plugin) and a script interpreter. [15:23:08 CEST] <Kadigan> (I'm probably very wrong there :D) [15:23:24 CEST] <Kadigan> Interesting. [15:23:51 CEST] <JEEB> well avisynth is "the original" OSS frameserver, and it just had its own scripting language [15:24:15 CEST] <JEEB> which was very crappy and people just got used to it (you f.ex. didn't have ifs etc but you did have ternary operators) [15:24:34 CEST] <Kadigan> I remember there was a good reason why I didn't learn avisynth. [15:24:51 CEST] <JEEB> while vapoursynth just takes python and utilizes it for the scripting part, although in theory you can make your own mappings against other languages [15:25:18 CEST] <Kadigan> Maybe because people usually recommended its use with VirtualDub, which was in itself mighty unstable... [15:25:38 CEST] <Kadigan> Hm, that's an interesting proposition actually. [15:25:55 CEST] <JEEB> virtualdub was one of the first things you could preview it with nicely [15:26:06 CEST] <JEEB> although later you had AvsP(mod) [15:26:09 CEST] <Kadigan> I do PHP work mostly, with the occasional dip into C/C++/C#. Maybe there's a mapping set for PHP out there. [15:26:41 CEST] <JEEB> nah, the only ones that were worked upon were lua and JS and neither was finished, IIRC [15:27:02 CEST] <JEEB> although I'm not sure since I remember some people using !python with vapoursynth in mpv [15:27:17 CEST] <furq> LuaSynth is a LuaJIT 2.0 binding for VapourSynth. It is currently a hacked together mess which was mostly created in a single sitting with very little thought, [15:27:20 CEST] <furq> well at least he's honest [15:27:36 CEST] <JEEB> right, that one [15:34:41 CEST] <furq> that seemed interesting for a minute there until i realised i'd also have to port >10kloc of vs scripts for it to be of any use [15:35:18 CEST] <furq> which will probably take more time than just dealing with vs being a bit slow [15:36:12 CEST] <JEEB> well I'm not sure if much of the slowness is due to cpython compared to anything that is bound straight into native things [15:36:50 CEST] <Venti^_> Kadigan: I think any HD tv should be able to handle high profile level 4 [15:43:40 CEST] <Venti^_> well, DVB televisions at least, americans might do things different [16:07:01 CEST] <Kadigan> I think so too, but I've seen stuff fall flat on its face if the USB drive didn't have a /DCIM/SOMESHIT/ folder structure. [16:07:21 CEST] <Kadigan> (hence I'm cautious about making assumptions of it "just working") [16:16:28 CEST] <Kadigan> Anyway, thanks again! [16:38:46 CEST] <Kadigan> I'm back. I have another issue. I have a video file that's supposed to be 4:3, but it's behaving as if it's something more like 5:3. I tried scaling it, but it produced output that's similar to the original. MPC-HC tells me the AR of the video is 1.772, and I need it to be 1.3[3]. How do I tell ffmpeg to change AR to that specific value as well? [16:39:49 CEST] <JEEB> post output of that file on a pastebin from ffprobe filename or ffmpeg -i filename [16:40:04 CEST] <JEEB> and link here [16:40:29 CEST] <Kadigan> http://pastie.org/private/qhpq8uay1pbnihamgvdidw [16:41:03 CEST] <Kadigan> Ah, that got the output. Wait. [16:41:29 CEST] <Kadigan> http://pastie.org/private/tq63edp6ereowjhc8aslw here is the source file [16:42:09 CEST] <Kadigan> Can I simply setdar=4/3 ? [16:42:37 CEST] <JEEB> if that thing is just wrongly flagged as 1:1, yes [16:42:39 CEST] <Protected> Hey all. Is there any way to get -progress to overwrite the file instead of appending to it? [16:43:30 CEST] <Kadigan> JEEB: setdar=4/3 worked. Yes, it was encoded from an old reel to PC by someone who didn't know what they were doing and was content with the result. [16:43:47 CEST] <Kadigan> It's a TV news broadcast from 1986 in Poland, so it couldn't have been 16:9 ;) [16:44:14 CEST] <JEEB> now just hope you know exactly how much was cropped out as well [16:44:26 CEST] <Kadigan> I believe it was simply squished. [16:44:38 CEST] <Kadigan> (ie. when I recode to 4:3, it looks okay and seems to have the entire picture) [16:44:52 CEST] <Kadigan> Yup, it looks good now. [16:44:55 CEST] <JEEB> well, just switching the aspect ratio doesn't change the content [16:44:58 CEST] <JEEB> btw, if you don't want to re-encode you can just use L-SMASH's tools (muxer/remuxer) to just modify the aspect ratio [16:45:15 CEST] <JEEB> since you've got an ISOBMFF (colloquially "mp4") file there [16:45:25 CEST] <Kadigan> I reencoded already, it's 13min long and SD. ;) [16:45:42 CEST] <JEEB> well, if you ever want to fix it without doing another lossy set of compression [16:46:10 CEST] <Kadigan> Yeah, I'll keep that in mind. Seems that most modern containers are very versatile. [16:46:52 CEST] <JEEB> for AVC you can have the SAR in the bit stream, but then if you have info on the container level that generally is specified to override that [16:46:52 CEST] <Kadigan> (I am aware, to an extent, of the difference between "encoding" (format?) and "container") [16:47:30 CEST] <JEEB> (ISOBMFF specifically says that the info in the boxes override whatever there might be in the bit stream) [16:47:50 CEST] <Kadigan> Bit of an aside: considering multiple streams, ordered chapters and multiple extra files (like subs, fonts etc.) - and assuming I don't know the capabilities of each - which container is better: mp4 or mkv? [16:48:55 CEST] <JEEB> if you need virtual timelines mp4 does support them but they definitely got more support in players on the matroska side (the idea stems from ISOBMFF's roots in MOV) [16:49:15 CEST] <JEEB> and I must say that virtual timelines in at least matroska aren't supported at all in libavformat [16:49:44 CEST] <JEEB> (things usually either have their own demuxer or they implement a wrapper around libavformat) [16:50:05 CEST] <JEEB> then you have segment linking which is separate but was often used with virtual timelines [16:50:39 CEST] <JEEB> as a feature it was present once again in mov originally, but more seen used in certain circles of matroska users. generally same comments apply with regards to libavformat not supporting it [16:51:17 CEST] <JEEB> for subs and fonts you in theory can do it with mp4 but generally more supported in matroska even though ASS is atrocious [16:53:47 CEST] <furq> 15:44:58 ( JEEB) btw, if you don't want to re-encode you can just use L-SMASH's tools (muxer/remuxer) to just modify the aspect ratio [16:53:57 CEST] <furq> you can modify the dar with ffmpeg [16:54:03 CEST] <furq> -c copy -aspect 16:9 [16:54:55 CEST] <JEEB> oh [16:55:00 CEST] <JEEB> I just remember that not working at some point [16:55:03 CEST] <JEEB> nice if it does nowadays [16:55:15 CEST] <furq> i'm not going to pretend to know it always works, but it's always worked for me [16:55:23 CEST] <JEEB> :) [17:19:42 CEST] <Threads> anyway to have subtitles added but not forced on ? [17:19:57 CEST] <Threads> i keep getting them added and they play at the same time with the file and i dont want that [18:49:33 CEST] <oerg866> Hello. I have a video recorded at 720x576 interlaced 25fps. However the source is an old video game outputting 720x288 at 50fps. How can I put each "field" into its own frame? The deinterlace filters seem to all apply some algorithm instead of just taking raw fields. [18:51:15 CEST] <furq> oerg866: https://ffmpeg.org/ffmpeg-filters.html#separatefields [18:51:34 CEST] <oerg866> Thank you so much :3 [19:19:24 CEST] <gordonfly> anyone have experience joining two videos? [19:23:23 CEST] <__jack__> sure [19:26:51 CEST] <gordonfly> what is the recommended way to do this through ffmpeg? [19:27:05 CEST] <gordonfly> can't seem to get this script to work:https://trac.ffmpeg.org/wiki/Concatenate [19:27:55 CEST] <gordonfly> is there a media format that works best for concatenation? [19:33:06 CEST] <NoobGuy1> Mavrik, thank you. I am working with a Dual Xeon E5-2450 (2 GHz 8 core), but it starts to choke on any preset last medium when transcoding live. Just looking for other options to take advantage of the other servers I have sitting idle. Doesn't look like there is any other choice sadly. [19:58:48 CEST] <podman> Hopefully a quick question. I have some files that require larger probesize and analyzeduration values than the default. In the documentation it says: A higher value will enable detecting more accurate information, but will increase latency. Is the latency significant? If the probe finds all of the information needed before it's reaced the probesize or [19:58:48 CEST] <podman> analyzeduration limits, will it stop or will it keep going until it's reached the probesize and analyzeduration values? [20:19:44 CEST] <DHE> some people just want the probe process to be fast. obviously it will need to read and decode more data if you crank up the numbers. [22:04:57 CEST] <KDDLB> quick question, the DTS decoder supports HD-MA? [22:05:50 CEST] <KDDLB> does the & support * [22:06:28 CEST] <JEEB> if it's new enough to have had it been replaced by the dts decoder library maintainer then yes [22:06:36 CEST] <JEEB> it supports the XLL lossless extension [22:06:55 CEST] <JEEB> (basically the old decoder was scrapped and replaced with the libdts one's internals) [22:07:26 CEST] <JEEB> since that I think support for DTS Express was also added [22:58:54 CEST] <DelphiWorld> SUP [23:01:32 CEST] <__jack__> ER [23:04:34 CEST] Action: DelphiWorld here with __jack__ [23:19:29 CEST] <Soltis> Is it possible to "stabilize" color with ffmpeg if, e.g., the camera's automatic color correction went nuts during filming? [23:21:12 CEST] <DelphiWorld> i'm still having issue streaming a hwaccel encoded H.264 to rtmp [23:31:35 CEST] <Kadigan> Just for kicks, how can I use ffmpeg to encode into Animation? And by "Animation" I mean the codec 'Animation', the one that allows me to export alpha channels at "Millions of colors+". [23:32:09 CEST] <Kadigan> Apparently Google is now too bloated to answer that question w/o telling me how to encode video animation into h264. [23:32:33 CEST] <JEEB> so what exactly do you require? 8bit per sample RGB with alpha lossless? [23:33:08 CEST] <Kadigan> I honestly never did that in ffmpeg, so all I know is that I need to re-code from ProRes 4:4:4:4 to Animation so that Premiere Pro accepts it with Alpha. [23:33:19 CEST] <JEEB> whatever that "animation" means [23:33:49 CEST] <JEEB> but I do know that most editors take in Ut Video (which is lossless) through the VFW component and supports alpha [23:34:03 CEST] <JEEB> for which I made an encoder in FFmpeg in 2012 (-c:v utvideo) [23:34:59 CEST] <Kadigan> JEEB: I don't know what You mean. I don't kow this Ut Video. [23:35:11 CEST] <JEEB> same goes for whatever you mean with this "animation" thing [23:35:25 CEST] <furq> https://en.wikipedia.org/wiki/QuickTime_Animation [23:35:26 CEST] <furq> this? [23:35:38 CEST] <JEEB> oh, quicktime RLE [23:35:41 CEST] <JEEB> that I've seen [23:35:44 CEST] <JEEB> mentioned somewhere [23:35:50 CEST] <Kadigan> https://discussions.apple.com/thread/1895735?tstart=0 [23:35:52 CEST] <JEEB> and yes, that article links to the dec and encoder [23:35:58 CEST] <JEEB> in FFmpeg [23:35:59 CEST] <furq> -c:v qtrle [23:36:00 CEST] <Kadigan> For all my purposes it has always been called "Animation" [23:36:08 CEST] <Kadigan> I've never seen it called QT RLE [23:36:23 CEST] <Kadigan> How do I tell qtrle to encode with alpha? [23:36:55 CEST] <JEEB> anyways, that seems in a way similar to Ut Video going by the RLE word. although Ut Video did have prediction as well [23:37:18 CEST] <JEEB> Kadigan: it should keep that as-is unless you tell it otherwise or it doesn't support the input pixel format you're giving to it [23:37:54 CEST] <JEEB> see the output pixel format that ffmpeg's command line ouptuts :P [23:38:24 CEST] <JEEB> supported formats for QT RLE are AV_PIX_FMT_RGB24, AV_PIX_FMT_RGB555BE, AV_PIX_FMT_ARGB, AV_PIX_FMT_GRAY8, AV_PIX_FMT_NONE [23:38:53 CEST] <JEEB> you probably want the ARGB one and then you'll want to make sure the YCbCr that comes from prores is properly converted [23:40:05 CEST] <JEEB> but yeah, I recommend Ut Video nowadays for windows and OS X video editing since the original author makes plugins for all major multimedia systems [23:42:24 CEST] <Kadigan> There a straightforward way to have ffmpeg iterate through a directory and convert all specified files? [23:42:27 CEST] <Kadigan> (Windows CMD) [23:43:06 CEST] <furq> does installing msys bash count as straightforward [23:43:30 CEST] <JEEB> just learn whatever windows scripting thing you want (or even python if you want something future-proof) [23:43:39 CEST] <JEEB> even batch can do for loops over files [23:43:44 CEST] <Kadigan> Okay, fine. ;) [23:44:02 CEST] <Kadigan> Bash loops are not straightforward, but I'll deal :D I was kinda hoping for some autodetection mode,\ [23:44:07 CEST] <Kadigan> but yeah, not the first rodeo. [23:45:41 CEST] <JEEB> isn't it something like `for i in $(find . -iname '*.ext'); do ffmpeg -i "${i}" -options! "out-${i}.ext"; done`? [23:45:51 CEST] <Kadigan> it's more like [23:46:02 CEST] <JEEB> it's batch that I actually don't remember [23:46:04 CEST] <Kadigan> `for /f "tokens=*" %f in ('dir /b') DO` [23:46:16 CEST] <JEEB> right, batch [23:46:17 CEST] <furq> find . -name *.mp4 -exec ffmpeg -i "{}" {}.mkv \; [23:46:23 CEST] <furq> obviously not in batch though [23:46:31 CEST] <furq> also with more quotes [23:46:41 CEST] <JEEB> you said bash so I put out a bash example :P [23:46:45 CEST] <Kadigan> furq I'm used more to `ls | while read $f; do` [23:46:50 CEST] <Kadigan> and just dealing with . and .. ;D [23:46:59 CEST] <furq> well you should stop being used to that [23:47:24 CEST] <Kadigan> it's shorter than `find . -type f -name "*.mov"` and ignores case. [23:47:24 CEST] <JEEB> and right, find had exec too [23:48:29 CEST] <JEEB> I think `find . -type f` would then do that what you want? [23:48:37 CEST] <JEEB> if you just want to take in all files [23:48:42 CEST] <furq> you don't need -type f if you're using -name [23:48:53 CEST] <JEEB> yes, but if he wanted the equivalent of just ls [23:48:58 CEST] <JEEB> without extra filters [23:49:00 CEST] <furq> also you don't need to specialcase . and .. if you use find, which i would hope removes any verbosity benefits [23:49:08 CEST] <Kadigan> furq: I know [23:49:29 CEST] <Kadigan> JEEB: doesn't matter; I'm on Windows right now, and on OS X I just write scripts (and then write them right) once and then forget. [23:49:37 CEST] <furq> also in that case why not just use for f in *; do [23:49:52 CEST] <JEEB> Kadigan: sure [23:50:12 CEST] <JEEB> and on windows the stuff you have by default are batch and then the VBscript and JSscript things [23:50:17 CEST] <JEEB> haven't tried the latter ones at all [23:50:19 CEST] <furq> and powershell [23:50:23 CEST] <JEEB> oh, right [23:50:24 CEST] <JEEB> that one [23:50:25 CEST] <furq> not that i think much of powershell [23:50:42 CEST] <furq> i know some people who swear by it but those are mostly people who get paid to edit group policy [23:51:13 CEST] <JEEB> yeah, the MS server stuff actually seems to have great integration to powershell [23:51:20 CEST] <JEEB> also you can ls through registry in it [23:51:35 CEST] <JEEB> so for some stuff it seems awesomesauce on windows in order to not have to use the GUI stuff :D [23:52:15 CEST] <furq> only in environments where you're prevented from installing a decent language [23:52:25 CEST] <furq> but then a lot of windows sysadmins are often in that environment [23:53:13 CEST] <JEEB> well I don't think you have the integration to those enterprisey things and registry as well (registry as a pseudo file system) in a lot of scripting systems [23:53:24 CEST] <JEEB> in which case you'd have to write that stuff yourself or grab it somewhere [23:53:38 CEST] <JEEB> so in that sense if shit's already provided it just makes sense to utilize [23:54:06 CEST] <Kadigan> PowerShell is nice in at-the-very-least that it supports UTF-8 right off the bat :D [23:54:16 CEST] <furq> you can edit the registry from most scripting languages i've used [23:54:23 CEST] <furq> probably not group policy and whatnot though [23:54:33 CEST] <JEEB> that's true > registry. I think python stdlib might have it even [23:54:40 CEST] <furq> yeah i think it does [23:55:11 CEST] <JEEB> but I like this about PS at the very least [23:55:12 CEST] <JEEB> PS C:\Users\jeeb> cd hkcu:\ [23:55:12 CEST] <JEEB> PS HKCU:\> dir [23:55:19 CEST] <JEEB> <stuff> [23:59:15 CEST] <KDDLB> JEEB TIL [23:59:24 CEST] <KDDLB> I no longer use Windows on a daily basis, though [23:59:47 CEST] <JEEB> I'm a mixed user, I get the cancer from multiple places ;) [23:59:57 CEST] <KDDLB> harsh words Dx [00:00:00 CEST] --- Sun Jul 24 2016

1 0

ffmpeg-devel.log.20160722
by burek 23 Jul '16

23 Jul '16

[01:00:49 CEST] <cone-147> ffmpeg 03Michael Niedermayer 07master:697b2bbd54d0: avcodec/ffv1dec: Test extradata_size instead of extradata for better robustness [01:14:55 CEST] <cone-147> ffmpeg 03Michael Niedermayer 07master:b9d77c1dfc28: doc/protocols: Fix seperated typo [01:15:07 CEST] <ubitux> http://ubitux.fr/pub/pics/ffcurvesdbg.png do ppl usually prefer a normalized legend [0;1] or sth like this? [01:19:05 CEST] <ubitux> so vs this: http://ubitux.fr/pub/pics/ffcurvesdbg-normalized.png [01:19:13 CEST] <ubitux> also, i could change the ticks [01:19:37 CEST] <ubitux> (context: adding an option to curves filter to debug them out) [01:21:40 CEST] <ubitux> http://ubitux.fr/pub/pics/ffcurvesdbg-normalized-01tics.png i like this one ... [02:35:29 CEST] <ubitux> can't delete files in trac? i messed up a bit [03:04:32 CEST] <jamrial> ubitux: someone with enough privileges probably can [03:17:13 CEST] <cone-147> ffmpeg 03Yong Lei 07master:8dcfe43de09b: avcodec/huffman: beautify: add space between #include and filename. [06:02:02 CEST] <cone-147> ffmpeg 03James Zern 07master:2afddfe9bf72: libvpxdec,cosmetics: rename common fns vp8 -> vpx [06:02:03 CEST] <cone-147> ffmpeg 03James Zern 07master:640c0962c09c: libvpxenc,cosmetics: rename common fns vp8 -> vpx [09:51:37 CEST] <midnight> Hi.Is anyone there? [09:53:51 CEST] <midnight> I read the source of the function av_image_get_linesize in imgutils.h.Its param plane confused me . [09:55:55 CEST] <andrey_turkin> what exactly confused you? [09:56:00 CEST] <midnight> it is unused,I wonder why it added ?Thanks. [09:56:19 CEST] <andrey_turkin> it is not unused in sources I have [09:56:27 CEST] <andrey_turkin> it cannot be unused in fact [09:56:58 CEST] <andrey_turkin> because linesize depends on a plane. E.g. for yuv420p chroma planes have smaller linesizes than luma plane [09:57:54 CEST] <midnight> when i wanted to get linesize, what param should I supply? [09:58:07 CEST] <andrey_turkin> well, plane number [10:00:22 CEST] <midnight> where can I find the number corresponding to some specific images? [10:02:28 CEST] <andrey_turkin> maybe you have some misconception about what this function does. Anyway, it seems like this discussion belong at #ffmpeg channel instead [10:06:18 CEST] <midnight> In fact, when I call this function av_image_fill_arrays in imgutils.h, I found that I have no idea about param dst_linesizes[4], [10:06:56 CEST] <midnight> And I think I can get inesizes by this funciton av_image_get_linesize. [10:09:22 CEST] <midnight> But i found soon I don't know the param plane either. [10:23:09 CEST] <andrey_turkin> do you know what planes are? [10:24:51 CEST] <Betablocker> hihi [12:57:15 CEST] <jermy> Probably more a mailing list question, but we've got a client wanting us to embed 608/708 captions in quicktime files - it can be done as a separate track with ccdp packets separate from the video, so somewhat easier than the AVC SEI NAL embedding option (or similar for MPEG2) [12:58:04 CEST] <jermy> Is there any formal process for ffmpeg bounties, or more a case of finding somebody who is interested? [13:06:46 CEST] <cone-224> ffmpeg 03Michael Niedermayer 07master:e83a01def3ad: avformat/teeproto: Fix memset sizeof [13:46:36 CEST] <omerjerk> hi Michael told me to replace this code - https://github.com/omerjerk/FFmpeg/blob/float_patch/libavcodec/alsdec.c#L15… [13:46:40 CEST] <omerjerk> with av_rb32 [13:46:55 CEST] <omerjerk> I'm not getting which function to use exactly. [13:47:05 CEST] <omerjerk> it's present in several header files [13:47:55 CEST] <Chloe> jermy: probably just send it to the mailing list. [13:48:51 CEST] <michaelni> omerjerk, its a macro [13:49:14 CEST] <omerjerk> yes. but I'll need to include some header file I guess. [13:49:24 CEST] <michaelni> omerjerk, libavutil/intreadwrite.h [13:49:46 CEST] <michaelni> AV_RB32 not av_rb32 [13:49:52 CEST] <omerjerk> okay. I was confused because there's one in bytestream.h too I guess [13:49:57 CEST] <omerjerk> ohkay [14:45:57 CEST] <cone-224> ffmpeg 03Petru Rares Sincraian 07master:fa393e30614b: fate: add test for extrastereo filter [15:43:34 CEST] <tmatth> kierank: do you have a list of the samples you ran through fffuzz? [16:08:54 CEST] <kierank> tmatth: nop [16:08:59 CEST] <kierank> just random samples I made [16:10:31 CEST] <tmatth> i guess more importantly, roughly what codecs you tested? [16:11:12 CEST] <kierank> h264, mpeg2, opus, ffv1 [16:11:43 CEST] <kierank> but this was before all the stupid merges in ffmpeg [16:11:47 CEST] <kierank> and rearrangements in libav [16:37:30 CEST] <cone-224> ffmpeg 03Steven Liu 07master:6f0746783124: doc/muxers: add hls_flags option description [16:37:31 CEST] <cone-224> ffmpeg 03Steven Liu 07master:41ad6d193ade: doc/muxers: improve hlsenc description [17:03:34 CEST] <cone-224> ffmpeg 03Sasi Inguva 07master:282477bf4534: libx264: Increase x264 opts character limit to 4096 [17:04:36 CEST] <JEEB> we should really normalize for x264-params... [17:05:23 CEST] <JEEB> x264opts maps things locally while x264-params just sends k/v pairs to libx264 [18:58:44 CEST] <cone-182> ffmpeg 03Jan Sebechlebsky 07master:8c997b3d3d1b: avcodec/mpeg4_unpack_bframes_bsf: Copy packet props [19:14:15 CEST] <cone-182> ffmpeg 03Jan Sebechlebsky 07master:1c46e53dcc68: avcodec/mpeg4_unpack_bframes_bsf: Check av_packet_from_data() return value [19:23:59 CEST] <cone-182> ffmpeg 03Carl Eugen Hoyos 07master:db8e8c97311a: lavc/intrax8: Use correct printf specifier for size_t on Windows. [21:25:36 CEST] <Chloe> Who maintains files which have no explicit maintainer? [21:29:17 CEST] <iive> Chloe: michaelni [22:01:15 CEST] <jamrial_> Chloe: nobody in particular, hence no explicit maintainer, but if you find a bug report it on trac, and someone that knows the code may look at it [23:59:30 CEST] <kode54> dammit [23:59:38 CEST] <kode54> oh, I found it [23:59:51 CEST] <kode54> I have a modified version of the mp3 stream parser that handles iTunes gapless info [00:00:00 CEST] --- Sat Jul 23 2016

1 0

ffmpeg.log.20160722
by burek 23 Jul '16

23 Jul '16

[00:07:04 CEST] <DelphiWorld> still zero luck [00:07:48 CEST] <DelphiWorld> even if i do format conversion in the vaapi filtergraph it's not applied [00:19:30 CEST] <DelphiWorld> furq: finaly that cmd work with both flv & mp4 but dont want to stream to rtmp. strange [00:19:38 CEST] <DelphiWorld> ffmpeg -vaapi_device /dev/dri/renderD128 -i udp://@239.100.1.2:1234 -acodec libfdk_aac -profile:a aac_he_v2 -b:a 64k -ar 48k -pix_fmt yuv420p -vf 'format=nv12|vaapi,hwupload,scale_vaapi=640:360' -c:v h264_vaapi -vb 800k -f flv rtmp://cdn.mybox.net/live/almajd [00:24:19 CEST] <furq> can you stream anything to that rtmp server [00:24:32 CEST] <DelphiWorld> furq: let me try to stream a video with copy [00:26:20 CEST] <DelphiWorld> yes furq it's streaming [00:30:16 CEST] <furq> encode an flv with those settings and pastebin the output of ffprobe [00:30:24 CEST] <DelphiWorld> ok [00:37:53 CEST] <DelphiWorld> i even lose the settings... lot of try's [00:50:44 CEST] <DelphiWorld> furq: are we missing hwdownload after the scale filter? [00:51:21 CEST] <furq> you can try it but i got the impression that was for using a software encoder [00:51:35 CEST] <DelphiWorld> ok so no need it [00:51:39 CEST] <DelphiWorld> i want fully hardware [00:51:58 CEST] <furq> i don't have a vaapi device to test with so i'm just reading the same docs as you [00:52:09 CEST] <DelphiWorld> furq: i can provide a box [00:52:42 CEST] Action: DelphiWorld is encoding the flv [00:52:57 CEST] <DelphiWorld> i even removed the scalling... [00:53:00 CEST] <DelphiWorld> ffmpeg -vaapi_device /dev/dri/renderD128 -i udp://@239.100.1.2:1234 -acodec libfdk_aac -profile:a aac_he_v2 -b:a 64k -ar 48k -vf 'format=nv12|vaapi,hwupload' -c:v h264_vaapi -vb 800k -f flv out.flv [00:54:10 CEST] <DelphiWorld> furq: http://paste.ubuntu.com/20374150/ [00:55:02 CEST] <jkqxz> Could the default 2 B-frames be bad for streaming? The reordering might cause problems somewhere, though I don't know what. (That is, try adding "-bf 0".) [00:55:38 CEST] <DelphiWorld> jkqxz: ok, where to add this -bf 0? after or befaure the -codec:v ? [00:56:00 CEST] <jkqxz> After -c:v, next to the bitrate. [00:57:28 CEST] <DelphiWorld> jkqxz: zero effect..., still shutting down [00:57:33 CEST] <DelphiWorld> ffmpeg -vaapi_device /dev/dri/renderD128 -i udp://@239.100.1.2:1234 -acodec libfdk_aac -profile:a aac_he_v2 -b:a 64k -ar 48k -vf 'format=nv12|vaapi,hwupload' -c:v h264_vaapi -vb 800k -bf 0 -f flv rtmp://cdn.mybox.net/live/almajd03 [00:58:24 CEST] <DelphiWorld> thought it work when multicasted, but rtmp refuse it [00:58:28 CEST] <jkqxz> Does it work with a software encode (libx264)? [00:58:38 CEST] <DelphiWorld> yes it work [00:58:44 CEST] <furq> bframes should make no difference with rtmp [00:59:13 CEST] <furq> i have no idea why outputting to an flv file would work, but rtmp wouldn't [00:59:57 CEST] <DelphiWorld> gonna make a flv based on that then try to rtmp it and see [01:00:21 CEST] <furq> if that doesn't work then i'd have to suspect there's a bug somewhere [01:00:27 CEST] <furq> er [01:00:29 CEST] <furq> if that does work, rather [01:01:52 CEST] <DelphiWorld> http://paste.ubuntu.com/20374905/ [01:01:53 CEST] <DelphiWorld> dont work [01:05:13 CEST] <DelphiWorld> i used -vn, audio works [01:05:17 CEST] <DelphiWorld> but -an, video dont work [04:22:47 CEST] <GyrosGeier> hi [04:23:05 CEST] <GyrosGeier> I'm trying to convert a capture from my laptop camera to mpeg4 [04:23:24 CEST] <GyrosGeier> it is horribly slow for some reason, something like 0.1fps [04:24:21 CEST] <GyrosGeier> encoding speed is normally around 10-20fps for me [04:24:31 CEST] <GyrosGeier> with 720p content [04:24:42 CEST] <GyrosGeier> now I have 640x480 [04:24:58 CEST] <GyrosGeier> Stream #0:0 -> #0:0 (rawvideo (native) -> mpeg4 (native)) [07:47:52 CEST] <mrelcee> trying to compile 3.1.1 freebsd 10.3 anyone care to help? [07:48:01 CEST] <mrelcee> +on [08:22:46 CEST] <relaxed> mrelcee: what's the problem? [08:24:44 CEST] <mrelcee> relaxed: hi thanks.. i successfully .configure. with my options but when i type make i get a long list of "line X: Need an Operator". and a fail [08:26:55 CEST] <mrelcee> http://pastebin.com/k1V4YPzv [08:27:58 CEST] <mrelcee> i installed the latest make after the first fail. changed nothing. I suspect I didn't upgrade anything there. [08:28:09 CEST] <relaxed> mrelcee: freebsd's `make` is not gnu make- use gmake [08:28:31 CEST] <mrelcee> ooo. it's compiling.. [08:30:12 CEST] <mrelcee> thanks for the help.. hopefully it all continues smoothly [08:35:16 CEST] <relaxed> you're welcome [08:37:58 CEST] <mrelcee> you're my new favorite person. :) [08:38:20 CEST] <mrelcee> i've been beaitng my head on this for a few hours.. [08:45:16 CEST] <mrelcee> It successfully compiled. yay. thanks again.. [11:00:12 CEST] <wjvds> I am trying to stream from v4l2 to ffserver but when I use the server (http://localhost:8090/feed1.ffm) as output I get the error Input picture width (1920) is greater than stride (736) Video encoding failed but when I output to a file I (test.mp4) everything works fine [11:02:44 CEST] <wjvds> My full command and output is http://pastebin.com/eQneFz3G [11:02:57 CEST] <mgbowman> is ffserver still maintained? [11:41:49 CEST] <platzhirsch> JEEB: asked on the mailing list and I guess you're right but speed could be utilized for monitoring purposes if I understood it now correctly, because if it falls below 1.0 the transcoding cannot keep up (in terms of live streaming) [11:45:30 CEST] <JEEB> yeah I was mostly noting that the quantizer value shouldn't be used [11:45:39 CEST] <JEEB> the rest could be utilized I guess [14:58:06 CEST] <osiris> hi guys [14:58:57 CEST] <osiris> honestly i am googling for ours already, is there a place, where I can download compiled ffmpeg version including vvapi and/or intel quick sync? [15:04:30 CEST] <jkqxz> Distributions may make builds with vaapi. A prebuilt version outside distribution setup doesn't really work because libva is really just a dlopen() wrapper. [15:07:09 CEST] <osiris> jkqxz: isnt it a intel based thing? I looked into the building process and its reall huge effort, and if static versions are delivered with some media servers, i thought an ffmpeg with quick snyc or something should be availeable somewhere as well [15:07:13 CEST] <jkqxz> The Intel stuff is non-free, so you can't distribute a binary linked with libmfx. Linking with the free mfx_dispatch can work, but you need a matching media sdk install to use it so it wouldn't really be very useful. [15:07:46 CEST] <osiris> ah, that explains a lost [15:08:24 CEST] <osiris> intel should cahnge their mind and give a better license modell, that sucks and they would benift from it [15:08:38 CEST] <osiris> s/benift/benefit [15:10:07 CEST] <jkqxz> The open source stuff with vaapi is all fine licence-wise, it's just that the driver dependency messes up making a static build. Does your distribution not contain a suitable version already? [15:11:01 CEST] <osiris> my media server is 15.10 ubuntu [15:11:08 CEST] <osiris> but will be 16.04 soon [15:11:12 CEST] <osiris> or maybe arch [15:11:41 CEST] <osiris> butt vaapi isnt built in the ffmpeg of ubuntu [15:11:46 CEST] <osiris> as far as i can check [15:12:35 CEST] <osiris> no --enable=vaapi [15:14:21 CEST] <osiris> and i have vaapi active with kodi already, but try to get transcoding working with ffmpeg now [15:16:10 CEST] <osiris> so do u know how I get ffmpeg to make use out of it? [15:16:16 CEST] <osiris> vainfo: VA-API version: 0.39 (libva 1.7.1) [15:17:02 CEST] <jkqxz> Hmm. Ubuntu 16.04 is still on the 2.8 branch and current development is on the 3.0 branch, so that doesn't help. (The vaapi support in the ffmpeg utility is recent; added between 3.0 and 3.1.) [15:17:35 CEST] <jkqxz> So I think you'd need to build it yourself to use on those distributions. [15:17:53 CEST] <osiris> damn [15:18:26 CEST] <osiris> usually there is always a "special" ppa [15:19:54 CEST] <jkqxz> There probably is somewhere, but nothing official as far as I know. [15:19:56 CEST] <osiris> but thx for answering, really apreciate it [15:20:21 CEST] <osiris> dont need official ;) [15:52:24 CEST] <osiris> this library installing is awfull [15:52:46 CEST] <osiris> feels like already 100 installed and still some missing [16:38:25 CEST] <soulshock> anyone played with Netflix's VMAF? I'm wondering if there's a way to avoid creating YUV files for each comparison and maybe generating them on-the-fly [17:41:43 CEST] <platzhirsch> xit [18:37:07 CEST] <whald> can someone tell me if a h264 frame that was handed to me through avcodec_receive_frame may still be read from the h264 decoder? [18:37:34 CEST] <whald> (i assume that it will at least not write to the buffers after giving me the frame) [18:40:55 CEST] <jkqxz> Yes. The H.264 decoder will keep a reference to it and read from it for as long as it has to be kept to decode other frames (which can be arbitrarily long). [18:43:02 CEST] <whald> jkqxz, i kind of expected that. thanks for the clarification. [18:44:16 CEST] <jkqxz> If you want to write on a frame from the decoder, av_frame_is_writable() can test whether you hold the only reference and av_frame_make_writable() will make a copy if necessary to leave you with a writable frame. [18:45:37 CEST] <whald> jkqxz, i'm decoding diretly to dma-bufs allocated by an intel gpu and think i have some cache cohereny problems causing artifacts on the screen. but the performance is sweet. :-) [18:47:31 CEST] <whald> jkqxz, i'm doing this exercise to get rid of unneccesary copies, so i'll have to find a way to do this another way. i this flushing the cpu caches or the like after avcodec_receive_frame but before telling the gpu to use the buffers might be enough [18:47:34 CEST] <jkqxz> Yeah, ok - that only works if your frames are in normal memory. [18:48:10 CEST] <jkqxz> You know that Intel GPUs contain a hardware decoder for H.264, which decodes directly into GPU surfaces? [18:49:02 CEST] <whald> jkqxz, i'm using this for yuv420p, but intel does not support yuv444p, so... [18:50:07 CEST] <jkqxz> Yeah, noone makes hardware for 4:4:4 decode. [19:25:31 CEST] <dustobub> Hey! I'm encoding an h264 mp4 and am having a small issue with the output byte size of the first frame vs. all other frames. The first frame always ends up being ~2x the byte size of other frames. Is there any way to control this behavior? My usecase requires that the first frame be very similar in size compared to other frame. Thanks. [19:26:28 CEST] <dustobub> ffmpeg -framerate 60 -i "C:\XYZ%4d.tif" -c:v libx264 -g 1 -vf "scale=3840:2160" -crf 19 -pix_fmt yuv420p C:\XYZ.mp4 [19:27:09 CEST] <dustobub> Using ffprobe I see that pkt_size of the first frame is always ~2x compared to other frames [19:50:55 CEST] <DHE> dustobub: key frames are bigger than other frames. this is quite normal. [19:51:56 CEST] <DHE> there is an option, intra-refersh, which will reduce keyframe size but it has other drawbacks [19:52:05 CEST] <DHE> oh, -g 1. didn't see that [19:54:28 CEST] <dustobub> DHE: but all of my frames are keyframes. Just the first frame always ends up being larger for some reason. [19:54:55 CEST] <dustobub> If I encode a dummy black frame as the first frame I see these results: [19:54:57 CEST] <dustobub> frame|pkt_size=2173|pict_type=I [19:54:57 CEST] <dustobub> frame|pkt_size=466255|pict_type=I [19:54:57 CEST] <dustobub> frame|pkt_size=430179|pict_type=I [19:54:57 CEST] <dustobub> frame|pkt_size=416652|pict_type=I [19:54:57 CEST] <dustobub> frame|pkt_size=411401|pict_type=I [19:54:58 CEST] <dustobub> frame|pkt_size=407174|pict_type=I [19:55:00 CEST] <dustobub> frame|pkt_size=405377|pict_type=I [19:55:02 CEST] <dustobub> frame|pkt_size=403207|pict_type=I [19:55:04 CEST] <dustobub> frame|pkt_size=401588|pict_type=I [19:55:06 CEST] <dustobub> frame|pkt_size=401200|pict_type=I [19:55:30 CEST] <dustobub> But without that dummy frame I see this: [19:55:36 CEST] <dustobub> frame|pkt_size=1377043|pict_type=I [19:55:36 CEST] <dustobub> frame|pkt_size=406953|pict_type=I [19:55:36 CEST] <dustobub> frame|pkt_size=407200|pict_type=I [19:55:36 CEST] <dustobub> frame|pkt_size=406647|pict_type=I [19:55:48 CEST] <dustobub> [...] [20:01:42 CEST] <dustobub> I posted some more details on stackoverflow - http://stackoverflow.com/questions/38532700/ffmpeg-first-frame-byte-size-ov… [20:01:59 CEST] <dustobub> But let me know if you have any ideas. Thanks. [20:15:22 CEST] <apoms> Hi all. Is it possible to setup and initialize an AVFormatContext without probing the input file? I would like to be able to preprocess a file once to discover its format and then each time I open it in the future I could just start decoding frames without having to probe for the container and stream information. [20:50:00 CEST] <CFS-MP3> Slightly offtopic... trying to get Intel QVC to work... I compiled ffmpeg OK, but it seems to require a linux kernel patch that might not be available for 4.4... any suggestion? [20:56:27 CEST] <jkqxz> CFS-MP3: Change your OS to the one supported CentOS version. (The Intel Media SDK is an immense pain on Linux. Unless you really need the stuff in it, use VAAPI to access your QuickSync hardware instead.) [21:04:36 CEST] <soulshock> hm what's the difference between -psnr and -filter_complex "psnr" ? [21:05:22 CEST] <soulshock> as in what does -psnr use as reference video and what does the filter use as reference video? [21:09:49 CEST] <CFS-MP3> jkqxz I was hoping for a solution that didn't imply switching to CentOS...that has too many implications at this point, so I think I'm going to pass until intel releases a decent patch [21:15:18 CEST] <kepstin> soulshock: the -psnr option to the x264 encoder causes x264 to print out the psnr between the input and output of the encoder. [21:15:40 CEST] <kepstin> i think [21:15:45 CEST] Action: kepstin should double-check that [21:16:55 CEST] <kepstin> might work with some other encoders too, not sure what the full list is [21:18:55 CEST] <soulshock> x264 is fine that's what I'm using. I'm asking because I'm getting wildly different values between -psnr and -filter_complex "psnr" [21:23:20 CEST] <soulshock> I guess what you're saying is the -psnr option only compares the frames sent to x264 whereas -filter_complex compares the input video to the output video. meaning -psnr does not take into consideration any resizing or deinterlacing [21:24:14 CEST] <soulshock> and -filter_complex "psnr" does include any additional filters applied to the source, i.e. deinterlacing and resizing (I'm upscaling to source resolution when calculating psnr) [21:37:03 CEST] <soulshock> now I'm thinking if it's possible to use filter "split" on output of x264 and write one to file and the other as input for psnr [21:41:42 CEST] <DHE> soulshock: filter_complex analyzes the 2 videos using ffmpeg filters. -psnr asks x264 for its opinion on the subject [21:42:26 CEST] <soulshock> ok. so I need filter_complex for what I want. thanks [21:43:39 CEST] <soulshock> hm maybe not. I'm confusing myself [21:44:22 CEST] <dustobub> per my CRF issue above I have found that the first frame when encoding with libx264 uses a QP value half of where it should be [21:44:24 CEST] <dustobub> [libx264 @ 05380b60] frame= 0 QP=14.92 NAL=3 Slice:I Poc:0 I:32400 P:0 SKIP:0 size=1485053 bytes [21:44:24 CEST] <dustobub> [libx264 @ 05380b60] frame= 1 QP=29.48 NAL=3 Slice:I Poc:0 I:32400 P:0 SKIP:0 size=361196 bytes [21:44:24 CEST] <dustobub> [libx264 @ 05380b60] frame= 2 QP=29.48 NAL=3 Slice:I Poc:0 I:32400 P:0 SKIP:0 size=359406 bytes [21:44:34 CEST] <dustobub> As discussed in http://stackoverflow.com/questions/38532700/ffmpeg-first-frame-byte-size-ov… [21:45:32 CEST] <dustobub> Does anyone know if it's possible to get CRF encoding to handle that first frame like other frames? Most people probably wouldn't care, but my use-case depends on having that first frame similar to other frames. [21:55:55 CEST] <audioblazer> Hello, i'm new to ffmpeg and am having a weird issue when accessing a input file I use in my command after I run the line [21:57:09 CEST] <audioblazer> It says there was invalid data found when processing the input (although was valid prior to the run) and and says moov atom not found [22:25:49 CEST] <Mavrik> audioblazer, you have a broken video file. [00:00:00 CEST] --- Sat Jul 23 2016

1 0

ffmpeg-devel.log.20160721
by burek 22 Jul '16

22 Jul '16

[00:00:02 CEST] <sanmarcos> PJSIP is rendering the OpenH264 decoded frames to OpenGL [00:00:19 CEST] <nevcairiel> you would need to ask the openh264 people then to fill the frame black [00:00:25 CEST] <sanmarcos> (there is the alternative of using ffmpeg instead of OpenH264) [00:00:28 CEST] <nevcairiel> with external libraries we have no control [00:00:51 CEST] <sanmarcos> so you say it is an OpenH264 issue [00:01:25 CEST] <nevcairiel> well for a certain definition of issue [00:01:45 CEST] <nevcairiel> its certainly their decision to init the frames zero instead of coloring them [00:02:15 CEST] <sanmarcos> right my theory is that when we start getting frames we are missing a keyframe, and it is all inited to 0 YUV, which turns out to be green for a few seconds [00:02:19 CEST] <sanmarcos> then we actually get video [00:05:16 CEST] <sanmarcos> ok, I either need to find where Openh264 is doing this, or find a way of being notified when we actually get proper video frames to display [00:05:39 CEST] <BBB> jamrial: can you explain what your stack alignment patch does? it seems like it gets rid of per-size stack alignment macros, doesnt that break e.g. on msvc where stack is aligned by 16 byte on 64bit, but not 32byte? [00:05:48 CEST] <BBB> jamrial: or do I fundamentally misunderstand what that stuff does? [00:07:18 CEST] <sanmarcos> https://github.com/cisco/openh264/issues/1163 [00:07:21 CEST] <TD-Linux> sanmarcos, you can also wait for the first I frame before feeding frames in [00:07:22 CEST] <sanmarcos> that seems to be related to what I am facing [00:07:35 CEST] <sanmarcos> TD-Linux: I will try to do that, do you have any hints on how to do so? [00:07:42 CEST] <sanmarcos> I am new to all of this [00:09:31 CEST] <TD-Linux> sanmarcos, for H.264 in RTP I think you have to parse the beginning of the H.264 stream. OpenH264 might have an API for this but I'm not sure. [00:09:48 CEST] <iive> sanmarcos: other would know better, but there is h264 feature that allows the stream to not have I-frames at all. Instread the bitstream contains a number of frames that have to be discarded, before the output should recover. [00:10:01 CEST] <TD-Linux> note that if you are missing an I frame, you can also request for the sender to send one via a RTCP message [00:11:02 CEST] <TD-Linux> iive, that's cyclic refresh, and it's not commonly used for this [00:12:15 CEST] <sanmarcos> iive: hmmm? ok [00:12:29 CEST] <sanmarcos> I will take a look at the options in pjsip I have available and will read some of the openh264 source code [00:12:35 CEST] <sanmarcos> doesnt help that pjsip only supports an old version of openh264 [00:16:44 CEST] <sanmarcos> https://trac.pjsip.org/repos/browser/pjproject/trunk/pjmedia/src/pjmedia-co… [00:23:37 CEST] <jamrial> BBB: i'm simply getting rid of the extra configure variables and unnecessary ifdeffery. the macros are functionally the same before and after the patch [00:24:02 CEST] <nevcairiel> so why did we have those in the first place [00:25:20 CEST] <jamrial> it used to be two, align8 and align16. i guess at some point some arch may have not supported one or the other, or whoever wrote that assumed there could be in the future [00:25:30 CEST] <jamrial> then i added align32 for avx stuff some time ago [00:25:38 CEST] <sanmarcos> ok, I commentd out the openh264 code, it is not openh264 [00:25:42 CEST] <jamrial> i now check and every arch that enables them enables them all [00:25:46 CEST] <sanmarcos> it is something before that is causing the opengl buffer to be green [00:26:08 CEST] <sanmarcos> I made oh264_got_decoded_frame and oh264_codec_decode do nothing [00:26:11 CEST] <BBB> jamrial: ok, very strange, I thought stack was only aligned by N bytes on some platforms [00:26:15 CEST] <BBB> but if it works its ok [00:26:48 CEST] <nevcairiel> well thats what those macros are for, doing higher alignment [00:48:05 CEST] <BBB> nevcairiel: I thought they were wrappers for if platform/compiler/whatever supports alignment by N, do this, else, do this hacky thing at runtime" [00:48:10 CEST] <BBB> nevcairiel: or is that something else? [00:48:38 CEST] <BBB> nevcairiel: and then the switch between if/else was the HAVE_LOCAL_ALIGNED_N values in config.h [00:48:43 CEST] <nevcairiel> they are, but apparently we dont need the complex size-specific complexity we had [00:48:55 CEST] <nevcairiel> hm too much complex [00:50:29 CEST] <BBB> well if this works on msvc 64bit I dont think I have anything to complaina bout [00:50:48 CEST] <nevcairiel> 64-bit never really had all the big troubles since it even comes with aligned stack [00:50:55 CEST] <BBB> right, but 16 [00:51:05 CEST] <BBB> so does that work for avx2 optimizations requiring 32byte? [00:51:56 CEST] <nevcairiel> yasm code would re-align when avx2 is used, and if any stack-data is passed into such a function then thats where LOCAL_ALIGNED_32 comes in [00:53:22 CEST] <nevcairiel> the only thing these macros did was and no longer do after the patch is allowing fine-grained configure control over which of the alignments is enabled [00:53:32 CEST] <nevcairiel> ... which was not used by anything in configure, as jamrial found out [00:55:39 CEST] <nevcairiel> anyway sleepz [00:58:00 CEST] <sanmarcos> an I-frame is a key frame? [00:58:35 CEST] <jamrial> i did not remove the per-size macros in the patch [01:00:11 CEST] <iive> sanmarcos: I stands for intra and that means stand alone. [01:01:03 CEST] <nevcairiel> although the definition of key-frame is a bit more complex in codecs like h264, an I frame is not necessarily a random-access-point [01:01:29 CEST] <iive> idr are [01:01:43 CEST] <JEEB> IRAP is nowadays the proper way to note it [01:01:48 CEST] <sanmarcos> inter frame is a keyframe [01:01:57 CEST] <sanmarcos> I-frame is inter frame? [01:02:04 CEST] <JEEB> intra. inter has prediction between things [01:02:30 CEST] <iive> internet - inter net , connected networks [01:02:41 CEST] <sanmarcos> so to properly decode H264 you first need the first intra frame [01:02:48 CEST] <sanmarcos> and then you can have frames that are diffs on that intra frame? [01:03:13 CEST] <JEEB> no [01:03:34 CEST] <JEEB> read the AVC specification to know what sort of intra frames are such that you can start random access from [01:03:51 CEST] <JEEB> one of them is IDR and the other one IIRC is intra+a specific SEI message (with open gop) [01:04:38 CEST] <JEEB> intra frame just means that you can (probably) decode *that* frame, it doesn't mean that you can decode the ones that come after it in coding order [01:04:40 CEST] <nevcairiel> the sei even works entirely without intra [01:04:52 CEST] <JEEB> ok [01:04:55 CEST] <TD-Linux> sanmarcos, I frame = keyframe = intra frame [01:05:00 CEST] <iive> sometimes you need 2 i-frames, because you have b-frames (bidirectional frames) [01:05:17 CEST] <nevcairiel> well you need some sort of intra data, but you can have only intra slices [01:05:20 CEST] <TD-Linux> iive, that's usually I and P [01:05:25 CEST] <iive> and h264 could use 15 different frames for predictions. [01:05:41 CEST] <TD-Linux> openh264 does not use B frames anyway [01:05:45 CEST] <JEEB> TD-Linux: and in your definition key != IRAP, right? [01:05:57 CEST] <JEEB> then I agree :) [01:06:27 CEST] <JEEB> some people (and I think libavcodec currently) tends to mish-mash these definitions :P [01:06:32 CEST] <TD-Linux> JEEB, right, and of course h265 has IDR vs I [01:06:51 CEST] <TD-Linux> the definitions are codec specific [01:07:32 CEST] <JEEB> well, HEVC just defines it better so I like to use its umbrella term with AVC as well [01:07:50 CEST] <JEEB> even though for AVC the stuff that's under definition of IRAP isn't just defined as a picture type [01:08:11 CEST] <JEEB> but rather there's various extra info which is required for something not IDR to be an IRAP [01:08:50 CEST] <JEEB> and HEVC has the same kinds of I and IDR, and then it has those other picture types that note different kinds of IRAP [01:09:01 CEST] <nevcairiel> doesnt hevc have those definitions for random access points as well under some other funny acronyms? [01:09:09 CEST] <JEEB> yes, BLA etc [01:09:17 CEST] <JEEB> but at least those are picture types [01:09:30 CEST] <JEEB> so you get the picture and know if it's an IRAP right away [01:09:41 CEST] <TD-Linux> HEVC has both slice types and frame types [01:09:43 CEST] <JEEB> without additional messages. except maybe for the rolling one [01:10:17 CEST] <JEEB> TD-Linux: a picture can be a field or a frame - let's just use the wording "picture" since that's what's used in the specs as well? :( [01:10:25 CEST] Action: TD-Linux like the generic term 'keyframe' for seekable frames [01:11:20 CEST] <JEEB> a keyframe is a picture that can be decoded by itself, it doesn't make any promises regarding anything that comes after in coded order [01:11:22 CEST] <TD-Linux> ugh now you brought up interlacing. time for bed. [01:11:31 CEST] <JEEB> unless you define it specifically [01:11:56 CEST] <JEEB> I think I actually defined the term once for matroska since it utilizes the wording "keyframe" which is often not specific enough [01:12:22 CEST] <TD-Linux> JEEB, matroska is now being standardized at CELLAR (I was at the meeting tuesday), so now's the time to fix it [01:13:15 CEST] <JEEB> this was a thing in '13 [01:13:16 CEST] <JEEB> https://lists.matroska.org/pipermail/matroska-devel/2013-September/004567.h… [01:13:36 CEST] <nevcairiel> thats typical for containers to use that term and flag [01:13:57 CEST] <JEEB> ISOBMFF called it a random access point at least, IIRC? [01:14:03 CEST] <JEEB> but yes, it's often used [01:16:53 CEST] <JEEB> also it seems like I defined the flag worse than I thought. mostly the "random access point" mention, after that it gets into the format specifics [03:25:14 CEST] <cone-940> ffmpeg 03Michael Niedermayer 07master:3af1aba2be02: avformat/mux: Fix some codecpar non uses [03:42:34 CEST] <Chloe> michaelni: so I've checked everything thrice, rebased, pushed to my test fork (which looks fine): https://github.com/joshdekock/FFmpeg/commits/master Do I just push to FFmpeg now? [03:46:16 CEST] <cone-940> ffmpeg 03James Almer 07master:ec0e888e3aeb: Changelog: move a misplaced entry to the correct section [03:46:17 CEST] <cone-940> ffmpeg 03James Almer 07master:66408fce493b: avformat: add an Ogg Video muxer [03:46:18 CEST] <cone-940> ffmpeg 03James Almer 07master:120f34b6ac06: avformat/oggenc: add vp8 muxing support [03:46:28 CEST] <Chloe> :/ [03:50:55 CEST] <michaelni> Chloe, yes once you rebase again ;) [03:51:17 CEST] <cone-940> ffmpeg 03Josh de Kock 07master:9134d2df5b17: libopenmpt: add subsong support [03:51:18 CEST] <cone-940> ffmpeg 03Josh de Kock 07master:4ed6edac7627: docs/demuxers: add libopenmpt section [03:51:19 CEST] <cone-940> ffmpeg 03Jörn Heusipp 07master:840df1f19314: libopenmpt: set stream duration and fix time base [03:51:20 CEST] <cone-940> ffmpeg 03Jörn Heusipp 07master:64131b87d64d: libopenmpt: Add "date" to metadata. [03:51:26 CEST] <Chloe> ahh scary [03:51:53 CEST] <Chloe> michaelni: I guess I dont need to comment on my own patches that they were push on the ML, or all? [03:52:03 CEST] <Chloe> pushed* [03:52:50 CEST] <michaelni> its probably best if you or someone does reply with an applied or something to the ML [03:53:04 CEST] <michaelni> avoids confusion [03:53:21 CEST] <Chloe> ok thanks :) [11:26:54 CEST] <ubitux> > It seems that kVTDecodeFrame_EnableTemporalProcessing is just a suggestion to VideoToolbox, and one that it ignores. That means that, for now, this code only outputs frames in the correct order for I-frame only video. [11:26:56 CEST] <ubitux> lol [11:26:58 CEST] <ubitux> what a shit framework [11:27:42 CEST] <ubitux> said differently, "it means that this code only outputs frames in the correct order when they are already ordered" [11:31:55 CEST] <nevcairiel> I/P should also not have re-ordering, should it [11:32:09 CEST] <ubitux> yeah that's the thing [11:32:11 CEST] <nevcairiel> sounds like the VT API is entirely useless [11:32:41 CEST] <ubitux> well you "just" have to reorder yourself [11:32:42 CEST] <ritsuka> the docs say kVTDecodeFrame_EnableTemporalProcessing is useful only if you want videotoolbox to deinterlace the video :| [11:32:54 CEST] <ubitux> ritsuka: wut [11:32:57 CEST] <ubitux> what doc btw? [11:33:41 CEST] <ritsuka> the headers [11:34:43 CEST] <ritsuka> "The kVTDecodeFrame_EnableTemporalProcessing bit indicates whether the decoder may delay calls to the output callback so as to enable processing in temporal (display) order." [11:34:46 CEST] <ritsuka> and in kVTDecompressionProperty_DeinterlaceMode_Temporal [11:34:52 CEST] <ritsuka> this mode is only used if kVTDecodeFrame_EnableTemporalProcessing is set, otherwise a non-temporal mode (eg, VerticalFilter) will be used instead [11:35:41 CEST] <ubitux> lol [11:51:07 CEST] <cone-147> ffmpeg 03Mark Reid 07master:541731488e2d: libavformat/rawenc: add dnxhr raw format extension [11:51:08 CEST] <cone-147> ffmpeg 03Michael Niedermayer 07master:9264bb7e79a5: avfilter/vf_psnr: Fix rounding error in average_max [14:16:07 CEST] <omerjerk> I'm porting some old code to ffmpeg and getting the following error - [14:16:08 CEST] <omerjerk> error: use of undeclared identifier 'x86_reg' [14:16:17 CEST] <omerjerk> is it okay to convert it to int? [14:21:41 CEST] <omerjerk> ah, nvm, just needed to incude libavutil/x86/asm.h [14:57:25 CEST] <omerjerk> can anyone tell me the significance of this constant? - https://github.com/FFmpeg/FFmpeg/blob/master/libavcodec/x86/constants.c#L84 [14:58:04 CEST] <omerjerk> I'm getting ff_pd_2 undefined error. so my be I can use some logic [15:05:12 CEST] <kierank> it's 1 as a doubleword repeated [15:05:21 CEST] <kierank> replace the 1 with a 2 [15:05:47 CEST] <omerjerk> oh. that will be all? [15:06:00 CEST] <kierank> yes [15:06:06 CEST] <omerjerk> thanks [16:46:24 CEST] <Chloe> I'm looking at https://trac.ffmpeg.org/ticket/5366, and it seems audiounits support runtime querying of both audiounits and their parameters. I'm just wondering how this would work with FFmpeg options, as the parameters are named. Maybe something like -par1 -par2...etc? Is there any support for custom option names and types? I guess I'm trying to take a shortcut here and implement a generic audiounits filter instead of one-per-au (there are 32 [16:46:24 CEST] <Chloe> builtin filters). [17:38:11 CEST] <nevcairiel> other things use a syntax like filter=option1=value1:option2=value2 [17:38:59 CEST] <nevcairiel> and we have functions to parse and split such strings into key/value pairs [23:18:17 CEST] <cone-147> ffmpeg 03Michael Niedermayer 07master:38cc5c298c98: avformat/mov: Do not allocate empty extradata [00:00:00 CEST] --- Fri Jul 22 2016

1 0

ffmpeg.log.20160721
by burek 22 Jul '16

22 Jul '16

[00:00:48 CEST] <CFS-MP3> the windows version? While ffmpeg is running in linux, what I'm personally using is windows. Anyway you guys can take a look at the stream with any player you want [00:01:02 CEST] <CFS-MP3> http://54.73.134.173/live/ngrp:Mosaic.stream_all/playlist.m3u8 [00:01:20 CEST] <CFS-MP3> it should be obvious to notice the problem [00:01:30 CEST] <furq> the windows version can open that url [00:05:33 CEST] <furq> this looks fine to me [00:05:45 CEST] <furq> aside from the fact that nigel farage is on my screen now [00:06:17 CEST] <CFS-MP3> furq watch for 30 seconds or so [00:06:24 CEST] <furq> i was watching for about two minutes [00:06:29 CEST] <furq> i didn't notice any playback issues [00:06:33 CEST] <CFS-MP3> mpv has the same problem, but it complains a bit [00:06:35 CEST] <CFS-MP3> V: 00:00:22 Dropped: 1 Cache: 8s+0KB [00:06:35 CEST] <CFS-MP3> [ffmpeg] skipping 1 segments ahead, expired from playlists [00:06:35 CEST] <CFS-MP3> V: 00:00:52 Dropped: 1 Cache: 6s+0KB [00:06:35 CEST] <CFS-MP3> [ffmpeg] skipping 1 segments ahead, expired from playlists [00:06:35 CEST] <CFS-MP3> V: 00:00:59 Dropped: 1 Cache: 0s+0KB [00:06:35 CEST] <CFS-MP3> [ffmpeg] http: Stream ends prematurely at 429240, should be 5935912 [00:06:35 CEST] <CFS-MP3> V: 00:00:59 Dropped: 1 Cache: 0s+0KB [00:06:36 CEST] <CFS-MP3> [ffmpeg/demuxer] mpegts: PES packet size mismatch [00:06:36 CEST] <CFS-MP3> V: 00:01:09 Dropped: 1 Cache: 10s+0KB [00:06:37 CEST] <CFS-MP3> [ffmpeg/video] h264: Invalid NAL unit 0, skipping. [00:06:37 CEST] <CFS-MP3> [ffmpeg/video] h264: corrupted macroblock 65 26 (total_coeff=-1) [00:06:38 CEST] <CFS-MP3> [ffmpeg/video] h264: error while decoding MB 65 26 [00:06:41 CEST] <furq> huh [00:06:48 CEST] <furq> i didn't get any mpv messages [00:07:13 CEST] <CFS-MP3> did you play it for a bit, like 30 seconds or so? [00:07:20 CEST] <furq> 23:06:24 ( furq) i was watching for about two minutes [00:07:25 CEST] <CFS-MP3> damn [00:11:13 CEST] <furq> have you got wowza doing adaptive bitrate transcoding or something [00:11:26 CEST] <furq> maybe that's the culprit [00:11:43 CEST] <CFS-MP3> furq yes... I'd love to blame wowza here, but it's running the rest of the streams perfectly [00:12:02 CEST] <furq> i take it you've got plenty of bandwidth client-side [00:12:20 CEST] <CFS-MP3> yes, it's obscene [00:12:34 CEST] <furq> although i'd expect a better message if mpv was dropping you to the next bitrate [00:13:02 CEST] <kepstin> I wonder if you might be hitting issues with blocking writes to the pipes between the ffmpeg processes (raw video frames are often larger than the pipe buffer) [00:14:33 CEST] <furq> i'll leave it idle for a while and see if mpv shows any warnings [00:14:45 CEST] <CFS-MP3> kepstin it happened without the pipes as well ... the pipes are an attempt to split the work between several ffmpeg processes if that was the problem, that's all [00:15:06 CEST] <CFS-MP3> I think the "expired for playlists" could be some kind of clue [00:15:18 CEST] <thebombzen> how often is the stuttering? [00:15:21 CEST] <CFS-MP3> This seems somewhat related: https://trac.ffmpeg.org/ticket/3483 [00:15:27 CEST] <CFS-MP3> thebombzen like every 15-20 seconds [00:15:33 CEST] <thebombzen> okay lemme watch a bit more than [00:15:35 CEST] <thebombzen> then* [00:16:51 CEST] <thebombzen> CFS-MP3: I just watched it with VLC on Linux for 1 minute w/o seeing stuttering [00:17:09 CEST] <thebombzen> I did File->Open network stream, and copy/pasted the URL [00:17:28 CEST] <thebombzen> VLC version is 3.0.0-git, i.e. I compiled it myself from source [00:17:46 CEST] <thebombzen> sounds like a client/network transmission error [00:24:12 CEST] <furq> i've had it running for ~10 mins and not got any warnings from mpv [00:24:59 CEST] <furq> also i don't think i've seen this episode of room 101 [00:36:41 CEST] <CFS-MP3> thanks for checking furq [01:09:19 CEST] <CFS-MP3> OK the new problem is that the DAR is going to shit in the composed mosaic :-) And setdar doesn't fix it [01:10:43 CEST] <CFS-MP3> ignore, I'm just too tired, sorry [06:54:55 CEST] <ferdna> guys whats a recommended bitrate for a ip camera? [06:55:06 CEST] <ferdna> i need to stream decent vp9 video [08:18:14 CEST] <Ashish_> hi all [08:18:26 CEST] <Ashish_> we installed ffmpeg on our centos server. [08:18:42 CEST] <Ashish_> we forgot to configure libfreetype module with it. [08:18:56 CEST] <Ashish_> can anyone help us to do it ? [08:33:41 CEST] <soulshock> https://trac.ffmpeg.org/wiki/CompilationGuide/Centos [08:34:08 CEST] <soulshock> when running ./configure remember to include --enable-libfreetype [08:38:23 CEST] <Ashish_> i have alreday installed ffmpeg [08:38:45 CEST] <Ashish_> do you have any idea how to compile it on installed version ? [08:48:18 CEST] <soulshock> no. you would have to recompile it from scratch [08:50:59 CEST] <Ashish_> ok [08:51:28 CEST] <Ashish_> do you have idea from where i can get source directory of libfreetype ? [09:01:57 CEST] <soulshock> I am guessing you can install the freetype2-devel package? [09:07:33 CEST] <Ashish_> is libfreetype and freetype are same ? [09:07:41 CEST] <Ashish_> libfreetype and freetype are same ? [09:20:07 CEST] <soulshock> yes [09:20:14 CEST] <soulshock> I should think sop [09:21:04 CEST] <Ashish_> ok [10:13:26 CEST] <DelphiWorld> yo guys! [10:13:32 CEST] <DelphiWorld> question guys... [10:13:41 CEST] <DelphiWorld> when should i use H.264 base, main or high profile? [10:13:46 CEST] <DelphiWorld> how could i chouse bethwan? [10:39:43 CEST] <dv_> are a52 and ac-3 really the same? why are there two names? [10:41:16 CEST] <dv_> iirc a52 is an atsc standard, and describes the use of ac-3 in atsc? [15:44:10 CEST] <netw1z> I compiled FFMPEG from src on raspberry pi 3 but im getting relocation error: /usr/local/lib/libavfilter.so.5 when i try to run it [15:50:20 CEST] <furq> netw1z: are you sure that's the library you compiled [15:51:20 CEST] <netw1z> i see the file in /usr/local/lib/libavfilter.so.5 [17:09:25 CEST] <platzhirsch> FFMPEG posts progress and also outputs progress metadata in verbose mode. Where can I read about the meaning of.. time, quality, speed? [17:16:28 CEST] <JEEB> not quality [17:16:34 CEST] <JEEB> that is something you can't calculate [17:17:00 CEST] <JEEB> it's most probably some sort of quantizer value that you shouldn't try to derive anything out of esp. since I'm not sure if it works with various library-based encoders nowadays [17:17:08 CEST] <platzhirsch> what does it then mean JEEB ? You get a q value for each variant [17:17:51 CEST] <platzhirsch> https://ffmpeg.org/pipermail/ffmpeg-user/2012-December/011978.html [17:24:14 CEST] <JEEB> platzhirsch: as I said it's whatever the avcodec stuff has as the quantizer value [17:24:33 CEST] <JEEB> I wouldn't base too much if anything on it, esp. if library-based stuff can be completely incorrect with it [17:43:36 CEST] <platzhirsch> still would be nice to understand it. Probably best to write to the mailing list [17:43:43 CEST] <platzhirsch> JEEB: [17:45:01 CEST] <JEEB> it's the quantizer value as noted by the avcodec struct. not sure if it's the last coded sample's or average. also if audio affects it or not [17:45:15 CEST] <JEEB> i guess you could only read the fabulous sauce [17:45:37 CEST] <JEEB> or ask on the ml if you don't know hpw to read the sauce [17:47:46 CEST] <platzhirsch> JEEB: but what about speed and progress/time? [17:49:37 CEST] <JEEB> that you might get some stats out of. also i think ffmpeg can nowadays output the stats to something specific so you don't have to parse the terminal output [17:50:25 CEST] <platzhirsch> JEEB: yeah, it's the -progress option, but still without proper documentation what each key means it's a bit blurry [18:01:11 CEST] <Jpacker> hello, I'm mixing two audio files together Using FFmpeg and it's working perfectly. I'm modifying the audio in it and allow the user to run the mixing code again, but am getting this error on the exact same files that work the first time. Any thoughts? [18:01:25 CEST] <Jpacker> [mov,mp4,m4a,3gp,3g2,mj2 @ 0xf7163000] Format mov,mp4,m4a,3gp,3g2,mj2 detected only with low score of 1, misdetection possible! [18:01:29 CEST] <Jpacker> [mov,mp4,m4a,3gp,3g2,mj2 @ 0xf7163000] moov atom not found [18:17:37 CEST] <Sashmo_> hey guys, I cant seem to get two different channel map layouts in my command, can anyone point out what I am missing? http://pastebin.com/tQSZGHLk [18:18:44 CEST] <Sashmo_> if I run only one audio, I can choose whichever layout I need, but when adding the second one, it always uses the last audio as the maping [18:47:31 CEST] <soulshock> Sashmo_: you are remapping the same channels twice [18:47:39 CEST] <soulshock> with only one input file? [18:47:44 CEST] <Sashmo_> right [18:47:46 CEST] <xx_pk_xx> subt [18:48:00 CEST] <Sashmo_> I want to make two outputs from the same source of 6 channels [18:48:25 CEST] <soulshock> ah [18:49:00 CEST] <soulshock> try -filter_complex "[0:1][0:2][0:3][0:4][0:5][0:6]amerge=inputs=6,split=2[aout1][aout2]" [18:49:52 CEST] <Sashmo_> soulshock: hmmm ok I'll try that [18:50:29 CEST] <soulshock> hm maybe that's a video only filter. I'm not sure [18:50:56 CEST] <soulshock> -filter_complex "[0:1][0:2][0:3][0:4][0:5][0:6]amerge=inputs=6, asplit=2[aout1][aout2]" [18:51:00 CEST] <soulshock> asplit being the key [18:51:09 CEST] <Sashmo_> yeah [18:51:11 CEST] <Sashmo_> its video only [18:51:14 CEST] <soulshock> https://ffmpeg.org/ffmpeg-filters.html#split_002c-asplit [18:51:38 CEST] <soulshock> "asplit works with audio input, split with video." [18:53:35 CEST] <Sashmo_> still getting the same [18:53:53 CEST] <Sashmo_> http://pastebin.com/7hpGRqn2 [18:55:44 CEST] <soulshock> need someone with a bit more ffmpeg knowledge than me then [19:21:18 CEST] <Sashmo_> still nothing.... http://pastebin.com/uNAGEHqm no matter where I put the map option, its still only using one [19:23:04 CEST] <kepstin> Sashmo_: what's ffmpeg's output with that command? [19:23:23 CEST] <Sashmo_> http://pastebin.com/v8apDmbh [19:23:26 CEST] <Sashmo_> kepstin: http://pastebin.com/v8apDmbh [19:24:27 CEST] <kepstin> ok, so it's making a single output file containing one video track and 2 audio tracks right now [19:25:08 CEST] <kepstin> what was your original goal? [19:26:50 CEST] <Sashmo_> kepstin: yup thats right, but one audio needs to be 5.1 mapping, and the other needs to be stereo [19:27:07 CEST] <kepstin> i guess your source has a 5.1 (6-channel) of audio, and you want the result to have two audio tracks - one 5.1, one stereo? [19:27:19 CEST] <Sashmo_> kepstin: 6 sepearte PCM audio [19:27:30 CEST] <kepstin> right, since it's mfx :/ [19:29:50 CEST] <Sashmo_> kepstin: it works fine normally, just mapping two different maps doesnt work, I must be missing something, I just cant figure out where it is [19:30:28 CEST] <kepstin> how do you want to convert the input audio to stereo? just grab front left/front right or use some downmixing? [19:32:56 CEST] <kepstin> i'm kind of confused by the fact that you're getting two stereo channels with that command, it looks like they should both be 5.1 [19:38:42 CEST] <kepstin> I suspect you probably don't want to use the '-channel_layout' option at all, but rather do all of the manipulation in filters [19:39:13 CEST] <kepstin> have the filterchain generate two outputs, one 5.1 and one stereo (perhaps downmixed via the pan filter), then the ffmpeg audio encoders will handle the rest [19:41:57 CEST] <kepstin> (probably makes sense to use the 'join' filter rather than 'amerge' too, so you have more control over the channel layout of the result) [20:01:34 CEST] <t4nk299> Apologies - ffmpeg n00b here until 2 days ago - I'm streaming the video display on my pinball game to twitch but dont have the CPU do do the x264 encoding. Can anyone recommend an appliance that could take a raw stream and reasonably run it through the codec and stream it? I know I can build a whitebox but I was looking for a decent appliance like an Intel NUC? I'd like 30FPS x264, with a lot of extra juice to resize windows an [20:03:19 CEST] <Sashmo_> kepstin: hmm, you think that will fix the mapping? [20:03:33 CEST] <Sashmo_> kepstin: I can get 2 audio outputs, just the mapping is incorrect, [20:04:07 CEST] <kepstin> well, as it is you'll get two outputs but both will have the same number of channels in the same order [20:04:23 CEST] <kepstin> if you want one to be 5.1 and one to be stereo, you'll have to use a filter to downmix [20:04:49 CEST] <kepstin> (and then have two outputs in the filter graph; one with the 5.1 and one with the stereo) [20:21:07 CEST] <shincodex> uh [20:21:15 CEST] <shincodex> does your shit compile in 64 bit mingw [20:28:20 CEST] <JEEB> shincodex: if you mean FFmpeg, yes. you can see how it gets built for both win32 and win64 with mingw-w64 on FATE [20:28:27 CEST] <JEEB> http://fatebeta.ffmpeg.org/ [20:30:44 CEST] <DelphiWorld> hi friends [20:30:49 CEST] <DelphiWorld> thank you kepstin [20:30:56 CEST] <DelphiWorld> vaapi is working [20:33:29 CEST] <Sashmo_> kepstin: but isnt filter complex allowing me to downmix? I get the output that i need, so I even did it as two seperate lines like "[0:1][0:2][0:3][0:4][0:5][0:6]amerge=inputs=6[aout1]" and then another "[0:1][0:2][0:3][0:4][0:5][0:6]amerge=inputs=6[aout2]" [20:44:17 CEST] <DelphiWorld> can't we stream a webm live? VP8 + OPUS? [20:45:19 CEST] <kepstin> the webm (matroska) format is streamable. Not one of the best choices for live streaming, since you need a header at the start to decode it - which means having a streaming media server that remuxes for each client. [20:45:29 CEST] <kepstin> probably better to use a segmented format or mpeg-ts. [20:45:54 CEST] <DelphiWorld> kepstin: thanks dude [20:46:04 CEST] <kepstin> (webm can be used with dash for livestreaming) [20:51:58 CEST] <DelphiWorld> kepstin: i could just push from ffmpeg to a dash remuxing server like nimble streamer [20:54:11 CEST] <wouter__> Hi, I tried following https://trac.ffmpeg.org/wiki/HWAccelIntro to use Quick Sync to encode a video track (command: ffmpeg -i crowd_run_1080p50.y4m -c:v h264_qsv -preset:v faster out.qsv.mp4) but I get the very descriptive error message "[h264_qsv @ 0x2afc3e0] Error initializing an internal MFX session"[ [20:54:52 CEST] <DelphiWorld> wouter__: hi dude [20:55:01 CEST] <DelphiWorld> i recomand you to forget qsv at this time, and use vaapi [20:55:04 CEST] <DelphiWorld> i am using it [20:55:11 CEST] <DelphiWorld> i tryed a lot but no luck with qsv [20:55:43 CEST] <wouter__> DelphiWorld: ok Ill try that [20:56:17 CEST] <DelphiWorld> wouter__: a lot easyer, and if you have the intel media server sdk just clean it cause it'll create a driver conflict [20:58:43 CEST] <wouter__> DelphiWorld: I am not using the intel media server SDK, but the libva and libdrm libraries [20:59:07 CEST] <DelphiWorld> wouter__: good for you so, and try libva ... you'll get happy like i did today [20:59:17 CEST] <wouter__> DelphiWorld: Replacing my command with ffmpeg -i crowd_run_1080p50.y4m -c:v h264_vaapi -preset:v faster out.qsv.mp4 gives me the error Impossible to convert between the formats supported by the filter 'Parsed_null_0' and the filter 'auto-inserted scaler 0' [21:00:18 CEST] <fujisan> ello i have a question i downloaded the snapshot of ffmpeg but now i see it after extraction on the desktop as exec how do i extract the codecs now im on OS X btw [21:00:34 CEST] <DelphiWorld> wouter__: hold on let me give my cmd: [21:01:27 CEST] <fujisan> hoi wouter [21:02:12 CEST] <DelphiWorld> LOL [21:02:37 CEST] <fujisan> what's so funny [21:02:48 CEST] <DelphiWorld> whois [21:02:54 CEST] <fujisan> Del phi world [21:03:11 CEST] <DelphiWorld> wouter__: see this: ffmpeg -vaapi_device /dev/dri/renderD128 -hwaccel vaapi -hwaccel_output_format vaapi -i input.mp4 -an -vf 'format=nv12|vaapi,hwupload' -c:v h264_vaapi output.mp4 [21:03:11 CEST] <fujisan> wablief [21:03:33 CEST] <wouter__> DelphiWorld: I just found that same command on the internet and it works :D thanks alot! [21:03:36 CEST] <fujisan> i need your codecs to make my nomachine faster VP 8 is slow -_- [21:03:39 CEST] <fujisan> H.264 [21:03:52 CEST] <DelphiWorld> wouter__: lulz, awesome! [21:03:57 CEST] <fujisan> wouter hoi [21:03:58 CEST] <DelphiWorld> thank to kepstin [21:04:11 CEST] <DelphiWorld> hold on wouter__ rebooting [21:04:32 CEST] <fujisan> (>".)> help me please someone if it's not too much trouble [21:04:47 CEST] <fujisan> ObsidianX: [21:07:17 CEST] <fujisan> Coderjoe: speak spicypixel for the Orphis spaceinvader s are going cuz of the flux spaam & SpaceManiac telex and telepathically ocrete d by K1rk Fletcher into a squarecircle kanzure friki insane codehero s Coderjoe and Chloe they also have a CounterPillow codebam fling [21:07:58 CEST] <fujisan> creative listening music while i wait for a reply [21:08:03 CEST] <fujisan> im good at these [21:08:11 CEST] <fujisan> if they annoy please tell me and i 'll stop [21:08:21 CEST] <DelphiWorld> hey guys [21:08:23 CEST] <DelphiWorld> back [21:08:25 CEST] <fujisan> also if you have an answer that would be cool :) [21:08:27 CEST] <Chloe> can you not mass highlight please. [21:08:31 CEST] <DelphiWorld> wouter__: good vappi so? [21:08:36 CEST] <fujisan> you missed my creative listening highlight [21:08:42 CEST] <fujisan> Chloe: okay stopped [21:08:50 CEST] <fujisan> thanks for the feedback [21:09:21 CEST] <fujisan> you are a hero tho Chloe [21:09:33 CEST] <fujisan> i really can tell since you replied the first [21:09:39 CEST] <fujisan> =] [21:10:07 CEST] <fujisan> i will pray for you when i have my campfire ritual tonight with my friends by the beach [21:10:13 CEST] <fujisan> and bless your name [21:10:24 CEST] <Chloe> fujisan: what are you trying to do? [21:11:01 CEST] <fujisan> im trying extract or download codecs from ffmpeg to put into a dir for nomachine to use with decoding [21:11:24 CEST] <fujisan> to save bandwidth and decrease my latency by 30% [21:12:41 CEST] <Chloe> fujisan: dont you just need to follow these instructions? https://www.nomachine.com/DT02L00065#4 [21:13:46 CEST] <fujisan> yes my issue is with the extractions of ffmpeg i did that and then i get an exec on OS X and run it and then i dont know what to do [21:14:12 CEST] <DelphiWorld> fujisan: ffmpeg isn't graphical interface [21:15:01 CEST] <Chloe> fujisan: then you need to install the codec license presumably. This doesn't look like an FFmpeg issue though [21:15:19 CEST] <fujisan> then i need to read the man page perhaps [21:15:27 CEST] <fujisan> arent there any binary downloads? [21:15:52 CEST] <fujisan> im a newb at commandline stuff [21:16:03 CEST] <DelphiWorld> fujisan: you want to convert video? [21:17:00 CEST] <Chloe> DelphiWorld; some program they're using needs FFmpeg codecs, they just need to follow the guide for that program, it's not an FFmpeg issue. [21:18:01 CEST] <fujisan> http://pastebin.com/iqG15vav this is my output after i run the extracted exec on OS X [21:18:31 CEST] <DelphiWorld> Chloe: ok, got your point now [21:18:48 CEST] <fujisan> Delphiworld it's for nomachine to make the overall experience smoother [21:18:53 CEST] <Chloe> fujisan: contact the nomachine support [21:19:16 CEST] <fujisan> no i dont need to all i need is to get the ffmpeg codecs [21:19:26 CEST] <fujisan> im not a paying customer [21:19:32 CEST] <fujisan> it's free software [21:19:46 CEST] <DelphiWorld> Chloe: maybe not good question, but is a 720P stream taking 7% of the CPU a good value? [21:20:12 CEST] <DelphiWorld> 7% of a core only [21:20:23 CEST] <Chloe> fujisan: they have a forum, ask there. [21:20:30 CEST] <Chloe> DelphiWorld: it depends [21:20:40 CEST] <DelphiWorld> Chloe: depand on what? [21:20:53 CEST] <Chloe> on a lot of things [21:20:58 CEST] <DelphiWorld> source = a 1080P and out a 720P [21:21:39 CEST] <fujisan> i found the binaries thanks [21:26:02 CEST] <fujisan> libavutil libavcodec [21:26:18 CEST] <fujisan> i need these ffmpeg files i only get the exec [21:26:33 CEST] <DelphiWorld> ffmpeg -vaapi_device /dev/dri/renderD128 -hwaccel vaapi -hwaccel_output_format vaapi -i udp://@239.100.1.2:1234 -acodec libfdk_aac -ab 64k -ar 48k -vf 'format=nv12|vaapi,hwupload' -c:v h264_vaapi -vb 8000k -s 640x360 -aspect 16:9 -f mpegts udp://@239.100.100.2:1234 [21:26:34 CEST] <fujisan> and then they are listed where are they stored? [21:26:37 CEST] <DelphiWorld> any commant guys? [21:29:31 CEST] <fujisan> CHloe this is an ffmpeg question tho [21:29:38 CEST] <fujisan> i need ffmpeg files [21:29:43 CEST] <fujisan> thats all i need now [21:29:50 CEST] <DelphiWorld> you can compile it yourself [21:29:53 CEST] <fujisan> not help with how to set up nomachine [21:30:05 CEST] <DelphiWorld> ffmpeg build is static, mean that the libs are integrated in the binary [21:30:17 CEST] <fujisan> oh okay [21:30:24 CEST] <fujisan> thanks [21:30:41 CEST] <DelphiWorld> so you can just take the step and compile it yourself, or maybe install brew [21:33:40 CEST] <fujisan> okay ty and how would brew help me [21:33:44 CEST] <fujisan> i already have brew installed [21:34:14 CEST] <DelphiWorld> maybe brew search ffmpeg and install the required lib? [21:34:20 CEST] <DelphiWorld> i never used brew on a mac [21:34:26 CEST] <fujisan> oh okay ty [21:35:37 CEST] <Chloe> fujisan: probably just: brew install ffmpeg [21:37:12 CEST] <jpacker> Hello, i'm working on mixing two audio files together with ffmpeg and have that working correctly, but if I run the same command on the exact same files twice, it gives me this error: http://pastebin.com/AmFgSX4S [21:37:15 CEST] <jpacker> Any thoughts? [21:39:15 CEST] <furq> jpacker: pastebin the command line as well [21:40:18 CEST] <DelphiWorld> see here, the -s arg is ignored,and giving me the original stream size 704x576! [21:40:29 CEST] <DelphiWorld> ffmpeg -vaapi_device /dev/dri/renderD128 -hwaccel vaapi -hwaccel_output_format vaapi -i udp://@239.100.1.2:1234 -acodec libfdk_aac -profile:a aac_he_v2 -b:a 96k -ar 48k -vf 'format=nv12|vaapi,hwupload' -s 640x360 -r 29.97 -c:v h264_vaapi -vb 800k -aspect 16:9 -f flv rtmp://cdn.mybox.net/live/almajd03 [21:41:08 CEST] <jpacker> Hey furq here is the command i'm using a Java wrapper to run it [21:41:09 CEST] <jpacker> http://pastebin.com/GwiCiFis [21:42:05 CEST] <furq> "-ac -2" looks wrong [21:42:14 CEST] <DelphiWorld> furq: me ? [21:42:50 CEST] <furq> no? [21:42:59 CEST] <DelphiWorld> ok [21:43:01 CEST] <DelphiWorld> rebooting [21:50:08 CEST] <jpacker> Furq that just sped up my processing time but am still having that issue. [22:00:43 CEST] <DelphiWorld> so guys [22:00:57 CEST] <DelphiWorld> why my cmdline is ignoring my -s arg? [22:01:05 CEST] <jkqxz> DelphiWorld: "-s" there isn't going to do anything because the data is in GPU surfaces at that point, not in normal memory where a scale instance can see it (not sure why it doesn't fail, actually). Try adding ",scale_vaapi=640:360" to the end of your filer chain instead. [22:02:15 CEST] <DelphiWorld> jkqxz: -scale or ,scale? [22:03:23 CEST] <jkqxz> ",scale", as part of the filter argument ("hwupload,scale_vaapi") [22:04:27 CEST] <DelphiWorld> jkqxz: see it now? [22:04:27 CEST] <DelphiWorld> ffmpeg -vaapi_device /dev/dri/renderD128 -hwaccel vaapi -hwaccel_output_format vaapi -i udp://@239.100.1.2:1234 -acodec libfdk_aac -profile:a aac_he_v2 -b:a 64k -ar 48k -vf 'format=nv12|vaapi,hwupload,,scale_vaapi=640:360' -c:v h264_vaapi -vb 800k -f mpegts udp://@239.100.100.2:1234 [22:05:15 CEST] <DelphiWorld> not working... [22:05:23 CEST] <DelphiWorld> [AVFilterGraph @ 0x35131c0] No such filter: '' [22:05:33 CEST] <jkqxz> You put two commas... [22:05:45 CEST] <DelphiWorld> Lol... hahaha! [22:06:23 CEST] <DelphiWorld> jkqxz: this is crazy tts... misspelling the "cola" [22:07:18 CEST] <DelphiWorld> coma... crap [22:07:23 CEST] <DelphiWorld> jkqxz: ok, it works! [22:07:33 CEST] <DelphiWorld> now how to specify the aspect ratio? [22:07:42 CEST] <DelphiWorld> should the "aspect" be enough? [22:07:57 CEST] <DelphiWorld> and the fps? [22:11:36 CEST] <DelphiWorld> also why can't i specify a H.264 profile? [22:11:46 CEST] <DelphiWorld> ffmpeg -vaapi_device /dev/dri/renderD128 -hwaccel vaapi -hwaccel_output_format vaapi -i udp://@239.100.1.2:1234 -acodec libfdk_aac -profile:a aac_he_v2 -b:a 64k -ar 48k -vf 'format=nv12|vaapi,hwupload,scale_vaapi=640:360' -c:v h264_vaapi -vprofile baseline -level 3.1 -vb 800k -f mpegts udp://@239.100.100.2:1234 [22:12:05 CEST] <DelphiWorld> [h264_vaapi @ 0x2558320] [Eval @ 0x7ffcae5726b0] Undefined constant or missing ' [22:12:05 CEST] <DelphiWorld> (' in 'baseline' [22:12:13 CEST] <DelphiWorld> [h264_vaapi @ 0x2558320] Unable to parse option value "baseline" [22:12:22 CEST] <DelphiWorld> [h264_vaapi @ 0x2558320] Error setting option profile to value baseline. [22:13:18 CEST] <jkqxz> -profile:v 66 [22:13:34 CEST] <DelphiWorld> LOL... what's 66? [22:13:47 CEST] <DelphiWorld> jkqxz: now another issue... lot of hw encoding issue [22:13:53 CEST] <DelphiWorld> ffmpeg -vaapi_device /dev/dri/renderD128 -hwaccel vaapi -hwaccel_output_format vaapi -i udp://@239.100.1.2:1234 -acodec libfdk_aac -profile:a aac_he_v2 -b:a 64k -ar 48k -vf 'format=nv12|vaapi,hwupload,scale_vaapi=640:360' -c:v h264_vaapi -vb 800k -f flv rtmp://cdn.mybox.net/live/almajd03 [22:13:56 CEST] <DelphiWorld> see the result here: [22:14:08 CEST] <DelphiWorld> Input stream #0:0 frame changed from size:704x576 fmt:yuv420p to size:704x576 fmt:vaapi_vld [22:14:25 CEST] <DelphiWorld> and it stop [22:17:09 CEST] <DelphiWorld> -profile:v 66 fail [22:17:11 CEST] <DelphiWorld> not found [22:18:10 CEST] <jkqxz> Why do you want baseline profile anyway? Nothing actually supports it. (You probably meant constrained baseline, which is -profile:v 576.) [22:18:21 CEST] <DelphiWorld> if i send to mcast mpegts udp it works, but if i send to rtmp i get that error. [22:18:49 CEST] <DelphiWorld> jkqxz: why do you think the flv container is refusing my stream? [22:19:51 CEST] <DelphiWorld> ffmpeg -vaapi_device /dev/dri/renderD128 -hwaccel vaapi -hwaccel_output_format vaapi -i udp://@239.100.1.2:1234 -acodec libfdk_aac -profile:a aac_he_v2 -b:a 64k -ar 48k -vf 'format=nv12|vaapi,hwupload,scale_vaapi=640:360' -c:v h264_vaapi -vb 800k -f flv rtmp://mybox.net/live/almajd03 [22:21:22 CEST] <DelphiWorld> Input stream #0:0 frame changed from size:704x576 fmt:yuv420p to size:704x576 fmt:vaapi_vld [22:22:04 CEST] <furq> he really needs a better computer [22:26:00 CEST] <DelphiWorld> back [22:26:08 CEST] <DelphiWorld> jkqxz: did you got any point on that flv issue? [22:27:14 CEST] <furq> DelphiWorld: -hwaccel_output_format yuv420p [22:27:28 CEST] <furq> or scale_vaapi=1280:720:format=yuv420p [22:27:50 CEST] <DelphiWorld> furq: where to put that format bro [22:28:14 CEST] <furq> -hwaccel_output_format goes before -i [22:28:17 CEST] <furq> the other one should be obvious [22:31:37 CEST] <DelphiWorld> jkqxz: even 576 profile not found [22:33:03 CEST] <DelphiWorld> furq: result: [22:33:08 CEST] <DelphiWorld> ffmpeg -vaapi_device /dev/dri/renderD128 -hwaccel vaapi -hwaccel_output_format vaapi -i udp://@239.100.1.2:1234 -acodec libfdk_aac -profile:a aac_he_v2 -b:a 64k -ar 48k -vf 'format=nv12|vaapi,hwupload,scale_vaapi=640:360:format=yuv420p' -c:v h264_vaapi -vb 800k -f flv rtmp://cdn.mybox.net/live/almajd03 [22:33:34 CEST] <DelphiWorld> ffmpeg: i965_gpe_utils.c:345: i965_gpe_set_surface2_state: Assertion `obj_surface->fourcc == 0x3231564E' failed. Aborted (core dumped) [22:33:34 CEST] <furq> that looks right [22:33:41 CEST] <furq> nice [22:33:50 CEST] <DelphiWorld> but core dump [22:33:51 CEST] <furq> i copied that off the libav wiki so i'm not sure how much it applies to ffmpeg [22:34:04 CEST] <furq> try with just one or the other [22:34:10 CEST] <DelphiWorld> i copied everything from ffmpeg [22:34:25 CEST] <DelphiWorld> i might try the hw one [22:34:26 CEST] <DelphiWorld> hold on [22:34:31 CEST] <DelphiWorld> send me the second please? [22:34:48 CEST] <jpacker> I just realized I had my call wrong furq, do you see anything wrong with this one? http://pastebin.com/h42bC8G2 [22:36:41 CEST] <DelphiWorld> furq: send me back the hwaccell one? or give me the page? [22:36:47 CEST] <DelphiWorld> pretty less doc [22:37:07 CEST] <furq> https://wiki.libav.org/Hardware/vaapi#Surface_Formats [22:37:18 CEST] <furq> the second code box [22:37:29 CEST] <DelphiWorld> ok [22:40:58 CEST] <DelphiWorld> furq: ok got it [22:41:24 CEST] <DelphiWorld> ffmpeg -vaapi_device /dev/dri/renderD128 -hwaccel vaapi -hwaccel_output_format yuv420p -i udp://@239.100.1.2:1234 -acodec libfdk_aac -profile:a aac_he_v2 -b:a 64k -ar 48k -vf 'format=nv12|vaapi,hwupload,scale_vaapi=640:360' -c:v h264_vaapi -vb 800k -f flv rtmp://mybox.net/live/almajd03 [22:41:30 CEST] <DelphiWorld> but publishing is not working [22:42:23 CEST] <DelphiWorld> all i get at the end is Press [q] to stop, [?] for help [22:42:31 CEST] <DelphiWorld> and it shutdown [22:46:18 CEST] <furq> pastebin the full output i guess [22:46:34 CEST] <DelphiWorld> furq: no problem. how to capture the entire output please [22:47:14 CEST] <furq> -report [22:47:29 CEST] <DelphiWorld> -report what will do? [22:47:39 CEST] <furq> -report [22:47:40 CEST] <furq> Dump full command line and console output to a file named program-YYYYMMDD-HHMMSS.log in the current directory. [22:47:54 CEST] <DelphiWorld> haha, good idea [22:47:54 CEST] <DelphiWorld> hold on [22:49:22 CEST] <DelphiWorld> furq: http://paste.ubuntu.com/20358926/ [22:54:21 CEST] <furq> looks like it's still outputting vaapi_vld and not yuv420p [22:54:42 CEST] <DelphiWorld> so what must we do... [22:55:05 CEST] <DelphiWorld> ffmpeg -v verbose -vaapi_device /dev/dri/renderD128 -hwaccel vaapi -hwaccel_output_format yuv420p -i h.mp4 -acodec libfdk_aac -profile:a aac_he_v2 -b:a 64k -ar 48k -vf 'format=nv12|vaapi,hwupload,scale_vaapi=640:360' -c:v h264_vaapi -vb 800k -f mp4 out.mp4 [22:55:13 CEST] <DelphiWorld> i'm trying this local files, let's see the change [22:56:01 CEST] <DelphiWorld> using that cmdline i can see the outputed format in mp4 is yuv 420p [22:59:00 CEST] <DelphiWorld> duno what to do... [22:59:56 CEST] <furq> it sure would be nice if this was documented somewhere [23:00:06 CEST] <DelphiWorld> yep [23:00:12 CEST] <DelphiWorld> but when i get it i'll write a doc [23:04:56 CEST] <DelphiWorld> back [23:06:17 CEST] <DelphiWorld> furq: fyi, if i output to a mpeg2ts i wont get that issue [23:07:22 CEST] <jkqxz> The way a hwaccel works is that it first initialises as if using the software decoder (so yuv420p output), then initialises the hwaccel which changes to its own output format. That gives you the "Input stream ... changed ..." message, though really it didn't actually change because there were no frames of the first form. [23:08:37 CEST] <jkqxz> And is also the reason for the strange 'format=nv12|vaapi,hwupload' setup, because the filter graph gets that same initialisation. [23:08:47 CEST] <furq> http://vpaste.net/6FLYC [23:08:52 CEST] <furq> that's the bit i assume is to blame [23:09:04 CEST] <furq> but if it works with other output formats then i have no idea [23:10:13 CEST] <DelphiWorld> furq, jkqxz: so what's your suggestions guys [23:10:32 CEST] <DelphiWorld> i am fetching from mcast, encoding then pushing to rtmp but no *LUCK* [00:00:00 CEST] --- Fri Jul 22 2016

1 0