[Ffmpeg-devel-irc] ffmpeg.log.20171112

[burek]

[00:00:01 CET] <blap> maybe it's mpv's fault
[00:13:03 CET] <durandal_1707> blap: do you in any way set volume by more than 100 in mpv?
[00:41:27 CET] <blap> no durandal_1707
[00:41:41 CET] <blap> what's wrong with this? ffplay -f lavfi dynaudnorm=g=31:f=500:r=0.8:p=0.90:m=2 file.mp3
[00:43:10 CET] <blap> hah! you got cellular automata in there! :D
[00:52:08 CET] <blap> https://ffmpeg.org/ffmpeg-all.html#dynaudnorm  doesn't tell me how to invoke
[00:53:11 CET] <blap> ah got it  ffplay -af  dynaudnorm=g=31:f=500:r=0.8:p=0.90:m=2 -i test.mp3
[01:00:44 CET] <thebombzen> ffmpeg fails to build with libva 2.0
[01:00:53 CET] <thebombzen> is there any fix in the meantime, other than "downgrade to 1.8"
[01:03:01 CET] <JEEB> I saw there was a trac issue on that if I recall correctly but no patches on the ML as far as I can see
[01:03:23 CET] <durandal_1707> isnt libva bad?
[01:03:40 CET] <JEEB> isn't it the va-api implementation?
[01:04:03 CET] <JEEB> I mean, it's probably the least bad multi-vendor HW dec/enc API, no?
[01:04:10 CET] <JEEB> which isn't a high border, I know. but yunno
[01:04:39 CET] <durandal_1707> i just remember one lib being crap...
[01:04:55 CET] <JEEB> a lot of libs are bad
[01:04:56 CET] <JEEB> :D
[01:05:02 CET] <furq> are you thinking of libmfx
[01:05:31 CET] <thebombzen> vdpau was the one that was crap
[01:05:37 CET] <durandal_1707> ah
[01:05:51 CET] <JEEB> vdpau was the least bad decoding API at one point
[01:05:55 CET] <JEEB> then it lost the maintainer IIRC
[01:06:11 CET] <JEEB> now nvidia is saying that they're "totally adding 10bit support to VDPAU and we totally have a maintainer again"
[01:07:02 CET] <thebombzen> hm, that's interesting
[01:07:12 CET] <thebombzen> so libva 2.0 works with ffmpeg-3.4-4 in arch but not ffmpeg-3.4-1
[01:07:24 CET] <thebombzen> did the Arch maintainer do a really quick-n-dirty patch to make it work?
[01:07:30 CET] <JEEB> no idea
[01:08:02 CET] <thebombzen> in the meantime I'm downgrading but I don't like this solution :(
[01:08:04 CET] <JEEB> I just don't see a "make it work on 2.0" patch on the ML (I have no idea how hard/easy it would be)
[01:08:56 CET] <JEEB> huh, the ticket says that 3.4 supposedly works (?!)
[01:08:57 CET] <JEEB> https://trac.ffmpeg.org/ticket/6828
[01:09:23 CET] <JEEB> unless that arch user is getting the packaged one
[01:09:36 CET] <JEEB> which might just like you say work with patches
[01:11:47 CET] <JEEB> post on that ticket if you can get any findings about things I guess
[01:11:56 CET] <JEEB> regarding patching etc
[01:18:58 CET] <Johnjay> i didn't know about nvenc and nvdec until today
[01:19:07 CET] <Johnjay> it says they will continue to support vdpau though
[01:22:35 CET] <SortaCore> what's the difference between h264_d3d11va and h264_d3d11va2?
[01:22:47 CET] <SortaCore> is the first for backward compatibility
[02:46:53 CET] <blap> thanks for help durandal_1707 - ffplay has no distortion with default dynaudio settings
[02:54:37 CET] <TikityTik> I'm trying to concat 3 flv files that are h264 and encode them to webm, but there's a pause between each video, how do i get rid of this pause?
[03:33:23 CET] <kingsley> TikityTik: Humble suggestion. Consider a video editor. Like kdenlive.
[04:02:36 CET] <kingsley> Where should I look for documentation on parameters and syntax for using frei0r effects as sources?
[04:02:48 CET] <kingsley> https://frei0r.dyne.org/ says ... "Filters: 85   -   Generators: 7   -   Mixers: 33"
[04:02:59 CET] <kingsley> but https://ffmpeg.org/ffmpeg-all.html only mentions one named "partik0l".
[06:48:23 CET] <ZeroWalker> okay mpv was able to playback my tcp stream, so i guess libmpv should be able to do the same on android. I can't really find it though, not familiar with android, but i am guessing the mpv is Not the libmpv and only the actual player. I want the player to be inbuilt so to speak
[11:35:25 CET] <ZeroWalker> when you stream rtp, how do you make it so the sdp is also available in the stream thingy, like "rtp://xx:xx/live.sdp"
[14:01:08 CET] <redrabbit> im looking for a tool to analyse a file and get a graph of bitrate over time
[14:01:57 CET] <redrabbit> which shows bitrate distribution
[14:31:02 CET] <DHE> I ended up making my own based on ffprobe -show_frames
[15:53:30 CET] <ZeroWalker> when you stream rtp, how do you make it so the sdp is also available in the stream thingy, like "rtp://xx:xx/live.sdp"
[16:46:25 CET] <JEEB> thebombzen: libva 2.0 breakage seems to have gotten suddenly fixed ;)
[16:57:07 CET] <kepstin> ZeroWalker: rtp doesn't transfer the sdp, you need to use some other protocol for that. Either look at something like rtsp or just transfer the sdp in some other way.
[17:10:26 CET] <ZeroWalker> was a bit confused hot to get rtsp up and running with ffmpeg, something about the protocol not being there
[17:11:01 CET] <ZeroWalker> i actually prefer tcp for my objective, but it seems android supports normal streaming protocols natively, so thought i would had a go at that
[19:02:31 CET] <redrabbit> so is there anything like that available ?
[19:03:39 CET] <redrabbit> does own solution is ready for other people ?
[19:03:43 CET] <redrabbit> your*
[19:07:51 CET] <jaggz> if I have a hundred or so audio and video files, foo.m4a foo.m4v ... bar.m4a bar.m4v, how do I assemble those into one audio+video file?
[19:08:11 CET] <jaggz> they're not numbered, but I have the list of them in order.
[19:08:51 CET] <jaggz> the audio and video files with the same name correspond to each other, so foo.m4a is at the same time as foo.m4v
[19:09:22 CET] <klaxa> write a script that reads them into variables and then for each pair: ffmpeg -i $foo -i $bar -c copy foobar.mp4
[19:09:38 CET] <klaxa> where $foo is foo.m4v and $bar is foo.m4a for example
[19:09:51 CET] <jaggz> my commandline will be immensely long
[19:10:16 CET] <klaxa> this is for each pair
[19:10:21 CET] <klaxa> rather short imo
[19:10:29 CET] <jaggz> can I do it at once, or do I need to make 50 (100/2) "%03d".mp4's then join those?
[19:10:36 CET] <klaxa> oh
[19:11:24 CET] <klaxa> hm... if you have them in separate lists, maybe try: ffmpeg -f concat -i videos.txt -f concat -i audios.txt -c copy foobar.mp4
[19:11:29 CET] <jaggz> for instance, if I put them as foo.m4a foo.m4v bar.m4a bar.m4v into mylist.txt
[19:11:43 CET] <jaggz> okay, will try that :)
[19:12:02 CET] <klaxa> you need them to be in the file as: file "foo.m4v" though
[19:12:04 CET] <klaxa> see https://trac.ffmpeg.org/wiki/Concatenate
[19:13:34 CET] <jaggz> eenteresting
[19:17:28 CET] <Trieste> hi, is there a way to merge a "static" video loop with a capture device for the purposes of streaming? I'm using this command: https://gist.github.com/anonymous/69c67e264fb29e2aee85fc47f993e160 - but it seems the video stream drops out as soon as the video ends
[19:26:42 CET] <jaggz> eenteresting
[19:53:25 CET] <jaggz> klaxa, shoot.. the m4v and m4a files are not a recognizable format
[19:53:34 CET] <_Vi> How do I make FFmpge print reduced info like with `-v warning`, but still with progress indication like "frame=  117 fps= 87 q=-1.0"?
[19:54:18 CET] <_Vi> Found myself: `-stats`
[19:54:29 CET] <another> _Vi: -hide_banner ?
[19:58:01 CET] <_Vi> another, Tried. It still prints stream info and mapping.
[19:58:14 CET] <_Vi> another, `-v warning -stats` prints just what I need.
[20:31:46 CET] <_Vi> How do I connect a source to multiple sinks? Failing command line: `ffmpeg -filter_complex "anullsrc[s];[s]anullsink;[s]anullsink"`. Is three a filter that copies a stream?
[20:34:26 CET] <_Vi> Found: `split`, `asplit`. Initially searched the docs for "copy", "duplicate" and "tee".
[22:05:06 CET] <ray__> hi everyone
[22:05:31 CET] <ray__> im trying to compile ffmpeg acordding to the guide
[22:05:36 CET] <ray__> but not success
[22:05:46 CET] <ray__> it gives me an error ERROR: x265 not found using pkg-config
[22:07:26 CET] <ray__> im using ubuntu 16.04
[22:11:36 CET] <sfan5> you're missing the x265 dev package
[22:14:21 CET] <RayZX> but the compilation of x265 proceeded without errors
[22:15:19 CET] <sfan5> what does ffbuild/config.log say?
[22:15:31 CET] <RayZX> https://pastebin.com/k6PDpaHN
[22:15:36 CET] <JEEB> most likely he built static and the C++ standard library lib is missing
[22:15:47 CET] <JEEB> --extra-libs="-lstdc++" is most likely fix
[22:15:55 CET] <JEEB> because the pc file doesn't contain it in Libs.private
[22:16:14 CET] <JEEB> oh, that wasn't it
[22:16:25 CET] <JEEB> the header is broken?
[22:16:46 CET] <JEEB> or they made the header finally C++
[22:16:54 CET] <JEEB> in which case gg
[22:17:12 CET] <RayZX> https://bitbucket.org/multicoreware/x265/issues/371/x265-not-found-using-pkg-config
[22:17:22 CET] <JEEB> that was from before
[22:17:28 CET] <JEEB> that has nothing to do with that
[22:17:42 CET] <JEEB> oh, yea. it has a similar thing afterwards
[22:17:52 CET] <JEEB> https://bitbucket.org/multicoreware/x265/issues/377/x265h-is-valid-c-but-not-a-valid-c-code
[22:17:56 CET] <JEEB> yea
[22:17:57 CET] <JEEB> "valid C++ but not C"
[22:18:13 CET] <JEEB> so yes, pick an older x265 version :P
[22:18:17 CET] <JEEB> they broke their shit
[22:18:19 CET] <sfan5> looks like they accidentally dropped C support
[22:18:22 CET] <JEEB> yes
[22:18:25 CET] <RayZX> lol
[22:43:14 CET] <RayZX> worked with x265 2.4, thanks
[23:51:27 CET] <kingsley> If you happen to know what to do when the mandelbrot source logs errors like...
[23:51:33 CET] <kingsley> [Parsed_mandelbrot_0 @ 0x814f6760] Mandelbrot cache is too small!te=16108.4kbits/s speed=0.0109x
[23:51:45 CET] <kingsley> feel free to enlighten me.
[23:53:49 CET] <kingsley> So far, all I found is Michael Niedermayer's commit at github.
[23:54:40 CET] <kingsley> michaelni: What do you think about the above?
[23:56:00 CET] <durandal_1707> kingsley: for how long duration you run it?
[23:56:29 CET] <kingsley> durandal_1707: Hours.
[23:56:40 CET] <durandal_1707> why?
[23:57:05 CET] <kingsley> Oops.
[23:57:25 CET] <kingsley> When you asked how long duration I run it, did you mean
[23:57:37 CET] <kingsley> a.) wall clock time, or
[23:57:47 CET] <kingsley> b.) length of video to render?
[23:59:16 CET] <durandal_1707> b)
[00:00:00 CET] --- Mon Nov 13 2017