[Ffmpeg-devel-irc] ffmpeg.log.20191012

[burek]

[00:00:31 CEST] <durandal_1707> -f rawvideo tells to write rawvideo
[00:00:39 CEST] <durandal_1707> overiding .y4m extension
[00:00:57 CEST] <durandal_1707> .y4m is yuv4mpeg format
[00:01:41 CEST] <void09> so -f yuv4mpeg?
[00:01:43 CEST] <durandal_1707> you see, ffmpeg can guess output format from extension, so you do not need to type -f matroska
[00:01:53 CEST] <durandal_1707> void09: not needed, just use extension
[00:02:10 CEST] <durandal_1707> and instead just type .mkv
[00:02:50 CEST] <void09> if i do -f yuv4mpeg I get [NULL @ 0x562163ad6940] Requested output format 'yuv4mpeg' is not a suitable output format
[00:03:07 CEST] <durandal_1707> yuv4mpegpipe
[00:03:29 CEST] <durandal_1707> ffmpeg -muxers will list all muxers
[00:03:40 CEST] <durandal_1707> ffmpeg -h muxer=yuv4mpegpipe
[00:03:48 CEST] <durandal_1707> will list help for that muxer
[00:04:17 CEST] <void09> ffmpeg -i "input.ts" -an -c:v rawvideo -f yuv4mpegpipe -vframes 1000 tmp.y4m
[00:04:31 CEST] <void09> [yuv4mpegpipe @ 0x55b22f12a940] ERROR: Codec not supported.
[00:04:32 CEST] <durandal_1707> -c:v rawvideo is not needed for y4m
[00:04:56 CEST] <void09> ah, ok. it's confusing. isn't it the same raw video as with yuv ?
[00:05:03 CEST] <durandal_1707> you try multiple times to outsmart ffmpeg, but fails miserably
[00:05:04 CEST] <Zipdox> What is the syntax for -x265-params?
[00:05:35 CEST] <durandal_1707> void09: it uses wrapped avframe codec instead for performance reasons
[00:05:55 CEST] <durandal_1707> Zipdox: consult documentation
[00:06:03 CEST] <Zipdox> I couldn't find it
[00:06:05 CEST] <void09> oh, so it's a different kind of raw video
[00:23:50 CEST] <Zipdox> Ok I've tried lots of things and I couldn't find any documentation on how to use -x265-params
[00:32:14 CEST] <Zipdox> hello?
[00:34:26 CEST] <void09> Zipdox: https://trac.ffmpeg.org/wiki/Encode/H.265
[00:34:33 CEST] <void09>  Generally, options are passed to x265 with the -x265-params argument. For fine-tuning the encoding process, you can therefore pass any option that is listed in the x265 documentation. Keep in mind that fine-tuning any of the options is generally not necessary, unless you absolutely know what you need to change.
[00:34:52 CEST] <pink_mist> the manpage has examples https://ffmpeg.org/ffmpeg-all.html#Options-29
[00:40:53 CEST] <Zipdox> I found a youtube video with some helpfull command
[00:42:51 CEST] <Zipdox> turns out the x265-params arg isn't useful, I need to specify the values through ffmpeg directly
[01:13:07 CEST] <kepstin> x265-params is useful if you're trying to set a parameter on the x265 encoder that ffmpeg doesn't have a option mapping for (there's a bunch of them)
[01:30:19 CEST] <void09> any way to detect errors in an mkv/h264 file that cause visible visual glitches ? My tv recording software sometimes reports a various number of errors, but I have yet to detect any visual problems with the files.
[02:47:13 CEST] <oneforall2> how do i find out how long a file mkv is ? want to cut off the last 2sec
[02:48:06 CEST] <oneforall2> -ss 00:00:00.00 -to
[02:48:21 CEST] <furq> -sseof 2
[02:48:35 CEST] <furq> oh wait you want to remove the last two seconds
[02:48:40 CEST] <furq> ffprobe then
[02:48:40 CEST] <oneforall2> yeah
[02:48:57 CEST] <oneforall2> ah thanks will try that
[02:50:09 CEST] <oneforall2> hhm can I put this with -ss  0.000000           start: 0.000000  ?
[02:50:55 CEST] <oneforall2> -ss 00:00:00.0000 I see more of this :)
[02:51:06 CEST] <furq> just omit -ss
[02:51:34 CEST] <furq> you could use 0 or 00:00 or 00:00:00 or 0.0 or whatever
[02:51:49 CEST] <furq> but don't actually do that in this case
[02:53:17 CEST] <oneforall2> ah thanks
[02:58:42 CEST] <oneforall2> looks like it worked ..
[02:58:49 CEST] <oneforall2> :)
[06:42:07 CEST] <kingsley> In the command...
[06:42:23 CEST] <kingsley> $ ffmpeg  -f lavfi -i mandelbrot=s=1920x1080 -frames 50 /tmp/mandel.webm
[06:42:33 CEST] <kingsley> what is "-f lavfi" for?
[07:17:43 CEST] <furq> kingsley: input format libavfilter
[07:17:54 CEST] <furq> mandelbrot is the name of a libavfilter video source
[07:17:58 CEST] <furq> https://ffmpeg.org/ffmpeg-filters.html#mandelbrot
[07:32:48 CEST] <kingsley> furq: Thank you.
[09:19:06 CEST] <ok1mgj> Hi all, i have trouble with transcoding multicast. input and outputs are mpegts streams, all is working fine until input stream changes SAR/DAR  16:9 to 4:3 or back.
[09:19:51 CEST] <ok1mgj> error is: Impossible to convert between the formats supported by the filter 'Parsed_scale_npp_0' and the filter 'auto_scaler_0'
[09:20:17 CEST] <ok1mgj> any idea how to survive this SAR/DAR change?
[09:25:17 CEST] <ok1mgj> ..and command is: ffmpeg -re -nostats -hwaccel cuvid -c:v mpeg2_cuvid -deint 2 -drop_second_field 1 -i udp://239.250.1.2:1234?overrun_nonfatal=1"&"fifo_size=57344"&"buffer_size=256000 -map 0:0 -map 0:1 -rc-lookahead 32 -profile:v main -c:v h264_nvenc -qmin 0 -qmax 50 -c:a libfdk_aac -b:a 128k -b:v 1900k -minrate:v 900k -maxrate:v 2900k -g 50
[09:25:18 CEST] <ok1mgj> -bufsize 6M -preset fast -vf scale_npp=h=ih:w=sar*iw -f mpegts udp://239.253.99.99:1234
[18:28:36 CEST] <Celmor[m]> I have a video I downloaded from Youtube via youtube-dl that neither ffmpeg nor mpv can correctly process, mpv for example just sticks around, dumping output about corrupted frames while consuming 100% cpu. redownloading the video didn't help. Meanwhile vlc plays it just fine...
[18:32:26 CEST] <Celmor[m]> complete output of youtube-dl, vlc, mpv and ffmpeg: https://pastebin.com/raw/spM3yayK
[18:33:15 CEST] <JEEB> yea that sounds like the bug with libaom's release version
[18:33:24 CEST] <JEEB> which is utilized for AV1 decoding in that case with FFmpeg
[18:33:32 CEST] <JEEB> while I bet that vlc build uses dav1d
[18:34:47 CEST] <Celmor[m]> so what'd be the fix? using a different decoder?
[18:34:49 CEST] <JEEB> current FFmpeg master has defaulted to dav1d as default since seemingly Google didn't release a new release of libaom
[18:35:01 CEST] <JEEB> and in general dav1d is faster
[18:35:20 CEST] <JEEB> but it's funny how google's decoder fails with google's streams because even chrome currently utilizes dav1d :P
[18:35:45 CEST] <Celmor[m]> so I would have to recompile ffmpeg and mpv from master?
[18:36:56 CEST] <another> or update your libaom
[18:37:10 CEST] <JEEB> that way the default gets switched yes. otherwise is your FFmpeg already has dav1d built in then --vd=libdav1d does fix it
[18:37:12 CEST] <Celmor[m]> Other AVC streams seem to work just fine so I wonder what's special about that video
[18:37:17 CEST] <JEEB> it's not AVC
[18:37:20 CEST] <JEEB> it's AV1
[18:38:00 CEST] <Celmor[m]> ah, same crap as vp9 which requires software decoding. got it
[18:38:13 CEST] <Celmor[m]> so I need to force AVC when downloading
[18:38:55 CEST] <another> if you explicitly want avc then you can download that, yes
[18:39:40 CEST] <Celmor[m]> Unrecognized option '-vd=libdav1d'.
[18:40:02 CEST] <Celmor[m]> ( did use double-dash )
[18:40:19 CEST] <JEEB> you can check the output of -vd=help
[18:40:29 CEST] <JEEB> if there's no libdav1d there then your FFmpeg was built without it
[18:40:34 CEST] <JEEB> (the one linked against your mpv)
[18:41:17 CEST] <Celmor[m]> ffmpeg -vd=help: Unrecognized option 'vd=help'
[18:41:50 CEST] <Celmor[m]> I guess you meant mpv?
[18:41:59 CEST] <pink_mist> yes, that's mpv arguments
[18:42:21 CEST] <Celmor[m]> I recall av1 never working before on my machine either
[18:43:33 CEST] <Celmor[m]> would this use ffmpeg with the other decoder? https://aur.archlinux.org/packages/mpv-build-git/
[18:43:47 CEST] <Celmor[m]> since it compiles ffmpeg into mpv from what I understnad
[18:44:34 CEST] <pink_mist> you should probably ask in #mpv instead
[18:44:37 CEST] <pink_mist> this is #ffmpeg
[18:44:51 CEST] <pink_mist> (note: I don't know the answer)
[18:48:10 CEST] <another> Celmor[m]: if you're using arch you just need to recompile ffmpeg
[18:48:21 CEST] <another> mpv uses libdav1d via ffmpeg
[18:49:15 CEST] <another> actually, it looks like ffmpeg is compiled w/ dav1d by default
[18:49:22 CEST] <another> *in arch
[18:49:33 CEST] <lordarkmemo> Hi nice to meet you all
[18:50:24 CEST] <lordarkmemo> im streaming a video using -f flv rtmp://server.  But, when the rtmp server (nginx) stop, the ffmpeg hangs forever
[18:50:49 CEST] <another> Celmor[m]: just put 'vd=libdav1d' into your ~/.config/mpv/config
[18:51:18 CEST] <cards> How can I get ffprobe -print_format to work?  It does not.
[18:51:23 CEST] <cards> > ffprobe -print_format csv file.mpg
[18:52:19 CEST] <JEEB> cards: you need to use show_xxx to get something in there
[18:52:30 CEST] <JEEB> like -show_format -show_streams -show_programs etc
[18:52:40 CEST] <JEEB> so you can pick exactly what you need
[18:52:45 CEST] <cards> JEEB: what is the default output so i can emulate those options
[18:52:53 CEST] <cards> default flags
[18:53:29 CEST] <cards> seems like -print_format should work on the default flags
[18:53:41 CEST] <furq> the default is to not print anything
[18:53:55 CEST] <cards> the default is certainly to print something, since something does print
[18:54:05 CEST] <furq> it prints the normal ffmpeg output on stderr
[18:54:10 CEST] <furq> -of only affects the stuff printed on stdout
[18:55:06 CEST] <cards> it prints Metadata: major_brand minor_version compatible_brands creation_time title encoder comment  Duration: Stream #: Metadata: creation_time handler_name etc
[18:55:23 CEST] <cards> which flags then explicite that print?
[18:55:47 CEST] <furq> i don't think any ffprobe flags affect that output
[18:55:48 CEST] <pragomer> hi. I am under Debian but relativly new to ffmpeg. My actual goal is to "visualize" mp3 to mkv via showwaves. My actual line looks like that:
[18:55:53 CEST] <furq> that's the same stuff you'd get from ffmpeg -i
[18:56:02 CEST] <cards> furq: how do I get that output into csv
[18:56:02 CEST] <pragomer> ffmpeg -i input -filter_complex "[0:a]showwaves=s=1280x720:mode=line,format=yuv420p[v]" -map "[v]" -map 0:a -c:v libx264 -c:a copy output.mkv
[18:56:11 CEST] <furq> you don't, but you probably want -show_format -of csv
[18:56:22 CEST] <furq> and/or -show_streams
[18:56:23 CEST] <cards> i will get all the same values into csv?
[18:56:27 CEST] <pragomer> how could I manipulate the colors and/or forms of the waveforms? is there a list or so?
[18:56:29 CEST] <furq> you'll get way more values into csv
[18:56:37 CEST] <cards> but i just want those.
[18:56:55 CEST] <cards> i'm confused why the default values are pegged as YAML and cannot be printed to CSV
[18:58:06 CEST] <cards> I'm actually looking at a YAML2CSV converter to get around this
[18:58:14 CEST] <pmjdebruijn> does anybody here have experience with the lenscorrection filter? it seems to produce extremely poor quality for me
[18:58:21 CEST] <furq> well that's not yaml so you probably won't have much luck with that
[18:58:28 CEST] <furq> or if it is valid yaml then it's not on purpose
[18:58:29 CEST] <cards> it is YAML
[18:58:32 CEST] <furq> so i wouldn't depend on it
[18:59:50 CEST] <cards> I'll try the outputs you suggested.  which of those only reads metadata, and which of those takes excess time to analize?
[18:59:54 CEST] <durandal_1707> pmjdebruijn: what you want to do?
[18:59:55 CEST] <furq> anyway you want some combination of -select_streams and -show_entries
[19:00:10 CEST] <pmjdebruijn> durandal_1707: just correct some lens distortion
[19:00:16 CEST] <amosbird> hi, how does ffmpeg generate this surrounding rectangle when recording? https://la.wentropy.com/UYZl.png  I'd like to have that rectangle
[19:00:24 CEST] <furq> or not -select_streams if you only want container-level metadata
[19:00:43 CEST] <furq> and the only option that decodes the entire stream is -count_frames iirc
[19:00:44 CEST] <pmjdebruijn> however, on straight lines, the lenscorrection filter does not interpolate them correctly, it produces jaggies
[19:00:54 CEST] <cards> thanks
[19:01:04 CEST] <furq> otherwise it'll always just analyse whatever you set in -analyzeduration/-probesize which is five seconds and 5MB
[19:01:09 CEST] <furq> by default
[19:01:12 CEST] <pmjdebruijn> although perhaps I just try the latest version of ffmpeg
[19:01:35 CEST] <durandal_1707> pmjdebruijn: you give custom paramters to filter?
[19:01:47 CEST] <furq> pragomer: https://ffmpeg.org/ffmpeg-filters.html#showwaves
[19:01:51 CEST] <cards> furq: what is the default output of ffprobe so that I can reference it by name
[19:02:06 CEST] <pmjdebruijn> durandal_1707: of course, otherwise lenscorrection is useless
[19:02:17 CEST] <furq> cards: https://www.ffmpeg.org/ffprobe.html#default
[19:02:39 CEST] <furq> if you mean the stuff printed on stderr then i don't think that has a name
[19:02:41 CEST] <pmjdebruijn> I can try to create a nice compact reproducable case if needed
[19:02:55 CEST] <cards> I mean does it have a name people refer to it by.  container metadata/duration information?
[19:03:19 CEST] <lordarkmemo> im streaming a video using -f flv rtmp://server.  But, when the rtmp server (nginx) stop, the ffmpeg hangs forever
[19:03:40 CEST] <furq> i've not seen a name for that
[19:03:55 CEST] <kepstin> amosbird: that's the x11grab input? The input code in ffmpeg just makes an x11 window that's rectangle shaped and draws some stripes in it. You're welcome to copy the code if you follow the license terms...
[19:03:57 CEST] <furq> it would probably be nice to have one better than "the input stuff printed on stderr"
[19:04:14 CEST] <cards> so that I can put a feature request to get it supported by -print_format
[19:04:36 CEST] <durandal_1707> pmjdebruijn: try lensfun filter
[19:06:05 CEST] <pmjdebruijn> yeah lensfun has good interpolation
[19:06:32 CEST] <pmjdebruijn> but lenscorrection is sufficiently poor, I'm wondering why its not clearly marked as deprecated
[19:07:06 CEST] <kepstin> cards: the output you get from ffprobe on stderr (without any -print_format)  isn't even generated by ffprobe itself, it's some log output from the ffmpeg libraries.
[19:07:07 CEST] <pmjdebruijn> oh... and my system ffmpeg doesn't have lensfun support yet
[19:07:41 CEST] <pmjdebruijn> oh and the ffmpeg downlodable static binaries don't either?
[19:08:19 CEST] <kepstin> iirc lensfun has a bunch of data files of lens presets, so bundling it in a static binary doesn't make much sense
[19:11:43 CEST] <pmjdebruijn> kepstin: the library should be able to do a transform without those presets, if given manual parameters
[19:12:32 CEST] <cards> kepstin: it's still useful handy information.  does ffmpeg have the output flags to request this information in csv then?
[19:12:39 CEST] <amosbird> kepstin: so that's a transparent window then
[19:12:49 CEST] <pmjdebruijn> and I don't particulary care about optically perfect correction, I just want to get rid of the bulk of the distortion, the interpolation issue caused by the lenscorrection filter are however worse than the distortion to begin with
[19:13:01 CEST] <durandal_1707> pmjdebruijn: perhaps lenscorrection use no interpolation but nearest neighbor?
[19:13:07 CEST] <kepstin> cards: all of the information from that log output is included in the other ffprobe output formats (you might need -show_format and -show_streams)
[19:13:09 CEST] <amosbird> kepstin: and that window cannot be covered, always on top
[19:13:17 CEST] <amosbird> I don't know how that is achieved
[19:13:21 CEST] <kepstin> amosbird: I assume so? read the code.
[19:13:29 CEST] <pmjdebruijn> durandal_1707: but can I influence that interpolation
[19:14:12 CEST] <amosbird> ok
[19:14:13 CEST] <pmjdebruijn> the output quality is sufficiently poor, that I'm a bit surprised it's there at all :(
[19:14:30 CEST] <furq> cards: you can get it all in csv but you'd need to specify each field manually
[19:14:35 CEST] <cards> kepstin: indeed. that's just way more information than I wanted
[19:14:46 CEST] <durandal_1707> pmjdebruijn: no, filter uses no interpolation
[19:14:54 CEST] <kepstin> cards: ffprobe has options to allow you to specify exactly which output fields you want
[19:15:04 CEST] <cards> furq: it's possible to cherrypick fields from -show_streams ?
[19:15:07 CEST] <furq> yeah
[19:15:23 CEST] <furq> get rid of -show_streams and just use -show_entries
[19:15:59 CEST] <furq> e.g. -show_entries format=duration,start_time
[19:16:29 CEST] <cards> interesting, thanks.  what's the config file for the windows compile, ie compare, wget has wget.ini and youtube-dl has config.txt
[19:18:49 CEST] <pmjdebruijn> hmmm interesting, I can't reproduce on input photos
[19:19:04 CEST] <dominiques> Hello
[19:19:41 CEST] <dominiques> PLEASE SOMEBODY, CAN you help me solve frame by frame cutting issue ---> /usr/bin/ffmpeg -i "/var/www/html/public/userfiles/input/r_AUTO_autv102_102-2019-09-09_13-29-50_3214qweqwe.mp4" -vf trim=start_frame=222:end_frame=351 -an "/var/www/html/public/userfiles/input/test.mp4"
[19:20:14 CEST] <dominiques> this have problem .... its freezing video first ~3 seconds then playing normally
[19:20:20 CEST] <dominiques> i need frame by frame precision
[19:20:29 CEST] <dominiques> please anyone can help ?
[19:22:53 CEST] <amosbird> btw, can we have different theme to x11grab?
[19:33:08 CEST] <amosbird> hmm, it's now called xcbgrab
[20:27:52 CEST] <kingsley> FYI:
[20:27:55 CEST] <kingsley> $ time ffmpeg  -f lavfi -i mandelbrot=s=1920x1080 -frames 50 /tmp/output.mp4
[20:29:50 CEST] <kingsley> took 26 seconds on a little, old 2 core AMD Athlon computer, but evidently only 5 seconds on a newer 8 core AMD Ryzen computer.
[20:30:51 CEST] <kingsley> I note the ratio of ffmpeg's run times is comparable to how many cores each computer has.
[20:31:53 CEST] <DanielTheKitsune> haha, and I was running ffmpeg on a fairly old CPU
[20:32:01 CEST] <kingsley> Do you happen to know if ffmpeg's rendering speed tends to increase with the number of available CPU cores?
[20:32:26 CEST] <DanielTheKitsune> kingsley: if they implement threading support, yes
[20:32:39 CEST] <DanielTheKitsune> otherwise, they'll just load one of your cores and run the others idle
[20:33:08 CEST] <DanielTheKitsune> modern encoders implement threading support, but there are some niche ones that don't, such as msvideo1
[20:35:33 CEST] <kingsley> DanielTheKitsune: When you wrote "if they implement threading support", did you mean... if ffmpeg and its encoder being used were compiled to support multiple threads and/or cores?
[20:35:45 CEST] <DanielTheKitsune> probably
[20:35:59 CEST] <DanielTheKitsune> although I'm not sure how to explain it, actually ;)
[20:37:56 CEST] <kingsley> Feel free to advise me on the pros and cons of spending extra money for more cores on a new computer that I expect to render video with, probably with the non-linear editor named "kdenlive" and its back end called "melt".
[20:38:51 CEST] <another> threading support depends on the encoder
[20:39:01 CEST] <DanielTheKitsune> hrm, meanwhile I encode AMR audio and MPEG4 video on a P4
[20:39:18 CEST] <DanielTheKitsune> the thing barely decodes h264 1080p
[20:39:26 CEST] <DanielTheKitsune> reaches 90% if doing it realtime
[20:39:31 CEST] <kingsley> ffmpeg used both of my little old computer's cores, so it seems reasonable to me to expect more cores should render faster.
[20:39:59 CEST] <another> as said, depends on the encoder
[20:40:24 CEST] <kingsley> The two options I'm leaning toward are
[20:40:26 CEST] <another> the default for .mp4 is h264 which usually defaults to x264 as the encoder
[20:40:32 CEST] <DanielTheKitsune> AMR encodes quickly and gets very low bitrate for voice-only things
[20:40:41 CEST] <kingsley> 2 CPUS, each with
[20:40:56 CEST] <another> x264 has quite good threading
[20:41:01 CEST] <kingsley> 1.) 8 cores, for a total of 16 cores, and
[20:41:17 CEST] <kingsley> 2.) 18 cores, for a total of 36 cores.
[20:41:26 CEST] <DanielTheKitsune> kingsley: do you have a fast server or something where you quickly encode video in? ;)
[20:42:50 CEST] <kingsley> However, my understanding is each core in the CPU with 18 would have less on chip L3 cache.
[20:43:07 CEST] <kingsley> Plus, the CPUs with more cores are more expensive.
[20:43:27 CEST] <DanielTheKitsune> I've got (remotely) a CPU witha quad-core i5 (god knows which generation) where I remotely encode videos or something, but it's just shared from somewhere
[20:46:08 CEST] <kingsley> another: So maybe I should ask someone to bench mark the encoder I tend to use, .webm, on the multiple core CPUs I'm considering buying, POWER9, to see if their implementations support multiple threads.
[20:46:38 CEST] <DanielTheKitsune> the thing is fast, it's nice
[20:47:06 CEST] <kingsley> DanielTheKitsune: At the moment, I only have a little, old AMD Athlon, which I'm considering upgrading to new hardware of specifications yet to be determined.
[20:47:52 CEST] <kingsley> DanielTheKitsune: Which "thing" is fast and nice?
[20:48:04 CEST] <DanielTheKitsune> Intel Core i5 @ 3 GHz
[20:48:13 CEST] <DanielTheKitsune> I dunno which generation, presumably 4th
[20:49:29 CEST] <kingsley> DanielTheKitsune & another: Thank you for sharing your thoughts. I should be able to proceed more wisely now.
[20:49:46 CEST] <DanielTheKitsune> quad-core beast, very fast compared to my good old P4 @ 2.8 GHz
[20:50:41 CEST] <DanielTheKitsune> I should get one
[20:51:11 CEST] <DanielTheKitsune> brb, work
[22:27:16 CEST] <dominiques> Hi
[22:28:08 CEST] <dominiques> can u help me with recording UDP every hour i need... 9:00:00 to 10:00:00 .. then 10:00:00 to 11:00:00 etc,,,,  ffmpeg -i udp://224.1.1.1:1234 -vcodec copy -acodec copy -f mpegts "output.ts"
[22:28:14 CEST] <dominiques> .
[22:29:00 CEST] Last message repeated 2 time(s).
[22:29:20 CEST] <dominiques> thanks for your helping
[22:40:41 CEST] <dominiques> anyone
[22:40:41 CEST] <dominiques> .
[22:40:45 CEST] Last message repeated 3 time(s).
[22:40:45 CEST] <dominiques> ?
[22:50:21 CEST] <Fenrirthviti> could you not spam the channel with . please? Be patient until someone can look and respond.
[22:55:20 CEST] <another> so what is your problem?
[23:00:01 CEST] <friki-> dominiques: -t to stop recording after 1h. And launch an sh wrapper for the ffmpeg command using cron setting the proper output name using date. That's my aproach
[23:03:23 CEST] <friki-> Probably ffmpeg can handle all together by itself, even setting a proper filename format
[23:06:15 CEST] <friki-> If it's some kind of human controled security recording, I suggest using "-f tee" to record+"re-broadcast". The human condol can be done on this "re-broadcast", and the security recording is also supervised :)
[23:19:40 CEST] <another> you could also look into the segment muxer
[23:45:55 CEST] <Celmor[m]> another: how do I check if my ffmpeg has that decoder?
[23:46:29 CEST] <Celmor[m]> just recompiled and I still appear not to have it
[23:46:46 CEST] <JEEB> if you mean libdav1d it needs to be available as a library
[23:46:57 CEST] <JEEB> you can test with ffmpeg.c and -codecs |grep "dav1d"
[23:47:33 CEST] <Celmor[m]> well, I installed using mpv-full-build-git and checked via `mpv -vd=help | grep dav1`
[23:47:55 CEST] <Celmor[m]> is there an --enable-... option I need for ffmpeg?
[23:49:10 CEST] <another> ffmpeg -codecs | grep dav1d
[23:49:43 CEST] <JEEB> Celmor[m]: yes, if it's not in that list then the FFmpeg that the mpv was linked against doesn't have it at least
[23:49:55 CEST] <JEEB> needs 1) libdav1d installed
[23:50:11 CEST] <JEEB> 2) I think it will get automagically found, but --enable-libdav1d is a good way to make sure configure fials if it's not there
[23:52:11 CEST] <Celmor[m]> alright, now compiling with it
[23:52:30 CEST] <Celmor[m]> another: that won't work as I don't generate an ffmpeg binary but rather mpv statically linked with it
[00:00:00 CEST] --- Sun Oct 13 2019