[Ffmpeg-devel-irc] ffmpeg.log.20171009

[burek]

[00:09:19 CEST] <kerio> how are you sending it to the tv?
[00:09:40 CEST] <kerio> whoops that was way in the backscroll nvm
[00:10:06 CEST] <JEEB> the TV I have supports both DLNA and miracast. not sure which the win10 built-in thing is using but probably miracast
[00:10:42 CEST] <kerio> isn't miracast just screen mirroring
[00:10:47 CEST] <JEEB> no
[00:11:30 CEST] <JEEB> although I do remember that being one of the primary use cases for it when I was working at a company implementing it
[00:12:35 CEST] <JEEB> (basically I'm pretty sure the "Cast to Device" thing in Win10 uses Miracast)
[00:12:57 CEST] <JEEB> although it also supports the screen sharing via it
[00:19:21 CEST] <kerio> yea but
[00:19:32 CEST] <kerio> doesn't that mean that it's supposed to be transcoded?
[00:24:57 CEST] <JEEB> no
[00:25:08 CEST] <JEEB> I mean, what's going over the pipe is H.264 and pals anyways
[00:38:41 CEST] <Djfe> hi, is it possible to tell ffmpeg to ignore 404s? I'm getting 404s while downloading an HLS over VPN and ffmpeg just skips to the next file of the playlist instead of retrying (at least it seems so). I know that 404s mean "the file is unavailable right now, but there might be something here available in the future" (in terms of HLS context)
[00:45:35 CEST] <Djfe> it's used in a vod context and not a livestream,so the segments won't disappear
[00:50:24 CEST] <DHE> no. that's broken. the file should be available, then the playlist is updated. vod is entirely more stringent
[00:54:56 CEST] <Djfe> ok thx
[00:56:01 CEST] <Djfe> bye
[01:35:54 CEST] <bodqhrohro1> Are now any plans of OnTextData flv packet implementation?
[02:57:59 CEST] <hendry> how do i quickly convert Matroska data to an mp4 that FCPX understands?
[03:02:51 CEST] <tdr> quickly is such a relative word, the gods would say 1000 yrs wasnt so long :)
[03:06:29 CEST] <DHE> I assume it means no transcoding. ffmpeg -i input.mkv -c copy output.mp4
[03:11:40 CEST] <Johnjay> question
[03:11:52 CEST] <Johnjay> what's the smallest size mp4 file that would probably play in VLC/wmp?
[03:11:54 CEST] <Johnjay> 200 KB?
[03:12:26 CEST] <Johnjay> by play I mean something like show the first frame
[03:27:31 CEST] <hendry> DHE: Could not find tag for codec pcm_s16le in stream #1, codec not currently supported in container <-- when i try `ffmpeg -i input.mkv -c copy output.mp4`
[03:32:20 CEST] <relaxed> hendry: try output.mov
[03:37:09 CEST] <hendry> relaxed: the resulting output.mov is just black in the import pane of FCPX
[03:48:09 CEST] <DHE> then you need to transcode the audio. aac codec at 64k per channel (eg: 128k for stereo) as bitrate shoul dbe okay
[03:49:46 CEST] <hendry> DHE: for MP4 ?
[03:51:33 CEST] <hendry> DHE: ffmpeg -i input.mkv -c copy -acodec aac output.mp4 # doesn't work on FCPX. No audio and black
[04:42:17 CEST] <FishPencil> Is there a way to combine a bunch of video thumbnails into one big image, say x images wide and y images high?
[04:45:28 CEST] <hendry> FishPencil: probably with convert http://dabase.com/e/01046/
[04:47:27 CEST] <FishPencil> So extract the frames with FFmpeg and create a large image with imagemagick? I'm surprised FFmpeg can't do this
[04:48:42 CEST] <furq> you can do it with ffmpeg
[04:48:44 CEST] <furq> !filter tile
[04:48:44 CEST] <nfobot> furq: http://ffmpeg.org/ffmpeg-filters.html#tile
[04:51:52 CEST] <FishPencil> Nice, is it then possible to divide up the entire stream into n number of thumbnails, tiling them in the same go?
[04:54:43 CEST] <furq> there's a thumbnail filter but you'll need to know the total number of frames in the file first
[04:54:59 CEST] <furq> also it buffers the entire selection in memory so you probably want to decimate or scale that beforehand
[04:55:05 CEST] <furq> it's generally not ideal
[04:58:50 CEST] <FishPencil> for tile do I need to know how many frames, or can I just have it do a set width amount and keep going down till it's done?
[07:20:11 CEST] <echelon> hey guys
[07:20:32 CEST] <echelon> is there a way to distribute encoding across network? like distcc?
[08:46:52 CEST] <frzs> Hi! I'm using ffprobe it to filter valid video files from a folder, but I found something strange: if I run the following command: " ffprobe.exe myimage.jpg " ffprobe will show " Duration: 00:00:00.04 " .
[08:47:34 CEST] <frzs> The problem is that in my case this means a valid video file, as I'm checking if(duration > 0) for a valid video file
[08:47:49 CEST] <e38383> 1 frame has this duration
[08:48:28 CEST] <e38383> maybe check if there is a video part
[09:14:35 CEST] <panpeter> good morning @ all :) i am totally new to ffmpeg (used gstreamer before). Nvidia says ffmpeg supports nvenc and nvdec completely. Im running ubuntu 16.04 with a gtx 750 (which supports at least second generation nvenc) if ill want to de- and encode an rtsp-stream in my own application, do i have to compile ffmpeg on my own or can i use ffmpeg from the default ubuntu repositories? how can i...
[09:14:37 CEST] <panpeter> ...check nvenc is working correctly?
[09:18:48 CEST] <stevenliu> ffmpeg -h encoder=h264
[09:20:43 CEST] <panpeter> stevenliu: this prints a wall of text (options) for the following encoders(?): libx264, libx264rgb
[09:20:47 CEST] <Nacht> panpeter: If you run ffmpeg, you see in the banner what configuration its build with
[09:20:53 CEST] <Nacht> panpeter: it should say --enable-nvenc
[09:21:43 CEST] <panpeter> Nacht: so it seems the default ubuntu package does not contain it
[09:22:23 CEST] <Nacht> panpeter: I'm not 100% sure tho. If I read the page about it, it does say the following.  "ffmpeg configured without --disable-nvenc "
[09:22:37 CEST] <stevenliu> look at the banner
[09:22:38 CEST] <Nacht> So I'm not sure if enable-nvenc is on by default
[09:22:55 CEST] <stevenliu>   configuration: --enable-fontconfig --enable-gpl --enable-libass --enable-libbluray --enable-libfreetype --enable-libmp3lame --enable-libspeex --enable-libx264 --enable-libx265 --enable-libfdk-aac --enable-version3 --cc='ccache gcc' --enable-nonfree
[09:23:03 CEST] <stevenliu> for example, my ffmpeg banner is this
[09:23:15 CEST] <Nacht> But simplest to test it with, just get a file and try to encode it
[09:25:27 CEST] <panpeter> Yes, "--disable-nvenc" is not activated, but also "--enable-nvenc" is not activated, it seems as a default it is disabled
[09:28:00 CEST] <panpeter> https://trac.ffmpeg.org/wiki/HWAccelIntro says if not --disable-nvenc is set, it will compiled with nvenc support :D
[09:28:31 CEST] <panpeter> but also it seems i need the NVIDIA Video Codec SDK
[09:30:25 CEST] <panpeter> Okay and i think shipping the NVIDIA Video Codec SDK would be a license flaw. So it seems nvidia doesn't want normal end-customers to have encoding and decoding capabilities used under linux systems
[09:32:50 CEST] <Nacht> From what I've heard here, using the GPU doesn't really improve much. Not experianced it myself yet though
[09:33:50 CEST] <panpeter> Nacht: ill do not need a rendered fps improvement, ill only need to hold the cpu free from rendering load
[09:37:22 CEST] <stevenliu> use nvenc is better more than CPU
[09:38:22 CEST] <Nacht> panpeter: I'm not sure how easy it is to get into the developer program. (ie: just signing up). But this document will help in the steps further: http://developer.download.nvidia.com/compute/redist/ffmpeg/1511-patch/FFMPEG-with-NVIDIA-Acceleration-on-Ubuntu_UG_v01.pdf
[09:38:29 CEST] <stevenliu> speex is high than cpu, for exmaple, tesla K20 , speex is 20x , but cpu is 1x, 1080p 10M to 1080p 2.5M
[09:39:54 CEST] <Nacht> panpeter: This might be intresting: http://git.videolan.org/?p=ffmpeg.git;a=commit;h=325e56479ff64c884f3bcccf922a7f7163488b89
[09:40:11 CEST] <Nacht> panpeter: Try the following: ffmpeg -encoders 2>/dev/null | grep nvenc
[09:41:46 CEST] <panpeter> Nacht: it prints nothing, also "ffmpeg -h encoder=h264_nvenc" and "ffmpeg -h encoder=hevc_nvenc" deliver "Codec xyz is not recognized by FFmpeg."
[09:42:04 CEST] <Nacht> Hmm. Doesn't seem you have the encoder installed then.
[09:42:33 CEST] <Nacht> panpeter: According to this: https://askubuntu.com/questions/778100/how-to-install-compile-nvenc-in-ubuntu
[09:42:39 CEST] <Nacht> if you compile it yourself, you will have it
[09:43:15 CEST] <panpeter> Nacht: the main problem is, ill have to ship a working hw-accelerated software to customers, so i think shipping the needed nvidia video codec SDK is restricted, but this may has to be checked by a lawyer
[09:43:50 CEST] <Nacht> panpeter: Yeah, that's def worth a check then. Might even be worth signing up as a developer at nvidia then
[09:44:28 CEST] <panpeter> Nacht: i already have done signed up as a developer ;)
[09:45:13 CEST] <panpeter> Nacht: ah i found some thing that goes into the right direction http://git.videolan.org/?p=ffmpeg.git;a=commit;h=325e56479ff64c884f3bcccf922a7f7163488b89
[09:46:45 CEST] <panpeter> Nacht: maybe only my ffpmeg version is to old. I think i can check out the rest myself, thank you very much
[13:05:18 CEST] <paradoxspiral> Is there an option when seeking with ms precision that will seek to the nearest frame? I'd like to seek to e.g. 1.86085. Or do I have to convert to the nearest frame myself?
[13:06:45 CEST] <JEEB> do note that if you want frame-exact seeks then you want something like ffms2
[13:06:54 CEST] <JEEB> which creates indexes based on FFmpeg's libraries and decoders
[13:07:19 CEST] <JEEB> because just seeking with av_seek_frame is not always frame-exact depending on the format.
[13:07:43 CEST] <JEEB> so if you need frame-exact seeks and don't want to reinvent wheels then take a look at ffms2
[13:07:53 CEST] <JEEB> https://github.com/FFMS/ffms2
[13:07:57 CEST] <paradoxspiral> Well, I'd like to keep it more simple, since this is for a Lua mpv script.
[13:08:20 CEST] <JEEB> well then you don't have access to the API at all
[13:08:25 CEST] <JEEB> libavformat/-codec
[13:08:26 CEST] <paradoxspiral> yeah :/
[13:08:35 CEST] <JEEB> I'm not sure what you were expecting to do
[13:08:54 CEST] <paradoxspiral> I'm invoking ffmpeg via os.execute
[13:09:11 CEST] <JEEB> because screenshots with mpv aren't possible or something?
[13:09:30 CEST] <JEEB> I'd like to know the actual use case without the mpv lua script part in general.
[13:10:05 CEST] <paradoxspiral> The script creates clips of the current video, and the start/end point is in the form of second.ms
[13:10:23 CEST] <JEEB> do you need stream copy or do you always re-encode?
[13:10:28 CEST] <paradoxspiral> re-encode
[13:10:39 CEST] <JEEB> ok, then if you are OK with slowness you can use -ss after -i
[13:10:43 CEST] <JEEB> which is slow but frame-exact
[13:10:55 CEST] <JEEB> or well, it should be frame-exact :P (since it decodes A->B until that point)
[13:11:33 CEST] <paradoxspiral> Does that re-encode the vid upto that point?
[13:11:42 CEST] <JEEB> no, it just decodes
[13:11:53 CEST] <paradoxspiral> then it's not tooo bad I gues
[13:11:55 CEST] <paradoxspiral> +s
[13:12:13 CEST] <JEEB> as in, what you really need is seek + decode, but in that case using ffms2 through some cli app/script would make more sense :P
[13:12:28 CEST] <JEEB> so I guess "just wait until it decodes isn't /that/ bad"
[13:13:37 CEST] <paradoxspiral> Yeah^^, thanks :3
[13:15:07 CEST] <JEEB> basically in the API of FFmpeg you can do av_seke_frame with the AVSEEK_FLAG_BACKWARDS , which should give you the closest random access picture from before the point. except when the format is not easily exactly seekable (say, MPEG-TS). of course you then have to decode until your point... at which point you're recreating parts of ffms2 :D
[14:46:12 CEST] <MacroMan> Is there anything I can do to lower output file size, other than reducing the sample rate (just audio)?
[14:46:29 CEST] <MacroMan> Sorry, bitrate, not sample rate
[14:46:47 CEST] <Mavrik> Lower the bitrate and number of channels
[14:46:55 CEST] <MacroMan> I'm already down to 32kbps, but any lower I lose audio quality.
[14:47:10 CEST] <Mavrik> Which format?
[14:47:13 CEST] <MacroMan> mp3
[14:47:23 CEST] <MacroMan> I think I'm at the floor limit, but I thought I'd ask
[14:47:26 CEST] <Mavrik> Can you choose another one?
[14:47:29 CEST] <Mavrik> Yeah, 32k is really low
[14:47:42 CEST] <Mavrik> Although switching to HE-AACv2 will probably give you quite a boost in quality
[14:49:54 CEST] <MacroMan> Mavrik: You'll have to give me a few minutes to read up on HE-AACv2
[14:50:00 CEST] <MacroMan> I have no idea what that is.
[14:50:21 CEST] <JEEB> a profile for AAC made for low bit rate audio
[14:50:41 CEST] <Mavrik> MacroMan: it's a high efficiency version of AAC made for <64kbit usecases
[14:50:43 CEST] <JEEB> although personally if possible I'd utilize opus instead if there is no device support required
[14:51:01 CEST] <Mavrik> It sounded significantly better for me than MP3/AAC-LC at those low bitrates and it's quite widely supported still
[14:53:49 CEST] <MacroMan> Is there a command flag to set to HE-AACv2 or opus?
[14:54:31 CEST] <MacroMan> Ah, think I've found it: `-c:a libfdk_aac -profile:a aac_he_v2`
[14:58:12 CEST] <Mavrik> MacroMan: yep, that looks right
[14:58:23 CEST] <Mavrik> you'll need to use HE-AAC (without v2) for mono content
[14:58:51 CEST] <MacroMan> OK thanks. Do I have to compile ffmpeg myself for HE-AAC support? I'm getting this error: 'Unknown encoder 'libfdk_aac''
[14:59:30 CEST] <JEEB> yes, fdk-aac isn't compatible with FFmpeg's license and thus you can't distribute the two linked together in binary form
[14:59:39 CEST] <Mavrik> Does the builtin AAC support he-aac?
[14:59:46 CEST] <JEEB> no idea, I would guess not
[14:59:51 CEST] <Mavrik> MacroMan: FDK is an AAC encoding library, similar to LAME for mp3
[14:59:52 CEST] <JEEB> atomnuker would know
[14:59:55 CEST] <MacroMan> Damn. I didn't really want to get this deep into this project.
[15:00:01 CEST] <JEEB> MacroMan: what's your client
[15:00:02 CEST] <Mavrik> Then just use mp3 :)
[15:00:06 CEST] <JEEB> as in what will play back the thing
[15:00:20 CEST] <MacroMan> I'll stick with that for now, thanks
[15:00:50 CEST] <atomnuker> use opus lol
[15:01:40 CEST] <Mavrik> Yeah, use opus if target supports it :P
[15:01:46 CEST] <Mavrik> Do mobile devices have decoders yet?
[15:02:04 CEST] <MacroMan> It only needs to be supported in Chrome and windows media player
[15:02:12 CEST] <MacroMan> I'll have a look at opus
[15:04:49 CEST] <JEEB> chrome supports opus just fine, WMP probably not
[15:08:29 CEST] <MacroMan> Unfortunately, the opus file isn't any smaller than mp3. I'll stick with that for now. Thanks for all the advice.
[15:08:54 CEST] <JEEB> well if you were setting a low bit rate only the quality would be better
[15:09:08 CEST] <JEEB> of course not the file size since you are defining the bit rate
[15:09:15 CEST] <JEEB> bit rate over duration is the file size
[15:37:26 CEST] <MacroMan> Wow, actually managed to lower the bitrate to 8kbps using opus, and it's still alright quality.
[15:37:37 CEST] <redrabbit> sample?
[15:37:40 CEST] <MacroMan> And about a 10th the file size of wav
[15:37:54 CEST] <redrabbit> you mean 80k right
[15:38:28 CEST] <MacroMan> These are voice recordings though, so the quality is not top priority.
[15:38:35 CEST] <MacroMan> What ever 7.0287kbit/s is
[15:38:43 CEST] <MacroMan> 7kbps?
[15:39:24 CEST] <redrabbit> kbit/s =/= kbps
[15:39:40 CEST] <MacroMan> Ah OK. Is that 70k then?
[15:40:22 CEST] <redrabbit> im also confused now
[15:40:22 CEST] <redrabbit> :p
[15:41:18 CEST] <MacroMan> kbit is kilobit and kb is kilobyte, I think.
[15:41:33 CEST] <MacroMan> Correct me if I'm wrong.
[15:42:08 CEST] <redrabbit> https://en.wikipedia.org/wiki/Data_rate_units
[15:42:55 CEST] <redrabbit> even more confusing lmao
[15:43:01 CEST] <MacroMan> kilobit per second (symbol kbit/s or kb/s, often abbreviated "kbps")
[15:43:16 CEST] <MacroMan> So Wikipedia are saying kbit/s === kbps
[15:43:53 CEST] <MacroMan> Bytes is kB/s
[15:44:43 CEST] <JEEB> that's a whole mess, just get used to kilobits (per second) when talking of bit rates
[15:44:57 CEST] <JEEB> (and dividing/multiplying by 1000)
[15:45:30 CEST] <MacroMan> lol. Whatever the correct symbol, I set the bitrate to 8 and it sounds fine, with an excellent file size.
[15:46:32 CEST] <Nacht> B = byte b = bit
[15:47:47 CEST] <redrabbit> 8kbps for audio sounds low
[15:48:05 CEST] <redrabbit> wav is 1440k
[15:48:43 CEST] <redrabbit> so maybe its really 8 kBps
[15:48:58 CEST] <JEEB> if it's -b:a 8k that's 8 kilobits
[15:49:03 CEST] <JEEB> and that's an OK bit rate for opus
[15:49:11 CEST] <JEEB> for certain use cases
[15:49:19 CEST] <MacroMan> From the man: Target bitrate in kbit/sec (6-256 per channel)
[15:49:20 CEST] <JEEB> it won't keep the audio the same, but it will not be horrible horrible stuff
[15:49:42 CEST] <redrabbit> can you up a small sample
[15:49:56 CEST] <MacroMan> You'll need to give me a few minutes
[15:50:09 CEST] <MacroMan> Most of the recordings are confidential
[15:50:12 CEST] <redrabbit> JEEB: he's saying "about a 10th the file size of wav"
[15:50:39 CEST] <JEEB> well you're thinking his source was of specific rate or bit depth
[15:50:42 CEST] <JEEB> which it might not be
[15:50:50 CEST] <MacroMan> Yes, opus: 0.8mb vs wav 15.1mb.
[15:50:54 CEST] <redrabbit> yep i might be another bitrate and just mono
[15:51:04 CEST] <JEEB> stop calling sample rate a bit rate
[15:51:24 CEST] <MacroMan> The original file doesn't have a bitrate as it's wav. The sample rate was only 8000Hz though
[15:51:31 CEST] <JEEB> there you have it :P
[15:51:34 CEST] <redrabbit> !
[15:51:36 CEST] <redrabbit> ok
[15:52:02 CEST] <JEEB> MacroMan: it does have a constant bit rate. sample rate * bits per sample * channel count
[15:52:08 CEST] <MacroMan> These files are produced by Asterisk btw, which is super weird with the way it handles audio
[15:52:36 CEST] <JEEB> but what redrabbit was mentioning as "bit rate" is not bit rate :P (sample rate by itself)
[15:53:28 CEST] <MacroMan> JEEB: Didn't know that. Ubuntu tells me n/a on the bitrate for wav files for some reason.
[15:54:03 CEST] <JEEB> well it's a constant amount of bits per second because it's not compressed, so it doesn't have a *compressed* bit rate as such :P
[15:54:12 CEST] <JEEB> but it does have a bit rate according to the conten
[15:54:13 CEST] <JEEB> *content
[15:54:19 CEST] <JEEB> raw amount of bits
[15:58:03 CEST] <MacroMan> Ah I see. Thanks.
[16:24:12 CEST] <wouterstreamit> I am trying to emulate avcodec_decode_video2 using avcodec_receive_frame. Am I correct that I should expect AVERROR(EAGAIN) if there is no data available yet and should poll until there is data available?
[17:57:54 CEST] <tomreyn> hi there. does ffmpeg already support the (proprietary) D(H)AV file format from Dahua (as used by their CCTV cameras)?
[17:59:20 CEST] <tomreyn> I have ffmpeg 2.8.11 here (ubuntu's) but -formats does not suggest to support it. I came across GPLv4 software http://tanidvr.sourceforge.net/ which states to support it, though.
[17:59:27 CEST] <tomreyn> *GPLv3
[17:59:55 CEST] <Johnjay> ok is there a simpler way to add silence to end of audio than this?
[18:00:05 CEST] <Johnjay> I don't mean creating a temp file 10SecSilence.mp3 or whatever
[18:00:07 CEST] <Johnjay> https://pastebin.com/UykBUzx0
[18:05:49 CEST] <durandal_1707> Johnjay: see apad filter
[18:06:50 CEST] <Johnjay> thanks
[18:09:30 CEST] <Johnjay> it wants samples, so would I do 48000 * N for N seconds of padding?
[18:11:34 CEST] <Johnjay> I have a webm audio input and need a mp3 output with padding at the end
[18:17:16 CEST] <durandal_1707> yes
[21:17:48 CEST] <fsphil> is there a way to include subtitles in an MKV file but not have them display by default?
[21:25:16 CEST] <Fenrirthviti> fsphil: Pretty sure that's a player setting, not something you can control in the source file.
[21:32:59 CEST] <CoreX> fsphil if i remember its -disposition:s:0 0
[21:33:14 CEST] <fsphil> trying that now
[21:38:13 CEST] <fsphil> no, no difference
[21:43:59 CEST] <nostrora> Google use ffmpeg for Youtube ?
[21:44:14 CEST] <durandal_1707> no
[21:44:40 CEST] <nostrora> there is an parameter i can add to ffmpeg to delete input file if encoding is correcty done ?
[21:44:43 CEST] <JEEB> you forgot them patching main() ?
[21:45:05 CEST] <JEEB> also how GOOG is now trying to send in all of their patches :P
[21:45:23 CEST] <durandal_1707> whut.?
[21:45:39 CEST] <relaxed> didn't they use a custom mencoder at on point?
[21:45:42 CEST] <JEEB> yes
[21:45:50 CEST] <relaxed> one*
[21:45:51 CEST] <JEEB> then they were basically renaming ffmpeg.c's main()
[21:45:55 CEST] <JEEB> and calling it from their app
[21:45:59 CEST] <JEEB> and/or library
[21:53:51 CEST] <nostrora> there is an parameter i can add to ffmpeg to delete input file if encoding is correcty done ?
[21:54:10 CEST] <JEEB> no
[21:54:24 CEST] <JEEB> use ffprobe with json output, decide by yourself
[21:56:11 CEST] <nostrora> ok :)
[23:43:42 CEST] <rabbe> anyone have a magewell capture card (1080P @ 60FPS -> USB3)? which ffmpeg command for saving mp4 to hdd will give me max quality?
[23:44:48 CEST] <dystopia_> capturing as raw video will give you best quality
[23:44:54 CEST] <dystopia_> but you probably don't want to do that
[23:45:41 CEST] <rabbe> no, probably h.264+aac .mp4 is best?
[23:46:04 CEST] <rabbe> i mean, best quality per reasonable size
[23:46:46 CEST] <rabbe> or mkv?
[23:47:00 CEST] <dystopia_> mkv and mp4 are just containers
[23:47:06 CEST] <dystopia_> what is it you are capturing?
[23:47:11 CEST] <dystopia_> tv, video games?
[23:47:26 CEST] <rabbe> yeah, so lets go with mp4. i think the customer will like that most
[23:47:37 CEST] <rabbe> it will be from a gopro mounted on a robot arm
[23:48:02 CEST] <rabbe> but the signal will go through a whdi link
[23:48:14 CEST] <llogan> how will the customer use these files? just for viewing, or uploading to YouTube, or editing, etc.
[23:48:35 CEST] <rabbe> just for viewing
[23:49:09 CEST] <llogan> i'm guessing you want something like: ffmpeg -i input -c:v libx264 -crf 18 -vf format=yuv420p -metadata title="Robot Video" output.mp4
[23:49:29 CEST] <llogan> see https://trac.ffmpeg.org/wiki/Encode/H.264
[23:49:37 CEST] <rabbe> alright
[23:49:49 CEST] <rabbe> thanks
[23:50:51 CEST] <rabbe> i did some testing and then i needed to tell the resolution
[23:51:30 CEST] <rabbe> but i did some check on magewells web page now and found they have some software to set up the resolution etc
[23:51:54 CEST] <rabbe> (i mean i had to specify in the ffmpeg command about the resolution)
[00:00:00 CEST] --- Tue Oct 10 2017