[Ffmpeg-devel-irc] ffmpeg.log.20190323

[burek]

[00:45:32 CET] <Matti-> i found a amazing program that scans mp3 files and tell which mp3 encoder was used.   if this could happen with mp3,  is it possible to have a program that does for AAC, and tells which AAC encoder used
[00:46:00 CET] Last message repeated 1 time(s).
[00:50:43 CET] <JEEB> some encoders leave visible metadata in the bit stream. if that is utilized it's possible to look into that. the problem of course is that there's no way to be sure if that thing is valid or not
[00:51:24 CET] <JEEB> if it's actual hints by how an encoder encodes thing, and not random metadata
[00:51:41 CET] <Matti-> jeeb would you like to know program name?
[00:51:55 CET] <JEEB> then that might work better as long as those structures are well-noted to be specific to that encoder
[00:52:30 CET] <JEEB> so far of the latter category I only know a single video encoder that has a known thing that you can spot it from even if metadata is cleared out :P
[00:52:52 CET] <JEEB> but I do not know if there's an automated application to find those hints out
[00:52:56 CET] <Matti-> jeeb it's not perfect but it can differenciate  lame/bladeenc/fdk/gogo/xing
[00:53:16 CET] <Matti-> i never heard of "gogo"
[00:55:40 CET] <Matti-> jeeb have you heard of all of them here:  lame/bladeenc/fdk/gogo/xing
[01:00:27 CET] <faLUCE> JEEB: just checked that MPEGTS NEEDS a starting video keyframe for audio-video stream
[01:04:13 CET] <BtbN> I don't see how that information is overly useful though?
[01:04:35 CET] <Matti-> btbn what information
[01:04:49 CET] <faLUCE> BtbN: it was a past discussion with JEEB
[01:05:02 CET] <faLUCE> (just for updating him)
[01:36:26 CET] <Matti-> what do you people think about PCM vs DSD ?  and does ffmpeg support DSD?
[01:56:46 CET] <buhman> Matti-: you mean PDM?
[01:56:53 CET] <Matti-> no
[01:57:02 CET] <Matti-> PCM and DSD
[01:57:11 CET] <buhman> those aren't comparable things
[01:57:15 CET] <Matti-> yes it is
[01:57:46 CET] <Matti-> buhman i am guessing you don't know what they are
[04:27:05 CET] <locsmif> Hi all. Trying to play an .m3u8 from local disk, i.e. ffmpeg -i C:\Temp\playlist.m3u8 -c copy -o output.mp4, but ffmpeg is giving me: [https @ 000000000053ed80] Protocol not on whitelist 'file,crypto'! c:\temp\playlist.m3u8: invalid argument - what am I doing wrong?
[04:28:55 CET] <furq> locsmif: -protocol_whitelist file,tcp,https
[04:34:09 CET] <locsmif> Now it's giving me a TLS error, but it works fine in VLC
[04:34:29 CET] <furq> if it's another whitelist error then add tls to that string
[04:35:23 CET] <furq> or just -protocol_whitelist ALL
[04:35:34 CET] <locsmif> TLS is not an URI scheme though
[04:36:07 CET] <furq> ?
[04:36:40 CET] <locsmif> Ffmpeg is trying to tell me it doesn't  like the TLS handshake with the server for some reason, right? (Besides, I just tried adding it to -protocol_whitelist anyways and as expected it failed)
[04:36:59 CET] <furq> what's the actual error
[04:39:21 CET] <locsmif>  Creating security context failed (0x80090302)
[04:39:46 CET] <locsmif> It works fine in VLC/Browsers, etc.
[04:41:28 CET] <ossifrage> I thought I had finally fixed my bframe muxing problems (computing DTS) only to find that chrome is presenting the frames out of order, joy
[04:41:34 CET] <locsmif> I think I should upgrade.
[04:43:45 CET] <furq> locsmif: if upgrading doesn't work then -tls_verify 0 should work around it
[04:43:49 CET] <furq> obviously that's not ideal
[04:48:31 CET] <locsmif> ffmpeg now works after upgrade. Switch back to -i "https://..." for the moment as I wanted to get rid of that factor in the equation
[04:53:00 CET] <locsmif> ffplay however doesn't wan to play the audio, because: SDL_OpenAudio (2 channels, 44100 Hz): WASAPI can't initialize audio client: CoInitialize has not been called., then:
[04:53:25 CET] <locsmif> SDL_OpenAudio (1 channels, 44100 Hz): WASAPI can't initialize audio client: CoInitialize has not been called. === followed by: No more combinations to try, audio open failed
[04:55:29 CET] <locsmif> But, okay, I used the nightly GIT build
[04:55:58 CET] <locsmif> Seems to be covered here: https://trac.ffmpeg.org/ticket/6891
[05:54:47 CET] <friendofafriend> Hi there, I'm getting "Invalid UE golomb code" when I'm playing an MPEG-TS stream created by ffmpeg.  Is there something I should add to my ffmpeg encoder's command line to fix it?
[10:36:19 CET] <Janosch__> I have a question regarding the conversion between a 'gray8' input and e.g. a 'yuv420p' output. I've learned that the 'Y' component in the YUV420p format is supposedly an 8-bit grayscale value. However, when i convert an 8-bit grayscale image to yuv using either ffmpeg.exe or 'sws_scale' from libav, the color is changed, so the spectrum doesn't include very dark and very bright values. When i convert back from yuv420 to gray8
[10:36:19 CET] <Janosch__> (e.g. as PNG output) the colors are almost correct again, with every 4th color gray-value missing from the histogram. why is that?
[11:53:02 CET] <__raven__> how to force a video pid in an flv container in cases the input rtmp is that bad that no video is recognized while probing?
[11:54:40 CET] <BtbN> Pretty sure you don't.
[11:54:45 CET] <BtbN> Increase the probe duration.
[11:55:47 CET] <pkunk> Well, that was much easier than I though it would be
[12:01:28 CET] <pkunk> As I discussed in the channel yesterday, I've modified the ffmpeg cli to support opening inputs in parallel
[12:02:56 CET] <pkunk> It works very well with only one issue.. The inputs are now out of order at the filter level, so in my mosaic use case the overlays are now in random with 0:v having the frame from the 5th input instead
[12:04:46 CET] <pkunk> Is there any function or OptionsContext struct field I can use to "request" a particular order for the streams probed in parallel
[12:17:07 CET] <BtbN> You wrote that code, how would anyone here know?
[15:18:52 CET] <kepstin> pkunk: you're gonna have to have a serialization step that collects all the asynchronously probed inputs in the order specified on the command line
[15:20:59 CET] <kepstin> Janosch__ is gone, but their problem was that they were converting from full range gray8 to limited range yuv420p :/
[16:45:11 CET] <SnakesAndStuff> I'm pretty sure I need to convert a -vf into a -filter_complex to do what I want... I'm trying to put a black box over an area of a video as well as speed up the video....
[16:45:39 CET] <SnakesAndStuff> I want to use -vf "color=black:581x53 [over]; [in][over] overlay=75:69 [out]" and "setpts=0.05*PTS" at the same time.... How do I combine these filters?
[16:48:56 CET] <furq> overlay=75:69,setpts=0.05*PTS
[16:53:21 CET] <SnakesAndStuff> I think I was putting it in the wrong spot....
[16:59:15 CET] <locsmif> friendofafriend: still no response huh
[16:59:23 CET] <locsmif> friendofafriend: I'd repost the question
[17:12:20 CET] <jkhsjdhjs> hey, i'm recording an rtsp stream with ffmpeg. i'm segmenting it in files à one hour. the first file starts at 0:00:00. however, the following files don't start at 0:00:00, but at 1:00:00, 2:00:00 etc. the problem is that all segments end after one hour, thus many segments have a starting timecode that is higher than the end timecode, e.g. a file starts at 4:00:00 and ends at 1:00:00. is there a way to make
[17:12:21 CET] <jkhsjdhjs> every segment start at 0:00:00? here's the full command i'm using to capture the stream: https://hastebin.com/ogaxiwequh.bash
[17:20:23 CET] <locsmif> jkhsjdhjs: tried -reset_timestamps ?
[17:20:36 CET] <jkhsjdhjs> nope, will do, thanks!
[17:22:12 CET] <locsmif> It belongs after -f segment amongst your other -segment_time / -segment_atclocktime options. I'm not sure if it works but I should hope so.
[17:23:43 CET] <jkhsjdhjs> nice, placed it after -segment_atclocktime
[17:27:59 CET] <jkhsjdhjs> yep, that seems to work :D
[17:30:29 CET] <jkhsjdhjs> do you perhaps know where i can find information on segmentation? i looked in man ffmpeg, but it contains the word segment just once. is there a seperate manpage for segmentation?
[17:32:04 CET] <JEEB> the manpages are all separated into parts
[17:32:06 CET] <JEEB> https://www.ffmpeg.org/ffmpeg-all.html
[17:32:11 CET] <JEEB> this is a link to the full one
[17:32:19 CET] <JEEB> the exact options depend on the module you're using
[17:33:33 CET] <jkhsjdhjs> alright, thanks!
[17:33:38 CET] <furq> jkhsjdhjs: https://www.ffmpeg.org/ffmpeg-formats.html#segment_002c-stream_005fsegment_002c-ssegment
[17:34:33 CET] <jkhsjdhjs> thank you too furq!
[17:36:23 CET] <furq> reset_timestamps is listed in there as doing the exact thing you want so it should work
[17:36:34 CET] <furq> i guess we'll find out in 44 minutes
[17:38:24 CET] <locsmif> jkhsjdhjs: great =)
[17:51:04 CET] <faLUCE> hello all
[17:53:07 CET] <JEEB> faLUCE: so which part of H.222 says you have to start the mux from a RAP?
[17:53:19 CET] <JEEB> (H.222 = MPEG-2 Systems, contains MPEG-PS and MPEG-TS)
[17:53:33 CET] <JEEB> or is it the FFmpeg muxer specifically and not a format specification limitation? :P
[17:53:40 CET] <faLUCE> JEEB: with RAP do you mean a video keyframe?
[17:53:54 CET] <JEEB> random access point, keyframe is just one, rap promises that you can decode what comes after
[17:53:59 CET] <JEEB> which is an important distinction
[17:54:13 CET] <faLUCE> [17:53] <JEEB> or is it the FFmpeg muxer specifically and not a format specification limitation? :P  <--- pheraps this one
[17:54:39 CET] <JEEB> yea, because as far as I looked at MPEG-TS I don't think the spec says you can't start the mux with a non-RAP :P
[17:55:24 CET] <faLUCE> JEEB: I told that because I had issues (-->warnings) in decoding a mpegts audio video starting with video non keyframes
[17:55:26 CET] <JEEB> in H.264 a RAP is either an IDR NAL unit, or the open GOP random access SEI + normal intra NAL :P
[17:55:28 CET] <DHE> considering mpegts is a streaming protocol, it sounds reasonable that it would allow it to happen that way, though from a muxer standpoint it sounds pretty stupid to do that...
[17:55:42 CET] <JEEB> faLUCE: that's normal
[17:55:49 CET] <JEEB> the decoder will warn you that it couldn't decode frames
[17:56:01 CET] <JEEB> you just keep feeding the decoder until you get a RAP
[17:56:07 CET] <JEEB> so yea, it's not a muxer limitation either then
[17:56:33 CET] <JEEB> you will get the same warnings when you start decoding a live TV station :P
[17:56:38 CET] <JEEB> until you get a RAP
[17:56:40 CET] <faLUCE> but I think that warning is not proper for an audio/video demuxer. I mean, the red color let you think about an error
[17:57:34 CET] <JEEB> yea, I think that's because it's hard to distinguish decoding errors before you have gotten your first RAP and after you have received your first RAP :P
[17:57:44 CET] <kepstin> assuming the demuxer sets the keyframe flag appropriately, you should be able to just drop packets until you get a RAP and only then start decoding?
[17:57:48 CET] <JEEB> so you can get errors in your logs until you get your first RAP, but that's 100% OK
[17:57:51 CET] <JEEB> kepstin: yes
[17:58:07 CET] <faLUCE> kepstin: but it's not mine demuxer, it's ffplay
[17:58:25 CET] <faLUCE> JEEB: ^
[17:58:39 CET] <JEEB> sure, ffplay.c doesn't skip packets
[17:58:53 CET] <faLUCE> then, that warning for mpegts in not proper
[17:58:57 CET] <faLUCE> is
[17:59:13 CET] <JEEB> yes, but now go around changing the messages to depend on whether you have already received your first RAP or not
[17:59:19 CET] <JEEB> have fun trying to figure that out
[17:59:30 CET] <JEEB> (from within the decoder)
[17:59:50 CET] <JEEB> also you will only get warnings/errors until you get your RAP, then it stops :P
[18:00:03 CET] <kepstin> you could patch ffplay to handle it better, but honestly who cares, it's just a demo player?
[18:00:03 CET] <JEEB> also if you really want to minimize that stuff, just skip the darn packets in your own API client
[18:00:27 CET] <JEEB> everything could be more perfect but is this really your main issue right now?
[18:01:00 CET] <faLUCE> JEEB: I'm not deprecating ffplay, I just noted it could be misleading
[18:01:19 CET] <JEEB> I don't even know what you mean with that, and quite honestly I don't care.
[18:01:43 CET] <faLUCE> JEEB: I care, instead. Because when I see red color msgs I think about an error.
[18:02:09 CET] <faLUCE> JEEB: I don't even know what you care, and quite honestly I don't care of what you care
[18:02:16 CET] <JEEB> please read my message some lines upwards where I note how non-simple it is to add a global state into the decoder
[18:02:29 CET] <JEEB> also many people utilize the functionality of starting to decode from non-RAP
[18:02:46 CET] <faLUCE> JEEB: I never said it's or it would be easy.
[18:02:58 CET] <JEEB> yes, now compare how much good you get out of it
[18:03:01 CET] <JEEB> compared to the effort
[18:03:23 CET] <JEEB> which is why I said "everything could be more perfect, but is this really your main issue right now?"
[18:03:24 CET] <faLUCE> JEEB: again, I'm not saying that ffplay should be patched for that
[18:03:34 CET] <kepstin> the individual pieces of ffmpeg are working correctly here - demuxer is returning packets from the start of the file, decoder is warning that it can't properly decode the first few packets sent into it.
[18:03:40 CET] <faLUCE> so, this discussion is pointless
[18:03:45 CET] <JEEB> thank you
[18:04:38 CET] <faLUCE> I just said it's confusing. then, what should or should not patched is not my decision, of course.
[18:05:12 CET] <faLUCE> and JEEB, read again: "[18:02] <faLUCE> JEEB: I don't even know what you care, and quite honestly I don't care of what you care"
[18:19:45 CET] <dannysantos> Hi people. I'm trying to do `ffmpeg -f alsa -ac 2 -i hw:0,0 output.wav` on my Manjaro but I get "[alsa @ 0x55e9429f9e80] cannot open audio device hw:0,0 (No such file or directory)"... please help
[18:21:20 CET] <faLUCE> dannysantos:  what is hw:0,0 ? an usb mic ?
[18:25:07 CET] <dannysantos> faLUCE, I don't know. my cat /proc/asound/cards gives me -> https://hastebin.com/egovelumuq
[18:25:33 CET] <faLUCE> dannysantos: where are you recording audio? on a microphone, usb camera or what??
[18:26:10 CET] <faLUCE> internal mic?
[18:26:25 CET] <dannysantos> right now I want to record the sound that I get from the speakers
[18:26:33 CET] <dannysantos> faLUCE,
[18:27:02 CET] <faLUCE> dannysantos: ok, but record with WHAT? the internal microphone?
[18:27:09 CET] <faLUCE> an usb camera? an usb mic?
[18:27:50 CET] <dannysantos> it is not the usb camera or usb headset microphone. I believe its the internal microphone then
[18:28:05 CET] <faLUCE> dannysantos: try  ffmpeg -f alsa -ac 2 -i default output.wav
[18:28:16 CET] <dannysantos> ok
[18:28:55 CET] <dannysantos> [alsa @ 0x564340037e80] cannot open audio device default (No such file or directory) default: Input/output error
[18:29:22 CET] <dannysantos> faLUCE
[18:30:50 CET] <c_14> dannysantos: check the output of arecord -L
[18:31:33 CET] <dannysantos> faLUCE c_14, https://hastebin.com/ohesegefam
[18:33:40 CET] <c_14> try using one of those devices (probably any of the sysdefault:CARD= devices), or because it appears you have pulse installed use the pulse input instead
[18:33:51 CET] <dannysantos> I want to get the interface of the sound that I get on my speakers to use it on my icecast stream to my girlfriend
[18:34:04 CET] <faLUCE> c_14: I think it's a jacl problem
[18:34:07 CET] <faLUCE> jack
[18:34:59 CET] <dannysantos> I dont have jack hardware, I only use usb
[18:35:07 CET] <faLUCE> but you have jack installed
[18:35:17 CET] <dannysantos> I have a usb camera and a usb headset that has a microphone
[18:35:36 CET] <faLUCE> jack is an audio server
[18:36:06 CET] <faLUCE> dannysantos: well, it's not a ffmpeg issue. You will have the same issue with skype, for example
[18:36:11 CET] <dannysantos> alright, I'm really new at this and mostly ignorant at this. what is your suggestion?
[18:36:45 CET] <faLUCE> dannysantos: wait, I start jack
[18:36:45 CET] <dannysantos> faLUCE, skype works well on my PC. already had tried it in the past
[18:36:50 CET] <c_14> So you don't want to record your mike, but rather the sound output of a program?
[18:36:58 CET] <dannysantos> yes
[18:37:22 CET] <dannysantos> the sound of the system in general
[18:38:05 CET] <dannysantos> for example, I have firefox and listening to a youtube music video, I want to stream that to my girl, but dont know which interface to use
[18:38:11 CET] <c_14> https://trac.ffmpeg.org/wiki/Capture/ALSA#Recordaudiofromanapplication
[18:39:54 CET] <dannysantos> c_14, will try that
[18:40:00 CET] <c_14> Or with pulse you could use the pulseaudio ui to create a monitor device for the audio output and record that with ffmpeg
[18:44:17 CET] <faLUCE> well, dannysantos,
[18:44:38 CET] <faLUCE> ffmpeg -f pulse -ac 2 -i default output.wav
[18:45:52 CET] <faLUCE> see if it works
[18:46:04 CET] <dannysantos> faLUCE, default: No such process
[18:46:36 CET] <faLUCE> dannysantos: how did you write the command?
[18:46:50 CET] <dannysantos> ffmpeg -f pulse -ac 2 -i default output.wav
[18:48:06 CET] <dannysantos> c_14, I dont have a pulseaudio ui on my Manjaro Deepin.
[18:48:27 CET] <faLUCE> dannysantos: install pavucontrol
[18:48:40 CET] <dannysantos> ok
[18:49:30 CET] <dannysantos> Im running the pavucontrol ui...
[18:50:13 CET] <faLUCE> dannysantos: at the voice "input devices" which devices are listed ?
[18:51:39 CET] <dannysantos> ok there are some, please wait
[18:53:17 CET] <dannysantos> CM108 Audio Controller Mono Analogic | Webcam C270 Mono Analogic | Front Microphone (unplugged) | Analog Input
[18:54:23 CET] <dannysantos> Im seeing "All except monitors"
[18:54:43 CET] <dannysantos> should I change this?
[18:56:41 CET] <faLUCE> dannysantos: basically, you have to follow that:
[18:56:43 CET] <faLUCE> https://askubuntu.com/questions/682144/capturing-only-desktop-audio-with-ffmpeg
[18:56:56 CET] <faLUCE> (using pavucontrol and ffmpeg)
[18:58:03 CET] <dannysantos> here is my pacucontrol screenshot -> https://ibb.co/dtdG0s7
[18:58:44 CET] <faLUCE> dannysantos: but you firstly said that you wanted to record from your mic, then you said you wanted to record from some other output
[18:58:56 CET] <faLUCE> so, what do you want to record?
[19:02:33 CET] <dannysantos> never said I want to record from my mic, i only said that I didnt want from my usb headset or usb camera so i believe it was the one suggestion left that you gave me: internal mic
[19:02:54 CET] <faLUCE> in any case: 1) if you want to record from your mic:    ffmpeg -f alsa -i pulse out.wav   .   2) if you want to record the output of a program follow the link I pasted
[19:03:21 CET] <dannysantos> I want to record the sound that I get on my speakers, not from my mics
[19:03:40 CET] <dannysantos> I believe my system doesnt have pulseaudio
[19:03:47 CET] <dannysantos> is that possible?
[19:04:32 CET] <faLUCE> dannysantos: yes, it has. Then just follow the link I pasted and replace  "ffmpeg -f pulse -i default output.wav" with "ffmpeg -f alsa -i pulse output.wav"
[19:04:49 CET] <faLUCE> (depending on the ffmpeg version)
[19:07:24 CET] <dannysantos> `ffmpeg -f pulse -i default output.wav` -> "default: no such process" | `ffmpeg -f alsa -i default output.wav` -> https://hastebin.com/uduyujidak.bash
[19:07:35 CET] <dannysantos> faLUCE
[19:10:34 CET] <faLUCE> dannysantos: is this the output of "ffmpeg -f alsa -i default output.wav"
[19:10:45 CET] <dannysantos> yes faLUCE
[19:11:13 CET] <faLUCE> dannysantos:
[19:11:15 CET] <faLUCE> [19:04] <faLUCE> dannysantos: yes, it has. Then just follow the link I pasted and replace  "ffmpeg -f pulse -i default output.wav" with "ffmpeg -f alsa -i pulse output.wav"
[19:11:16 CET] <faLUCE> [19:04] <faLUCE> (depending on the ffmpeg version)
[19:11:29 CET] <dannysantos> sorry
[19:11:32 CET] <faLUCE> the right command should be:
[19:11:39 CET] <faLUCE>  "ffmpeg -f alsa -i pulse output.wav"
[19:12:34 CET] <dannysantos> faLUCE, https://hastebin.com/apeninucub.bash
[19:14:29 CET] <faLUCE> dannysantos: did you follow the sequence in the link? 1) open pavucontrol  2) start ffmpeg
[19:15:31 CET] <dannysantos> I open pavucontrol, then when running ffmpeg I get a ffmpeg error crash, that I have shown you, so I cant proceed
[19:16:40 CET] <dannysantos> on pavucontrol tab Record I get "no application is currently recording audio" message
[19:16:59 CET] <faLUCE> it seems a permission problem, wait
[19:17:03 CET] <dannysantos> and it is showing "Applications"
[19:17:48 CET] <faLUCE> dannysantos: what if you run "sudo ffmpeg -f alsa -i pulse output.wav"  ?
[19:18:18 CET] <dannysantos> it seems to be working
[19:18:23 CET] <dannysantos> with sudo
[19:18:42 CET] <faLUCE> ok, please confirm that all works and I tell you how to fix permanently
[19:22:46 CET] <dannysantos> ok faLUCE, following the askubuntu.com link I got it working ;)
[19:22:59 CET] <faLUCE> ok, now:
[19:23:53 CET] <faLUCE> "cat /etc/group  | grep audio"
[19:24:09 CET] <faLUCE> paste the output
[19:24:21 CET] <dannysantos> audio:x:92:fernando
[19:24:27 CET] <dannysantos> my current user is daniel
[19:24:37 CET] <dannysantos> there is only fernando there
[19:24:50 CET] <faLUCE> sudo gpasswd -a daniel audio
[19:25:06 CET] <dannysantos> alright
[19:25:18 CET] <dannysantos> is it everything?
[19:25:24 CET] <faLUCE> logout and login again and see if it works without sudo
[19:25:31 CET] <dannysantos> ok
[19:28:52 CET] Last message repeated 1 time(s).
[19:28:52 CET] <dannysantos> I have issues faLUCE
[19:28:56 CET] <dannysantos> this is my error
[19:29:32 CET] <dannysantos> https://hastebin.com/yopaqoqiki.bash
[19:29:45 CET] <faLUCE> dannysantos: sudo gpasswd -a pulse audio
[19:30:08 CET] <faLUCE> then paste the output of "cat /etc/group | grep audio"
[19:30:21 CET] <dannysantos> "the user pulse does not exist"
[19:30:32 CET] <faLUCE> paste the output of "cat /etc/group | grep audio"
[19:31:08 CET] <dannysantos> audio:x:92:fernando,daniel
[19:31:23 CET] <dannysantos> `whoami` -> daniel
[19:31:40 CET] <dannysantos> should I reboot just to be sure?
[19:32:25 CET] <faLUCE> dannysantos: "cat /etc/group | grep pulse"
[19:32:42 CET] <dannysantos> there is no output
[19:33:19 CET] <faLUCE> then something is broken with your installation. What if you login as "fernando" ?
[19:33:40 CET] <dannysantos> that is my dad account
[19:33:49 CET] <FlipFlops2001> Why is the "acompressor" threshold parameter expressed in DECIMALS rather than DECIBELS(dB) like every other hardware or software compressor/limiter in the world?
[19:34:01 CET] <dannysantos> whould I su as fernando on the command line and test it?
[19:34:12 CET] <faLUCE> dannysantos: yes
[19:34:36 CET] <faLUCE> but, in any case, it should grep the pulse user
[19:34:44 CET] <faLUCE> wait
[19:35:31 CET] <dannysantos> ALSA lib pulse.c:243:(pulse_connect) PulseAudio: Unable to connect: Connection refused
[19:35:31 CET] <dannysantos> [alsa @ 0x55b2f950ee40] cannot open audio device pulse (Connection refused)
[19:35:31 CET] <dannysantos> pulse: Input/output error
[19:35:43 CET] <faLUCE> dannysantos: ok:    "adduser daniel pulse"
[19:36:06 CET] <faLUCE> then "adduser daniel pulse-access"
[19:36:17 CET] <dannysantos> what will that command do?
[19:36:23 CET] <dannysantos> create a pulse user?
[19:36:45 CET] <faLUCE> no, they will add daniel to the group "pulse"
[19:36:51 CET] <dannysantos> ok
[19:36:52 CET] <faLUCE> and to the group "pulse-access"
[19:36:57 CET] <dannysantos> ok
[19:37:24 CET] <dannysantos> but I dont have a pulse group right now
[19:37:27 CET] <faLUCE> then logout/login
[19:39:06 CET] <dannysantos> cat /etc/group | grep -i pulse -> no output
[19:39:15 CET] <faLUCE> that's pretty weird... we could create this group but it should have been already created with your OS
[19:39:35 CET] <faLUCE> what is your distro?
[19:39:45 CET] <dannysantos> I believe I will use sudo on ffmpeg
[19:39:53 CET] <faLUCE> no, don't mess things with sudo
[19:40:02 CET] <faLUCE> what is your distro?
[19:40:14 CET] <dannysantos> Manjaro Deepin x86_64
[19:41:33 CET] <dannysantos> so, you believe the right action is to create the pulse group and add daniel user to that group
[19:41:48 CET] <dannysantos> ?
[19:42:04 CET] <faLUCE> no, I don't believe that. it should already be created with the installation of your OS/software
[19:42:20 CET] <dannysantos> faLUCE, ok
[19:45:19 CET] <faLUCE> dannysantos: let's try by addind the groups. If it won't work, we will remove them:
[19:45:23 CET] <faLUCE> groupadd -g 58 pulse
[19:45:37 CET] <faLUCE> groupadd -g 59 pulse-access
[19:46:06 CET] <faLUCE> usermod -a -G audio pulse
[19:46:31 CET] <dannysantos> the user pulse doesnt exist
[19:46:51 CET] <dannysantos> maybe daniel?
[19:47:15 CET] <faLUCE> useradd -c "Pulseaudio User" -d /var/run/pulse -g pulse -s /bin/false -u 58 pulse
[19:47:52 CET] <dannysantos> ok, now it works the last command
[19:48:11 CET] <dannysantos> the usermod command
[19:48:20 CET] <faLUCE> (then repeat ok, logout/login, then grep "cat /etc/group | grep pulse"
[19:50:14 CET] <dannysantos> audio:x:92:fernando,daniel,pulse
[19:50:14 CET] <dannysantos> pulse:x:58:
[19:50:14 CET] <dannysantos> pulse-access:x:59:
[19:50:25 CET] <faLUCE> ok, try the ffmpeg command
[19:51:42 CET] <dannysantos> error
[19:51:45 CET] <dannysantos> https://hastebin.com/qamumahoru.bash
[19:52:46 CET] <faLUCE> dannysantos: does it still work with sudo ?
[19:52:51 CET] <dannysantos> faLUCE, maybe daniel should be on pulse group instead of audio?
[19:53:05 CET] <faLUCE> dannysantos: no...
[19:53:21 CET] <dannysantos> yes, it doesnt give error with sudo
[19:53:25 CET] <dannysantos> faLUCE,
[19:54:08 CET] <faLUCE> dannysantos: how did you install the audio?
[19:54:19 CET] <faLUCE> stuff?
[19:54:34 CET] <dannysantos> I havent messed with any audio, just installed the distro
[19:54:49 CET] <dannysantos> and am using what it came with
[19:55:22 CET] <dannysantos> maybe on pavucontrol some interface maybe blocked
[19:55:25 CET] <dannysantos> no?
[19:55:46 CET] <faLUCE> dannysantos: it seems a bug of the distro
[19:55:50 CET] <faLUCE> wait
[19:56:45 CET] <dannysantos> I mean locked
[19:57:23 CET] <faLUCE> dannysantos: did you install your distro for some real time audio stuff?
[19:58:33 CET] <dannysantos> no
[19:58:39 CET] <dannysantos> faLUCE,
[20:00:58 CET] <faLUCE> dannysantos: then use sudo meanwhile, and send a bug/missing feature report to your distro
[20:01:42 CET] <friendofafriend> Hello everyone, I'm getting "Invalid UE golomb code" when playing an MPEG-TS stream created by ffmpeg.  Is there something I should add to my ffmpeg encoder's command line to fix it?
[20:02:36 CET] <dannysantos> hmmm.... bugs.manjaro.org stop existing and I dont know what to write in this bug report...
[20:02:51 CET] <dannysantos> if I use the manjaro forum
[20:03:44 CET] <faLUCE> dannysantos: then use a better distro :-)
[20:04:15 CET] <dannysantos> loool
[20:04:20 CET] <dannysantos> ok
[20:04:44 CET] <dannysantos> I'm happy with it right now faLUCE . Thanks man for all your help
[20:05:22 CET] <faLUCE> dannysantos: yw
[20:05:24 CET] <faLUCE> ;-)
[20:05:29 CET] <dannysantos> :D
[20:31:14 CET] <dannysantos> faLUCE, everything is working but my interface sound volume is too low. do you know a way to increase it?
[20:31:51 CET] <dannysantos> i'm not using ffmpeg now, i'm using ices to create an icecast stream...
[20:33:30 CET] <dannysantos> maybe some hack to increase the pulse interface volume...
[20:35:20 CET] <faLUCE> dannysantos: you have the volume control il pavucontrol
[20:35:40 CET] <dannysantos> yes, it still is too low. i have it on max
[20:36:31 CET] <faLUCE> dannysantos: see also the volumes by executing  "alsamixer"
[20:38:25 CET] <dannysantos> faLUCE, the capture is on max
[20:38:35 CET] <tdr> use alsamixer then press f6 to get the right card/device
[20:38:39 CET] <dannysantos> of pulseaudio.... need more
[20:39:00 CET] <tdr> pulse will still be very quiet is the lower-level alsa is turned down
[20:40:26 CET] <dannysantos> ok, which card should increase? loopback?
[20:40:54 CET] <tdr> it depends what your options are.  check each of them
[20:41:10 CET] <tdr> probably isn't loopback, it would be your actual hard/device that you're getting input from
[20:41:22 CET] <tdr> s/hard/card
[20:42:18 CET] <dannysantos> there are so many that im afraid of messing with this
[20:42:36 CET] <tdr> you mean in the card list when you run alsamixer and press f6 ?
[20:42:39 CET] <dannysantos> there are some usb ones, i know that those arent
[20:42:55 CET] <tdr> what kind of device is it you're using for input?
[20:43:58 CET] <dannysantos> im not using microphones, im using software apps for the input Im outputing sound that came from my pc programs
[20:45:21 CET] <tdr> ok so it would either be output channels (do the programs sound super quiet when you're not capturing?) or the levels for the input channels if you are using loopback or similar.
[20:45:39 CET] <tdr> there's 1000 ways to do sound on linux, so its hard to say "it's right there" for sure
[20:46:00 CET] <tdr> (just alsa, pulse on alsa, oss, jack ... so many ways to mix layers)
[20:46:45 CET] <alexpigment> hey guys, it's been a long time since i've been around, but i have a question that hopefully someone knows the answer to
[20:47:11 CET] <alexpigment> i've got a video from a DVD source that I believe came from a VHS source but the field order got messed up in the process
[20:47:18 CET] <alexpigment> so the interlacing lines are jumping up and down
[20:47:29 CET] <alexpigment> anyone have any idea how to fix this type of issue?
[20:47:42 CET] <dannysantos> tdr, the program sound on my pc are not super quiet, only on my icecast stream
[20:48:21 CET] <dannysantos> on my iceast stream im using "pulse" as a input device
[20:48:24 CET] <tdr> dannysantos, ok so if it set up some loopback type virtual devices, check the levels on that/them
[20:50:24 CET] <dannysantos> tdr, on pavucontrol im using the "monitor of cm108 audio controller digital stereo"
[20:50:47 CET] <dannysantos> but on alsamixer I cant find that card/device
[20:57:37 CET] <dannysantos> tdr, on alsamixer I should be looking for the capture section?
[20:58:13 CET] <tdr> it could be "capture" volumes down, sure
[20:58:31 CET] <tdr> it could be just pcm channels on some device it setup, i dont know the internals of how your software cofigures thing
[21:04:56 CET] <dannysantos> on pavucontrol i have "ALSA plug-in [ices]" . Can it be this plugin fault?
[21:05:10 CET] <dannysantos> tdr
[21:06:06 CET] <tdr> that makes me think it's setting up some "virtual" channels and should have their own volume controls in pulse and alsa when you look inside alsamixer
[21:07:56 CET] <dannysantos> i dont seem to be finding that virtual channels. im searching on the "select sound card"
[21:08:06 CET] <dannysantos> in alsamixer
[21:11:06 CET] <dannysantos> the only one that I'm finding that is not the usual is Loopback and maybe that is because I have modprobe snd-aloop pcm_substreams=1 following other tutorial
[21:12:22 CET] <Matti-> " legacy WAVE file has format type 1 but bits-per-sample=24"    i get this warning:   what is  "format type 1" mean
[21:19:29 CET] <dannysantos> ok I got it tdr
[21:20:10 CET] <dannysantos> thank you for the patience, I had to increase the master and decrease my headset audio output
[21:29:07 CET] <tdr> nice, glad its working, good job
[21:39:44 CET] <Matti-> what would sound better to human ears?   24bit 48khz wav file or   16 bit 96khz  wav file?
[21:46:18 CET] <another> define better
[21:47:13 CET] <Matti-> another closer to true analog sound
[21:48:45 CET] <another> chances are you can't hear a difference
[21:54:30 CET] <alexpigment> a lot of things are in 24bit 48khz in the professional world
[21:54:51 CET] <alexpigment> but that's really just because there's more headroom for mixing
[21:55:03 CET] <Matti-> alexxpigment i am sure professionals use 24bit 96khz
[21:55:08 CET] <alexpigment> of course
[21:55:14 CET] <alexpigment> but they also use 24/48
[21:55:21 CET] <alexpigment> no one uses 16/96 in the professional world
[21:55:40 CET] <Matti-> why would they record in 24/48 if they can record in 24/96
[21:55:49 CET] <alexpigment> at any rate, 24/48 probably also sounds better too, although it would be difficult to hear the difference
[21:56:03 CET] <alexpigment> Matti - don't ask me. i see it in professional contexts
[21:56:21 CET] <alexpigment> but i would *guess* that it's because certain hardware doesn't do 24/96
[21:58:00 CET] <alexpigment> could also be that broadcast/video mastering will ultimately be 48khz anyway so there's no need to waste the extra bits
[21:58:27 CET] <alexpigment> for purely audio - there's a lot of people that argue it's better to record at 44.1 or 88.2 rather than 48 or 96
[21:58:41 CET] <Matti-> alexpigment really? i never heard of that
[21:58:46 CET] <Matti-> why do they say that
[21:58:52 CET] <alexpigment> yeah, it goes back to CD production
[21:59:07 CET] <alexpigment> 88.2 > 44.1 is much cleaner than 96 > 44.1
[21:59:20 CET] <Matti-> alex i see; that makes sense
[21:59:33 CET] <another> if you can hear the resampler, get a better resampler
[21:59:50 CET] <alexpigment> right, but resampling at non-integer values is always imperfect, technically speaking
[22:00:00 CET] <alexpigment> better to know your final medium and choose accordingly
[22:01:07 CET] <Matti-> 88.2 > 44.1 would be cleaner, that makes total sense
[22:01:17 CET] <alexpigment> i'll be the first to admit though that this is all academic cork-sniffing stuff. a lot of stuff is bounced from 24/96 to 16/44.1 and it sounds great
[22:02:05 CET] <alexpigment> the bit depth is really just for editing though. 24-bit is better than 16-bit, but if you're never going to change the volume of the track after recording, it doesn't matter
[22:02:11 CET] <Matti-> not sure why dvd-video started using 48khz
[22:02:24 CET] <alexpigment> i feel like i knew at some point, but i can't recall anymore
[22:02:39 CET] <another> you sniff corks? ^^
[22:04:31 CET] <another> afaik 48kHz is the recommended sampling rate when you do video
[22:04:33 CET] <alexpigment> haha. no, i'm not a cork sniffer. there are some quality things i care about (proper interlacing/deinterlacing, maintaining original frame rates, keeping high bitrates, etc) but i consider myself to have pretty discerning ears and I have no problem with a 320kbps MP3, for example
[22:04:46 CET] <another> i think aes recommended that
[22:04:57 CET] <alexpigment> right, 48khz for video unless you're doing some sort of blu-ray audio thing
[22:05:26 CET] <alexpigment> specialized blu-ray audio to be clear. 48khz is still the normal standard there
[22:08:47 CET] <another> alexpigment: no wine for you then ;)
[22:08:52 CET] <alexpigment> haha
[22:09:11 CET] <alexpigment> i'm a beer guy anyway. my wife is the wine drinker ;)
[22:12:11 CET] <dingwat> fwiw 24/48k (rather than 24/96k) is better when bandwidth is an issue. I work with systems that transport huge channel counts over standard layer 3 networking layers, and doubling the bandwidth would cripple the channel count needlessly
[22:13:48 CET] <dingwat> also 48k is a lot nicer relative to 44.1k from a filter design perspective, it's a lot harder to design a well-behaved filter that that's flat out to 20k and then drops 60 or 80dB at 22.05kHz versus 24kHz
[22:13:51 CET] <dingwat> just my two cents
[22:15:18 CET] <alexpigment> dingwat: for the original context, Matti was asking about using 24/48 vs 16/96
[22:15:35 CET] <dingwat> I'd prefer 24/48k over 16/96k, 48k is plenty enough sample rate unless you're timestretching heavily, whereas the extra dynamic range of 24 vs 16 is significant and comes in handy in a lot of contexts
[22:15:41 CET] <alexpigment> i don't anyone who uses 16/96 in a pro context
[22:16:07 CET] <dingwat> I don't either. 16 bit converters aren't too common these days
[22:16:24 CET] <dingwat> the audiophile folks are pushing 32bit converters, in fact...
[22:16:35 CET] <dingwat> and the converter manufacturers, I guess..
[22:16:54 CET] <Matti-> then why does dsd uses 1bit 2.8mhz ?  which suppose to sound better than pcm 24bit192khz
[22:16:55 CET] <alexpigment> yeah, even basic software like audacity defaults to 32-bit (as well as audition, which i use). a lot of data, but no reason not to use it
[22:17:07 CET] <alexpigment> dsd is a weird beast
[22:17:14 CET] <alexpigment> SACD i assume you're talking about
[22:17:17 CET] <dingwat> AKMs latest chipsets do have very impressive performance, but you have to have very good pres and the rest of your signal chain has to be bombproof for it to matter
[22:17:22 CET] <Matti-> alexpigment yes
[22:17:33 CET] <alexpigment> it doesn't relate to PCM at all, and i have no idea why it sounds as good as it does
[22:17:36 CET] <dingwat> Matti-: you can't really compare DSD to PCM directly, they're different beasts
[22:17:51 CET] <Matti-> dsd use massive mhz sampling rate
[22:17:59 CET] <Matti-> but only 1 bit
[22:18:13 CET] <dingwat> DSD is essentially the direct output of a delta sigma converter, and is great when you are driving a delta sigma converter
[22:19:05 CET] <alexpigment> on the other hand, modern PCM formats at high bit depths and sample rates should sound better
[22:19:17 CET] <alexpigment> i think SACD is only advantageous compared to CD redbook audio
[22:19:46 CET] <alexpigment> but it is a standardized format with an established catalog, so that's an inherent benefit
[22:19:50 CET] <dingwat> Matti-: for further reading, start with the wiki page on delta sigma: https://en.wikipedia.org/wiki/Delta-sigma_modulation
[22:23:07 CET] <dingwat> it's also worth pointing out that modern DS converters are operating at very high orders with extreme noise shaping, and most of the reading material focuses on first or second order DS architectures, which is fine, but in general, "real" audio converter design is a sort of black magic and there's not a lot of folks out there who can design the stuff that AKM/Cirrus/others are making
[22:23:29 CET] <dingwat> It's not difficult to build your own first, second, maybe third order DS
[22:24:01 CET] <dingwat> But at the higher orders it's very difficult to get them to be stable and well-behaved (or so I'm told, I'm not good enough at math to mess with that stuff)
[22:25:57 CET] Action: alexpigment understands nothing about what dingwat just said, but appreciates the high level of knowledge
[22:26:01 CET] <alexpigment> ;)
[22:26:14 CET] <Matti-> alexpigment me neither
[22:26:37 CET] <dingwat> Sorry, I'm an audio nerd and this shit is cool :)
[22:27:01 CET] <Matti-> dingwat do you think dsd is superior to pcm
[22:29:50 CET] <dingwat> No, I don't think either one is superior. In general, PCM is a more versatile format, but DSD (more generally, pulse-density) is an attempt to eliminate certain parts of the signal chain, with inconclusive results
[22:30:28 CET] <alexpigment> i agree about the inconclusive results :)
[22:31:13 CET] <dingwat> Matti-: IIRC, 24/48k PCM theoretically has more DR than DSD, and it's also much more widely supported, so IMO PCM should be used unless there's a very specific reason not to
[22:31:33 CET] <alexpigment> having said that, if you were debating on releasing a "definitive" version of something, i'd do blu-ray audio rather than SACD
[22:31:48 CET] <Matti-> dingwat DSD much much higher sample rate
[22:31:54 CET] <Matti-> dsd has*
[22:32:12 CET] <alexpigment> a lot of that sampling rate is wasted on making up for the 1-bit depth though
[22:32:22 CET] <dingwat> yes but it's not the same "kind" of sample rate
[22:32:34 CET] <dingwat> You cannot compare them directly
[22:32:39 CET] <Matti-> what does bluray-audio use? PCM?
[22:32:47 CET] <alexpigment> several formats, PCM is one of them
[22:32:53 CET] <alexpigment> you can do 24/192 even
[22:32:53 CET] <Matti-> what is other format
[22:32:59 CET] <alexpigment> although 24/92 is more common
[22:33:17 CET] <alexpigment> DTSMA, DolbyHD
[22:33:24 CET] <alexpigment> or whatever dolby calls their "lossless" format
[22:33:29 CET] <alexpigment> truehd
[22:33:54 CET] <alexpigment> of course there's regular dolby digital, but that's not really worth mentioning for high quality audio
[22:33:56 CET] <Matti-> then what is the difference between bvluray audio vs  dvd-audio?
[22:34:01 CET] <Matti-> sounds very similar
[22:34:12 CET] <alexpigment> sampling rates mostly i think
[22:34:20 CET] <alexpigment> but you run out of space quickly at 24/96
[22:36:26 CET] <alexpigment> i was trying to do the bitrate calculation - i think 5.1 24/96 PCM is like 13mbps or so (i may have done that wrong) but it also reminded me that DVDs have a bitrate maximum
[22:36:53 CET] <alexpigment> 10mbps total stream bitrate, but realistically much lower for universal compatibility
[22:36:58 CET] <alexpigment> 10.8
[22:42:02 CET] <another> wiki says max 9,8 MBit/s
[22:42:07 CET] <Matti-> "FLAC supports linear PCM samples with a resolution between 4 and 32 bits per sample. FLAC does not support floating point samples"      what is  "floating point samples" mean
[22:42:31 CET] <another> 5.1 24/96 should come to ~13,8 MBit/s
[22:42:47 CET] <alexpigment> floating point usually means decimal vs integer
[22:43:29 CET] <Matti-> but what does that mean in real life
[22:44:15 CET] <alexpigment> for normal humans? nothing that i'm aware of
[22:44:25 CET] <alexpigment> dingwat probably has an answer though
[22:45:35 CET] <alexpigment> another: that's for video. there's a total data rate though that is 10.8 i think
[22:45:39 CET] <alexpigment> or maybe 10.08?
[22:45:47 CET] <alexpigment> i don't remember. it's been a while since i worked with dvd in a professional context
[22:56:37 CET] <another> i think i got that number from the dvd-audio article
[22:57:34 CET] <Matti-> how comes  ffmpeg -i 24bit.flac 24bit.wav  converts  24bit.flac to 24bit.wav fine  but   flac binary does not
[22:57:57 CET] <Matti-> how is ffmpeg does better than flac binary
[22:59:04 CET] <alexpigment> another: ah. i suppose it's probably maximum bitrate for a single stream then (video or audio)
[23:05:22 CET] <Matti-> how comes  ffmpeg.exe -i 24bit.flac 24bit.wav  converts  24bit.flac to 24bit.wav fine  but   flac.exe -d  24bit.flac 24bit.wav  does NOT
[23:11:08 CET] <Matti-> sorry i made a mistake
[23:11:26 CET] <Matti-> ffmpeg.exe -i 24bit.flac 24bit.wav  is not working  it's creating  16bit.wav
[23:12:44 CET] <Matti-> ffmpeg.exe -i 24bit.flac 24bit.wav  is not working  it's creating  16bit.wav:   why is that?
[23:15:28 CET] <another> because you didn't specify the format. therefore ffmpeg guesses it from the extension of the output file
[23:15:40 CET] <Matti-> another how do i do that
[23:16:37 CET] <alexpigment> -sample_fmt s24 i would assume
[23:17:15 CET] <alexpigment> although i would *think* it should pass it along
[23:17:18 CET] <another> -c:a pcm_s24le
[23:17:37 CET] <alexpigment> makes sense. if you're outputting pcm_s16le, it's going to force 16-bit
[23:17:59 CET] <Matti-> what is the difference between -c:a pcm_s24le and -sample_fmt s24 then
[23:18:06 CET] <alexpigment> does ffmpeg assume s16le for .wav?
[23:18:45 CET] <another> yes
[23:18:47 CET] <alexpigment> Matti: well, that's the generic way to force a bit depth, but with PCM in ffmpeg, you generally have to use the correct c:v to get the right bit depth anyway
[23:18:59 CET] <another> alexpigment: ffmpeg -h muxer=wav
[23:19:05 CET] <alexpigment> ah
[23:19:23 CET] <Matti-> what does  ffmpeg -h muxer=wav  do?
[23:19:28 CET] <alexpigment> i kinda just assumed it would auto choose a little endian format based on the input depth
[23:20:04 CET] <alexpigment> -h is help and you're getting the specific help related to the wav muxer
[23:20:23 CET] <Matti-> oh okay
[23:20:37 CET] <alexpigment> it says "default audio codec: pcm_s16le"
[23:21:00 CET] <Matti-> -c:a pcm_s24le  works,  but why isn't ffmpeg smart enough to know that  flac is 24bit and use  -c:a pcm_s24le  automatically
[23:21:32 CET] <another> i guess because wav is not a codec but a container
[23:21:42 CET] <alexpigment> well, you're converting, so it's gotta choose a codec, and it seems like it chooses that exact codec by default when the output extension is wav
[23:22:19 CET] <Matti-> another  ffmpeg is smart enough to  detect that flac is using 88.2 khz  and convert automatically to  88.2 khz wav
[23:22:38 CET] <alexpigment> the sampling rate is separate from the codec fwiw
[23:23:23 CET] <alexpigment> although i do hear what you're saying. it seems like it could detect all of the above and select the appropriate codec
[23:23:35 CET] <alexpigment> equivalent codec, i mean
[23:35:14 CET] <Dodutils> hi guys, I am trying to use ffmpeg as some kind of double RTSP flow redirector (-c copy mode no re-encoding), I need ffmpeg to read an RTSP H.264 source and do direct-to-disk recording (easy part) but I also want it to be used as some kinf of proxy for this same stream to be used by some other RTSP client (not FFMPEG to restream but act as server) but FFServer is dead project so anyone have some idea about how to do it ?
[23:38:34 CET] <Dodutils> the final goal of this is to have some "lossless" video quality (original quality from RTSP source) but also let something like OpenCV read same stream to detect movements and write timestamps to a text file so I can then cut the saved file stream to keep only parts with movements, but I do not want camera to provide two streams at same time (bandwidth and camera ressources limitation).
[23:45:37 CET] <Dodutils> OpenCV is an example but could be a videosurveillance software too this is why I need this direct-to-disk + "proxy/server" approach
[23:47:22 CET] <faLUCE> Dodutils: then you want a rtsp server on demand, right?
[23:48:00 CET] <tdr> Dodutils, "motion" software already does what you're suggestion
[23:48:05 CET] <Dodutils> and second question is there a way to monitor that ffmpeg is not frozen ? I mean the camera tream may go wrong and ffmpeg will stop to read the stream but will not quit, my idea is to check for the direct-to-disk file last touch and if no touch for a few seconds then kill/restart ffmpeg.
[23:49:01 CET] <tdr> Dodutils, https://motion-project.github.io
[23:49:24 CET] <Dodutils> motion can do direct-to-disk ?
[23:49:32 CET] <faLUCE> Dodutils: too many things, let's proceed by order
[23:49:54 CET] <faLUCE> first of all, do you want a rtsp server on deman, right?
[23:50:19 CET] <Dodutils> @faLUCE : not exactly I want ffmpeg to read one stream, save it realtime direct-to-disk but also act as RTSP server source for this same stream
[23:50:41 CET] <faLUCE> Dodutils: yes, so it will do a rtsp server on demand
[23:50:44 CET] <faLUCE> with multiple clients
[23:50:51 CET] <Dodutils> one client only
[23:54:24 CET] <faLUCE> Dodutils: wait
[23:57:34 CET] <Dodutils> @tdr : you are correct I read "Passthrough recording from many IP cameras" never checked that feature but I do not want to use Motion as motion detection software unless Motion can also "cut" the passthrough file to keep only movement parts (with 2 second pre-buffer)
[23:58:40 CET] <faLUCE> it seems that the rtsp support is in a bad state, can you guys confirm that?
[00:00:00 CET] --- Sun Mar 24 2019