USB audio on a Raspberry PI

Dashboard
1

Hi all,
I am trying to stream audio from a radio scanner connected to my pi via usb. In simple terms it is acting as a usb microphone. So far i have had no luck at all getting it to work! can anyone provide me with a working example flow/guide? The other issue I am having is my iPad Pro does not play any audio from node red dashboard. any ideas on that?

cheers

2

Hello,
Does your scanner show up under Linux like a standard microphone/audio interface under alsa? If you do a arecord -l on the command line does show up as an input device or does it stream the data in some kind of raw format over serial?

Johannes

3

Yea it does

4

If shows up with alsa you can install sox and my node-red-sox-utils nodes from github which are in beta right now to interface with it directly from nodered:


In the record node the scanner should appear in the source dropdown if it’s a valid audio interface giving you the option to record the audio and handle it in nodered.
5

Thankyou, this works! however the audio is really choppy! have you got any ideas to fix this?

my flow is attached

[{"id":"1a1e0044.8a069","type":"ui_audio","z":"a96ef2d9.1d179","name":"","group":"a29ae445.1c44c8","voice":"","always":true,"x":820,"y":60,"wires":[]},{"id":"93930129.3b614","type":"sox-record","z":"a96ef2d9.1d179","name":"","buttonStart":"msg","inputs":1,"inputSource":"2,0","byteOrder":"-L","encoding":"signed-integer","channels":"2","rate":16000,"bits":16,"gain":"0","lowpass":8000,"showDuration":false,"durationType":"forever","durationLength":0,"silenceDetection":"nothing","silenceDuration":"2.0","silenceThreshold":"2.0","outputFormat":"stream","manualPath":"","debugOutput":false,"x":470,"y":60,"wires":[["55c39df.09d1964"],[]]},{"id":"ec08b42f.b38208","type":"ui_switch","z":"a96ef2d9.1d179","name":"","label":"Web Audio","tooltip":"","group":"a29ae445.1c44c8","order":4,"width":3,"height":1,"passthru":true,"decouple":"false","topic":"","style":"","onvalue":"start","onvalueType":"str","onicon":"","oncolor":"","offvalue":"stop","offvalueType":"str","officon":"","offcolor":"","x":310,"y":60,"wires":[["93930129.3b614"]]},{"id":"3a439393.82dfcc","type":"inject","z":"a96ef2d9.1d179","name":"","props":[{"p":"payload"}],"repeat":"","crontab":"","once":false,"onceDelay":0.1,"topic":"","payload":"start","payloadType":"str","x":170,"y":40,"wires":[["ec08b42f.b38208"]]},{"id":"9b883a6d.b0f358","type":"inject","z":"a96ef2d9.1d179","name":"","props":[{"p":"payload"}],"repeat":"","crontab":"","once":false,"onceDelay":0.1,"topic":"","payload":"stop","payloadType":"str","x":170,"y":80,"wires":[["ec08b42f.b38208"]]},{"id":"55c39df.09d1964","type":"wav-headers","z":"a96ef2d9.1d179","name":"","action":"add","channels":"2","samplerate":"16000","bitwidth":16,"x":650,"y":60,"wires":[["1a1e0044.8a069"],[]]},{"id":"a29ae445.1c44c8","type":"ui_group","name":"Radio Controls","tab":"701681b5.a061b8","order":2,"disp":true,"width":6,"collapse":true},{"id":"701681b5.a061b8","type":"ui_tab","name":"Ship Spy","icon":"fa-ship","disabled":false,"hidden":true}]
6

Please put your flow inbetween ``` or use the </> button from the toolbar so that your Flow will be formatted properly and can be imported.
Are you trying to livestream the audio? How are you consuming the stream?

7

Yes the audio is streaming to my dashboard

8

I have no experience with this. The raw stream output is not really ment for listening, it’s normally fed to other software/nodes for recording or processing. Is the dashboard node you are using ment for playing streams?
Because if it’s not explicitly ment for receiving a stream of audio chunks it would play every chunk like it was an individual file which could explain the choppiness you are experiencing.

9

I'm using a node to add the wav headers. Is there a way of buffering the audio so its not choppy?

10

image
image829Ă—99 8.26 KB

I have this layout

11

As I said I think the problem is that you are streaming audio chunks and the dashboard node is playing each and every chunk individually and not treating it as a stream. So the problem is on the dashboard side.

Edit:
Here is a link to thread where exactly this issue is discussed but I think never got resolved. Read from the post I linked onwards and you will see it’s exactly the issue I expected (fun fact the node you use to add wav headers was developed because of this):

So unless somebody picks up the proposed changes by Bart and does a working pull request on the dashboard audio out node it looks like it just doesn’t support streams.

12

@dceejay did any updates get pushed to fix this?

13

Hi Johannes,
I never got it working without distortions on my Android phone. So that was for me end of story unfortunately...
Bart

1 Like
14

Damn! so its a non starter then?

15

You could try implementing those changes @BartButenaers was working on locally in your installation and see if it works for you as he was only having problems with his android phone and not his desktop if i remember correctly. So it might work good enough for you although it would be a hack of sorts.

1 Like
16

Don’t think I’ve ever seen a PR

17

As Johannes said, you can use it as a starting point if you want. I have spent - at the time being - LOTS of spare time on it, but I only got it working on browsers running at Windows but not at Android (where the sound was very bad). So useless for a pull-request.

You can easily fool your eyes by showing images fast enough, but it seems to be very difficult to fool your ears: the slightest overlap or gap will be hearable.
I was very disappointed that I didn't managed to get it working cross platform, but perhaps you have more luck. And perhaps meanwhile somebody else has developed some kind of library that you can integrate into Node-RED.
Good luck with it!

18

I think your mods are beyone my skills :frowning: im not interested in android, only IOS or windows

19

the stream doesnt need to be exact real time,could i record for 1 second then play that recording? then i have a buffer for network latency

20

Than you would have a gap in the audio every second and a latency that would build up with every audio chunk that arrives as your problem is that the way it is right now a new play process is started for every chunk that arrives and instead of a continues process that consumes the chunks there is a new one each time. So the reason for the choppiness is this time that it takes to end the old process and start a new each and every time. So sending bigger buffers wouldn’t solve the problem.