Mathematica doesn't recognize USB mic (Raspberry Pi / Linux)

[Swamp Thing]

Raspberry Pi / Linux...

My USB mic shows up in the PulseAudio volume control, and also works OK with Audacity and arecord.

However, it doesn't show up in Wolfram Mathematica -- $AudioInputDevices returns a null list.

Anything I can do to get the mic working in Mathematica?

 

[jedishrfu]

Have you tried asking at the wolfram site? They are the makers of Mathematica and there may be a ticket open on it or a way to fix it.

I found this reference on another site that you can try:

https://askubuntu.com/questions/134...-input-device-in-settings-only-in-pavucontrol

or this site for Linux Mint:

https://forums.linuxmint.com/viewtopic.php?t=381489

 

[jedishrfu]

I had some issues with lapel mics on Raspberry PI Linux. The problem was due to the mics having the same USB address with no way to change them. The company didn't think someone might want to do stereo recording with their lapel mics.

In our case, we switched to another mic manufacturer. Our project was to make a headset for hearing impaired speakers who could pick out an audience member asking a question by simply looking in their direction and the two mics acted as directional mics beamforming the audio where you pointed your head.

 

[jedishrfu]

I forgot to add that we beamformed in the PI, and the one noticeable flaw was the time it took to reach your ears. It was enough to make it sound like an echo.

You'd hear the person speaking and then the beamformed audio stream of them speaking slightly behind them. This was due to the speed at which the PI could process the mic input, add them together, and then push the result to the headset.

 

[Swamp Thing]

Is it hard to get a good beam shape over an acceptable range of audio frequencies?

And re my original question, my use case works ok if I run a shell command from Wolfram, where I run "arecord". Then Mathematica can open the recorded file and get to the audio samples. But I will keep looking for a more elegant way.