Feature Request - USB Audio Support for 'Class Compliant' Devices (via ALSA?)

[Jajola]

Hey Matthew,

Firstly, thanks for this project (/product) - It's incredibly useful. We really love the martim01 as a test device in our workshop, where we use external NTP/DAD converters to route it an AES67 stream of whatever we patch to the converter.

I wondered if you could look at implementing support for Class Compliant USB Audio devices, potentially via Linux's ALSA tools!?

Personally, I have very little comprehension of how large or small a task this would be, but it feels like it would be a neat step to enabling wider support for other flavours of professional I/O - though I appreciate there could be some fundamental architectural issue as to why you haven't done it already...

As it stands, we need to swap multiple hats & configuration for balanced Analog I/O vs. S/PDIF (which we then need to convert to/from AES3), and it rules out support for formats like MADI, ADAT etc.

It also currently doesn't allow for testing mixed I/O, because I don't think you can stack Pi Audio HATs in that manner & you can't select one for out & another for in (as far as I can tell). An example use case would be utilising the Signal Generator Output to Analog, and monitoring an AES input - perhaps testing an external converter/router/DAW...

In any case - it's really awesome - thanks for making it available to the audio community!

Cheers,

JJ

[MatthewMartinBBC]

Hi JJ,
Thanks for the feedback, I'm glad you're finding the PAM useful.
For taking non-AES67 input (i.e. the PAM Soundcard setting) the PAM uses the ALSA library (via the portaudio library). I have in the past plugged an IP/ISDN audio codec in via USB and it worked fine, appearing as another input in the Soundcard setting.
I haven't tried multi-channel soundcards though or ones that support MADI, ADAT etc. Assuming Linux itself can see them via ALSA I would expect the PAM to be able to - though currently the PAM expects it's soundcard input/output to be stereo rather than any more channels.
I don't think this should be a too difficult feature to add therefore (at least for multichannel analog/aes3 devices).

I'll add it to the "roadmap".
Cheers,

Matt