[Hi all, I sent this to the alsa-user list about 2 months ago, and haven't seen any replies. Hoping someone here might know? Thanks in advance.]

Hi all,

I'm looking at ways to try and share a sound card over a network in a lossless and low-latency manner.

The application here will be to be able to use the sound card on a Raspberry Pi to interface to an amateur radio transceiver, and use that interface to send things like slow-scan television from a more powerful workstation.

For this to work, it is utterly *essential* that no audio compression is used. So PulseAudio, JACK, etc… unless I'm very much mistaken, are not options as they use lossy audio CODECs for compression. Audio itself might only be 8kHz at 16-bit linear PCM, mono… so not terribly high bit-rate, and even latency is negotiable, but lossy compression is likely to mess things up.

PulseAudio I'd imagine is also asking for trouble where digital modes are concerned, as it'll be trying to re-sample everything to keep a consistent sample rate. I want as little extra DSP on streams as possible. I'd prefer to let PulseAudio just handle "audio" only.

I'm considering writing an ALSA PCM plug-in to accomplish this, that would expose a "virtual" sound card on my workstation, and allow me to stream it (by custom protocol, RTP, whatever) to another machine. In researching this, I've stumbled on the `shm` plug-in.

https://www.alsa-project.org/alsa-doc/alsa-lib/pcm_plugins.html#pcm_plugins_...

It apparently talks to something called `aserver`. `man aserver` gives me nothing. Google search just takes me to its source code. There seems to be scant information about what this does, how to configure it, etc.

Alternatively, it'd be nice to know how `aserver` and the `shm` plug-in communicate, maybe I can write my own client daemon that acts like `aserver` and looks after streaming to the desired end-point.

Alternatively, if people feel I'm barking up the wrong tree, I'm willing to consider alternatives.

Regards,