I have been interested in trying to setup a system like this for our
radio station for a while now. How are you planning to do all this,
through jack.udp? If so, it'd be nice if somebody wrote a gui or
I'm completely bypassing the linux network layer and perform
direct communication between the network card(s) and the audio device within
the host's kernel space to keep latencies and CPU consumption at a minimum.
The gui will control the routing between audio card and NICs through an IOCTL
(or even mmapped space).
The Host is not supposed to do any audioprocessing. Exception may by
mastering.
The ultimate goal is to have a fanless Mini-ITX in the studio running the GUI
and the nodes processing the audio next door.
The sequencers on the nodes are themselves gui-less and be controlled
from a single GUI on the host. Tracks being identified by NIC:Node:Track .
Rosegarden4 seems to be a good candidate for something like that, because it
has the separation between gui and sequencer already.
Ralf