Hi again!
I have been prototyping a new system for a couple of months now, and it has prooven to work really well.
Voxcommando is running on my mainframe, as well as on some nodes (Intel compute-sticks). Each node communicates with the mainframe via EventGhost, and the other way around. Eventghost is able to trigger Voxcommando-events in other rooms as well as on the mainframe (For example, I could tell the node in my bedroom to tell whoever is in the kitchen to turn the stove on by sending an utp command to the kitchen-node and eventghost on that node triggering a VC.Speak event with the payload of the eventghost-message).
The microphones on the nodes are really bad, but as I described in this (
http://voxcommando.com/forum/index.php?topic=2777.0) post, I will be testing a couple of (hopefully) better microphones very soon.
This integration takes a bit more work than just running independent nodes, but I believe that this works a lot better as this method solves the issue of where the sound is coming from and where to send the response (if you were to stream audio to a mainframe) as well as the fact that it gives you to say "Turn on the lights" and it will just turn on the lights in the room that you are currently in (room-unique commands).
The system will, naturally, allow you to say "Turn on the lights in the hallway", and after it hears that the node in the hallway will receive a request to turn on the lights (if there is no node in that room then it will send the request to the mainframe, which is able to control the whole house).
Hope that I have inspired you as you always inspire me!
Oskar
