By way of context, this method was developed during a session between Charles Shriner and I. Charles is in Indianapolis, and I’m in Ithaca, NY. Sessions between players at longer distances obviously won’t be able to achieve the same low latency that we were (in our case, it was about 50 ms.) The description of our method will be pretty terse — in part because it’s pretty simple, and in part because the details have vanished from my memory (the session was about a month ago).

• Like JamKazam, you need a hard-wired Ethernet connection. WiFi won’t work.
• The initial (pre-experimentation) settings should be these. All are accessible from the OPTIONS tab that appears when you click the gear wheel (settings) in the upper left corner of the Sonobus window.
  • Default connection should be PCM 16 bit. “Change all connected” should be checked.
  • Default jitter buffer should be set to 1 ms and the options for buffer change (dropdown list to the right of the default jitter buffer setting) should be set to Auto-Up.
  • Auto-adjust drop threshold should be set to 2 seconds
  • Use specific UDP port should be checked, partly because if you’re running a firewall on your router you’re going to need it, and partly because I think that with some models of router or cable modem it reduces latency.
  • Override device sample rate should be checked.
• Moving to the AUDIO tab:
  • Sample rate should be 48K (make sure this is matched to the audio interface that’s feeding whatever computer you’re using, or your AUM session under iOS)
  • I use a 256K buffer. Not sure that’s necessary, but it probably helps.
• Now, as to the experimentation:
  • Fiddle with the buffer and jitter numbers until you find the lowest setting that does not result in noise or dropouts.