this week i’ve read a bit around the topic of how Audio over IP works and it turns out moving the audio is a solved problem. you use RTP and an appropriate codec, which might be linear 24 bit PCM if you’re using AES67 hardware or AAC if you’re on Airplay. The difficult bit is the signaling. i’ve successfully figured out how to nail up an RTP steam using both ffmpeg and gstreamer and using both PCM and Opus. I’ve learned to write SDP files, and I’ve found VLC is way too fussy to talk to any multicast stream I could make. Also, I figured out how to send a SIP invite by hand, although I think from here I need code to do a full call establishment due to deadline timings.
Anyway, not sure where I’m going with this but it’s probably gonna be the dumbest mashup possible. Like a dial-in multicast Opus stream of my stereo. One thing that does seem clear is that you could probably, with a bit of cunning, initiate a phone call as a MITM, essentially making two SIP dialogs and then optionally dropping yourself out of the signaling circuit. Other open questions include whether you can SIP-dial directly to various providers rather than going through the PSTN which, to a large extent, is now mostly selling phone numbers rather than connectivity.
