Signal and voice separately (voip). Is it real?

Question

Telephony

"Traffic" telephony consists of two parts. The signal and the voice itself. Everyone knows this. The signal can be sent separately from the voice. If we do this, then I don't quite understand. How will the voice "understand" which ip it should address?! The voice itself is planned to be encrypted using zrtp. I sketched the same questions on the diagram itself for clarity https://www.draw.io/?lightbox=1&highlight=0000ff&e...

Answer 1

Why are you reinventing SIP and H323? In them, the signal and voice go separately.

Answer 2

If you don’t want to read RFC, then at least look for articles about SIP on Habré, everything is perfectly chewed there.
I will add:
For many SIP operators (serious), the servers processing signaling and media are different servers, or different server clusters.
In another question, I mentioned SBC and you said you read about it.
So, SBC may or may not hide the addresses of media servers.
In the first case, all traffic goes through the SBC, in the second case, it tells the client where to send the voice.