[Asterisk-Dev] IAX timing of VOICE frames

[Lee Howard]

What is the right way for an IAX client (softphone) to time the sending 
of VOICE frames to the peer?

The voice frames are 20ms in length, of course, but I find that if I 
send them 20ms apart that very short periods of silence happen in the 
other end's audio stream (the audio sounds "choppy").  If I send a VOICE 
frame exactly when I receive a VOICE frame then that works out quite 
well... so long as the other end initiates it, but that is not always true.

Any insight appreciated.

Thanks,

Lee.