I'd like to ask for some pointers on how to optimize my Jitsi settings for echo cancellation.
My test system consists of 2 OpenVPN-connected computers. There's no firewall between the 2 hosts and I created "RegistrarLess" SIP accounts on both clients so that I can make direct SIP calls between the 2.
I can establish the call and have a conversation but there are quite a few problems:
1) audio is somewhat scratchy and "electronic" (may be due to network latency or selected audio codec so I'm leaving that for later testing)
2) when anyone of us activates the webcam, audio becomes horribly scratchy to a point that I can't hear the other party talk. Again, this can be due to a network overload so I'll fine.-tune that later on. However, I find that strange since both peers have around 400Kbps upload bandwidth dedicated to this test.
3) when the "scratchy" voice finally settles down, both parties experience echo problems. Both Jitsi softwares use Pulseaudio with echo cancellation and noise reduction. However, there doesn't seem to be any echo cancellation because everything one says is repeated in full, even after several minutes. I'm wondering if Jitsi's echo cancellation only works with a particular audio codec?
Any hints/tips as to what I can try to do to improve my Jitsi experience?