I have an asterisk 13.22, as a test bed. I have two phones connected to it, and A calls B - when I speak in both phones at the same time, the audio sent to the other end is degraded (distorted, and volume reduced) in some sections of the overlap.
I have captured the audio frames before and after asterisk, via libpcap. Both channels are chansip channels, in ALAW. The audio before asterisk is fine, and what is sent to the other party is degraded.
I’m a bit puzzled, as I would expect asterisk to just pass along the RTP packets without touching them, as there is no translation.
Why is this happening, how could I prevent it ?
That’s double-talk. You’ve got acoustic echo cancellers on both ends working at the same time. That’s what happens.
You haven’t provided anything that would explain that behaviour and you haven’t told us what is technically wrong with the media stream. I don’t think anyone can give a sensible answer, although I would look for severe resource constraints.
Even resource constraints won’t affect the volume.
@malcomd : I dont see that as an echo cancellation issue on any end - I capture the rtp stream before and after asterisk, and the incoming stream is fine, while the outgoing stream shows volume reduced and sound distorted
@david55&: the audio is inaudible on the other end, volume is low and sound distorted - see: https://ibb.co/eRN0LK
Also, though the test is done with the same audio on both ends, the issue also occurs when two people speak different sentences at the same time, so not an echo issue