For some time I wanted to make a paging/intercom system based on asterisk servers. The main idea is like this:
Main asterisk <— IAX2 (alaw/gsm(possibly opus)) —> asterisk on endpoint device <— chan_alsa—> speakers
The main problem with this idea is that when I dial an endpoint device, my voice from the speakers sounds like it’s being chopped it peaces by a helicopter engine. Huge portions of speech are missing. At first, I tried a raspberry pi as and endpoint device and blamed the quality for low cpu performance. But yesterday I repeated the same setup with an intel atom embedded board and got 100% the same result,
What am I doing wrong here? Testing sound with aplay and other console based players show good sound quality.