We have a conference hotline with roughly 65 participants at all times. On occasion we’ll have small spots where the audio will be garbled, it sounds like packet loss. When we do packet captures the audio coming in from the speaker is clear. However, we did notice that there was some packet loss from other participants that were not speaking. We have to assume that the problem happens when the audio is mixed by Asterisk. I already have jitter buffer applied to the conference bridge, and I have set they sync rate to 10 ms. Is there anything else I can try to make the audio smooth?
Contains LLM output.
Google’s AI Overview says:
In the context of Asterisk, “sync rate” or “Synchronization Rate” doesn’t have a standard, widely accepted meaning within the software itself. It’s not a configuration setting or a commonly used term when discussing Asterisk’s performance or functionality.
Which setting are you referring to?
Also, my gut feeling is that 10ms is far too low for most setting, and rather low for packet sizes on RTP. It would certainly be far too low for any jiitter buffer setting.
I’m talking about the mixing_interval found in confbridge.conf under [general].
Update:
I have set the mixing_interval to 40 ms and it seems to have really improved the audio quality of the conference and has reduced overall CPU consumption by about 2%.
Update again:
Even though the audio quality overall sounds really good, the intermittent garbled audio problem still remains. We believe we have tracked the root of the problem to hardware devices, specifically the EMPhone causing RTP drift.