> We’ve made significant improvements in noise-suppression and auto-gain-control as well as slight improvements in echo-cancellation to provide you with a better overall experience.
What does that even mean?
What does that even mean?
https://wiki.archlinux.org/title/Firefox/Tweaks#Disable_WebR...
proved helpful. Wish I'd known sooner.
Take for example echo suppression. For this to work, the algorithm has to be aware of the audio that's being output in real-time. We can't have the server do that, of course, and it would be very inefficient to do this in JavaScript (which also wouldn't have access to audio output by other tabs or processes).
So this has to be something done by the browser, in a very low latency and efficient way.
As far as I understand it, Chrome and Firefox have both been doing browser side DSP in WebRTC for a while. There is a way to disable this however.
It isn't opt-in at the browser level, but there is a standard in WebRTC to disable this across all browsers.