I just wanted to share a tip here, as my microphone is shit. In pulseaudio I could fix with that module, now in pipewire I just imported the module, configuring the file ~/.config/pipewire/pipewire-pulse.conf.d/99-echo-cancel.conf with:
context.exec = [
{
path = "pactl"
args = "load-module module-echo-cancel aec_method=webrtc aec_args=\"analog_gain_control=0 digital_gain_control=1\""
}
]
After this, restarting pipewire with systemctl --user restart pipewire pipewire-pulse, the source appeared and the sound from the Echo-Cancel Source is much cleaner than my microphone.
As I don’t hear anything talking about this, this might be valuable information.
This is a really cool tip!!
You might like this project if you’re interested in a more aggressive filtering option: https://github.com/werman/noise-suppression-for-voice/
I’ve used it in the past to clean up a table mic for voice chat and worked pretty well. Requires a lot more config work than your approach does though.
(Now on my main account)
Yeah, I used that before, but it used a lot of cpu (Idk why). As I searched for an alternative for my old setup on pulseaudio, I found I could just load pulseaudio modules, so I made this post, because I couldn’t find anyone talking about just loading that module, only the werman module with rnnoise.
The module is still cool though! Thanks for sharing.