StewOS
Source

stewos.audio

Declared in modules/nixos/audio.nix · 8 options

# stewos.audio.enable

Whether to enable audio.

Type
boolean
Default
false
Example
true
Declared in
Set by

# stewos.audio.noiseCancellation.channels

Channel layout the filter runs at, which is also the layout applications see. A microphone carries voice, so mono is normally right; a stereo capture device is downmixed into it.

Type
one of "mono", "stereo"
Default
"mono"
Declared in

# stewos.audio.noiseCancellation.description

Name the filtered source is shown under in mixers.

Type
string
Default
"Noise Canceling Source"
Declared in

# stewos.audio.noiseCancellation.enable

Whether to enable microphone noise cancellation.

Type
boolean
Default
false
Example
true
Declared in
Set by

# stewos.audio.noiseCancellation.gracePeriod

Milliseconds of audio to keep passing after voice activity stops, so the tail of a word is not cut off. This is the plugin's own default and costs no latency: it only holds the gate open for longer.

Type
integer between 0 and 1000 (both inclusive)
Default
500
Declared in

# stewos.audio.noiseCancellation.package

Package providing the noise suppressor as a LADSPA plugin at lib/ladspa/librnnoise_ladspa.so, exporting the noise_suppressor_mono and noise_suppressor_stereo labels.

Type
package
Default
pkgs.rnnoise-plugin
Declared in

# stewos.audio.noiseCancellation.retroactiveGracePeriod

Milliseconds of buffered audio to keep from before voice activity is detected, so the start of a word survives the detector's own delay. The plugin defaults this to 100; it is 0 here because the buffer is added straight to the latency of every capture path on the machine.

Type
integer between 0 and 200 (both inclusive)
Default
0
Declared in

# stewos.audio.noiseCancellation.vadThreshold

Voice activity threshold, as a percentage. Raising it suppresses more background noise at the cost of clipping quiet speech.

Type
integer or floating point number between 0.0 and 99.0 (both inclusive)
Default
50.0
Declared in