k2k audio Parameters
| Parameter | Range | Default |
|---|---|---|
| Threshold (dB) | -80.0 – -20.0 | -60.0 |
| Smoothing | 0.0 – 1.0 | 0.5 |
Threshold (dB) — Magnitude threshold (in dB) below which content is considered noise floor, −80 to −20. Lower (more negative) values capture only the very quietest content (clean noise floor without bleed from program material); higher values capture more of the quiet content (breath, room tone, low-level reverb tails). −60 dB is a good starting point for clean recordings; raise to −40 to −30 dB to catch breath and ambience along with the floor.
Smoothing — Time smoothing of the threshold detection, 0–100%. 0% applies the threshold per-frame strictly (clean separation but can produce per-frame artifacts on borderline content). 100% applies heavy smoothing (transitions are gradual but the separation is less precise). 50% is balanced — clean separation with smooth, artifact-free transitions.
About Noise Floor Extractor
Every recording has a noise floor: the constant low-level content that’s not the intended program material — mic self-noise, preamp hiss, room tone, HVAC rumble, distant traffic. This node identifies that content (everything below threshold_db) and extracts it as a separate stream. The two outputs are: the noise floor (extracted) and the program material above the floor (rejected). Use it for: noise-reduction workflows (process the noise stream with heavy smoothing or silence it entirely, then recombine), separating ambient/room tone for independent processing, or as a pre-processing step before further analysis where you want to ignore the noise floor. For dynamic noise gating per-band rather than full extraction, see the Spectral Gate (Dynamics).
Generated 2026-05-05 from K2K_Dev@96730bdc by scripts/gen_lexique.py. Edit _intros/ or _overrides/, not this file.