Skip to main content
k2k audio logo k2k audio

Back to Dynamics
Documentation tree

Spectral Ducker — Per-frequency-band sidechain ducking

Ducking pulls one signal's level down in response to another.

Parameters

ParameterRangeDefault
Threshold-92.0 – 0.0-20.0
Range-96.0 – 0.0-20.0
Attack0.1 – 100.05.0
Hold0.0 – 500.050.0
Release10.0 – 1000.0100.0

Threshold — Per-band level (in dB) at which ducking engages. When the input exceeds this in a given band, that band’s gain gets pulled down by range dB. Set this where you want the ducker to start reacting — typically around −20 to −30 dB on program material.

Range — Maximum attenuation in dB applied per band when fully ducked, −96 to 0. Sets how far the ducked bands drop. Subtle settings (−6 to −10 dB) gently sit one element behind another; aggressive settings (−40+) duck the offending bands almost completely.

Attack — How fast ducking engages once a band’s level exceeds threshold, in ms. Fast attack snaps the duck on quickly — useful for crisp ducking in time with hits. Slow attack creates a softer, more gradual pull-down — useful for ducking against sustained material.

Hold — Minimum time in ms the duck stays engaged after triggering, even if the input drops below threshold. Prevents flutter on borderline signals. Useful for keeping the duck active through a full word or hit rather than chattering on every micro-pause.

Release — How fast the band returns to normal level after hold elapses, in ms. Fast release recovers quickly (lively, more obviously rhythmic). Slow release smooths the pull-up, more transparent. Match release to the source — quick for percussion, slow for vocals and pads.

Additional controls

Bands — Number of frequency bands processed independently, 1–32. With many bands the ducker only pulls down the specific frequencies competing with the input — leaving everything else untouched. With few bands it behaves like a traditional broadband ducker. 8–16 bands is balanced; 32 is surgical; 4–8 for obvious multi-band character.

Detection Mode — How each band’s level is measured. Peak — instantaneous per-bin level. RMS — short-window average. RMS is smoother and more program-aware; Peak is more reactive.

Sidechain Listen — Solos the detection signal so you can hear what’s driving the ducker. Off (default) is the normal output; on lets you audition the trigger so you can tune threshold and bands to react to the right material. Turn off before mixing — this is a setup tool, not a normal output state.

About Spectral Ducker

Ducking pulls one signal’s level down in response to another. The classic example: voiceover ducking music — when the voice plays, the music drops; when the voice stops, the music returns. Spectral ducking goes further: instead of pulling the whole music down, it only pulls down the frequencies the voice occupies, leaving the rest of the music intact. The result is a more natural mix where elements share spectral space without competing — kick and bass, vocal and pad, lead and support — without obvious gain pumping.


Generated 2026-05-05 from K2K_Dev@96730bdc by scripts/gen_lexique.py. Edit _intros/ or _overrides/, not this file.