Audio Dynamics · screenshot coming soon
What this node does
Audio Dynamics applies one dynamics processor to an audio track, chosen by mode: a compressor (evens out loud/quiet parts), a gate (mutes anything below a threshold, e.g. room hiss between words), a limiter (a hard ceiling that stops peaks from clipping), or a de-esser (tames harsh “sss” sibilance). It takes aCOMFYTV_AUDIO snapshot (or a COMFYTV_VIDEO, whose audio track is used) and outputs processed COMFYTV_AUDIO plus an fx_spec. It has a ▶ Run: the actual processing is done by FFmpeg (acompressor / agate / alimiter / deesser). If no source is wired, the node still emits just the fx_spec so it can be chained into an FX Chain and rendered later.
To feed native ComfyUI AUDIO in or out, insert a Bridge — see bridges.md.
When to use it
- Even out a voiceover where some phrases are loud and others fade away (compressor).
- Silence background hiss or breaths between spoken lines (gate).
- Guarantee a track never clips above a safe ceiling before export (limiter).
- Soften piercing “s” and “t” sounds on a vocal recording (de-esser).
Parameters
mode
Which processor to run:compressor (default), gate, limiter, or deesser. The parameters below apply differently per mode, as noted.
threshold_db
The level (in dB) at which the effect kicks in. Range -60 to 0 dB, default -20. For the compressor and gate, signal crossing this threshold gets acted on; for the limiter it becomes the output ceiling. Not used by the de-esser. Lower (more negative) = the effect engages on quieter material.ratio
How hard the compressor or gate acts once the threshold is crossed. Range 1 to 20, default 4. Higher = stronger gain reduction (e.g. 4:1 compression). Not used by the limiter or de-esser.attack_ms
How fast (in milliseconds) the processor reacts to a signal crossing the threshold. Range 0.01 to 2000, default 20. Shorter = clamps transients faster; longer = lets initial punch through. Applies to compressor, gate, and limiter (the limiter internally clamps this to 0.1–80 ms).release_ms
How fast the processor lets go after the signal drops back. Range 0.01 to 9000, default 250. Applies to compressor, gate, and limiter (the limiter clamps this to 1–8000 ms).makeup_db
Extra output gain (in dB) added after compression to bring the level back up. Range 0 to 24, default 0. Compressor only.knee
How gradually the compressor/gate transitions around the threshold. Range 1 to 8, default 2.83. Higher = a softer, smoother onset. Used by compressor and gate.intensity
De-esser strength. Range 0 to 1, default 0.5. Higher removes more sibilance. Only used when mode isdeesser.
Outputs
Tips
- Limiter needs
alimiter. Some FFmpeg builds are compiled without it; the node raises a clear error if that filter is missing. - The mode you pick decides which knobs matter — e.g.
ratio,makeup_db, andkneedo nothing in limiter or de-esser mode. - For a “safe export” pass, run a limiter last in the chain with
threshold_dbnear -1 to -3.
Related nodes
- Audio Loudness — for target-level normalization (EBU R128 / peak / RMS) rather than moment-to-moment dynamics.
- Audio EQ — tone shaping; often paired before compression.
- FX Chain — renders multiple
fx_specsteps (including this one) in a single pass.