Effect Parameters Convention
Every effect builtin in akkado — delays, reverbs, modulation, comb,
filters, distortion, dynamics — exposes the same two mix parameters
at the end of its argument list: dry and wet. The mix is always
out = dry_in * dry + processed * wetper channel. The convention means you can swap from “in-line” to “parallel” mixing on any effect by changing one number, without remembering each effect’s bespoke parameter shape.
Two defaults, by category
Effects fall into two musical categories with different sensible defaults:
Category A — Additive (parallel-mix effects)
Delays, reverbs, modulation (chorus/flanger/phaser), comb. The dry signal is musically essential; the effect adds character on top. Default: dry=1.0, wet=0.5 — a balanced parallel mix
out of the box.
// Reverb with default balanced mix
saw(220) * ar(trigger(2)) |> freeverb(@, 0.5, 0.5) |> out(@)
// Fully wet for send/return
saw(220) * ar(trigger(2)) |> freeverb(@, 0.5, 0.5, wet: 1.0, dry: 0.0) |> out(@)
// Subtle chorus over the dry signal
saw(220) |> chorus(@, 0.5, 0.5) |> out(@)Category B — Transform (in-line effects)
Filters, distortion, dynamics. The effect replaces the signal.
Default: dry=0.0, wet=1.0 — pure processed output, back-compat
with how filters and distortion have always worked.
Parallel processing (NY compression, parallel distortion, blended filter) is one keyword away:
// Pure filter (default — dry=0, wet=1)
saw(220) |> lp(@, 800) |> out(@)
// Parallel ("New York") compression
saw(220) |> comp(@, -12, 4, dry: 0.5) |> out(@)
// Parallel distortion — mix some grit under the dry signal
saw(220) |> fold(@, 1.2, dry: 0.7, wet: 0.4) |> out(@)Quick reference
| Family | Category | Default dry | Default wet |
|---|---|---|---|
Delays (incl. pingpong) | A | 1.0 | 0.5 |
| Reverbs | A | 1.0 | 0.5 |
| Modulation (chorus, flanger, phaser) | A | 1.0 | 0.5 |
| Comb | A | 1.0 | 0.5 |
| Filters | B | 0.0 | 1.0 |
| Distortion | B | 0.0 | 1.0 |
| Dynamics (compressor, limiter, gate) | B | 0.0 | 1.0 |
Breaking change — delays
Prior to this convention, every delay variant (delay, delay_ms, delay_smp, tap_delay, tap_delay_ms, tap_delay_smp, pingpong)
defaulted to fully-wet output. They are now Category A
(dry=1, wet=0.5).
If you have a patch that called delay(in, time, fb) and expected
100%-wet echoes, you’ll now hear the dry signal mixed in with the
echoes at half volume. Restore the prior behavior by passing wet explicitly:
// Old behavior (fully wet, no direct signal in the delay return)
saw(220) |> delay(@, 0.5, 0.4, dry: 0.0, wet: 1.0) |> out(@)See the changelog entry under ## [Unreleased] for the full list of affected builtins.
Tips
- Send/return architecture: send to an effect with
dry=0, wet=1, then mix the returned signal at the bus. Works the same way on every effect. - Wet-only chain element: when stacking effects, set
dry=0on the intermediate stages to avoid summing the unprocessed signal multiple times. - Parallel processing: any Category-B effect becomes a parallel
effect by setting
dry > 0. There’s no separate “wet/dry mix” knob —dryis that knob.