Constant-rate ADSR release keeps the fall rate fixed, so releasing from a lower sustain takes less time

When a user changes the sustain level, an ADSR designer chooses how decay/release respond: constant-time (recompute the rate so the segment always spans the nominal duration) or constant-rate (keep the fall rate fixed, so a segment covering less distance simply finishes sooner). EarLevel’s Part 2 picks constant-rate: the release time is defined for a full-scale fall from 1.0 to 0.0, so releasing from a 0.5 sustain takes less time to complete while the rate of fall stays the same. Constant-rate mirrors how a capacitor discharges — always at the same rate per unit time regardless of starting level — and it avoids recomputing the coefficient every time the sustain changes.

Examples

Release time set for a 1.0->0.0 fall; with sustain at 0.5 the actual release finishes in roughly half that nominal time, but the slope (rate of fall) is unchanged. A constant-time design would instead recalculate the coefficient so it always took the full nominal release time.

Assessment

What is the audible/behavioural difference between constant-rate and constant-time release? Under constant-rate, why does releasing from a 50% sustain finish faster than the nominal release time?