Handbook for Sound Engineers

Consoles 901

either the positive or negative peaks are more, some-
times greatly, pronounced.
Attack times are measured and quoted either in
microseconds or milliseconds (i.e., the time constant of

the reservoir capacitor versus charging resistor, which

also corresponds approximately to the time a transient

takes to be controlled) or alternatively in dB/ms which

is the rate at which the attenuation changes.

25.12.1.3 Release Time

The purpose of a release time constant is manifold, but

in the case of a peak limiter its value is primarily to

minimize distortion, much as the attack time. If the

incoming sine wave train is above the threshold and the

limiter is trying to contain it, and if the release time is

short, the gain will tend to recover between each indi-

vidual crest of the sine wave. In effect, this is the

reverse of the attack-time distortion. Although there is

the brutality of a fresh set of attack-time-related distor-

tion on each crest, it must not be forgotten that as the

attenuation releases there is a less traumatic but never-

theless real change in shape of the rest of the waveform

as its amplitude changes within a cycle.

Release times are normally maintained much longer

than attack times. With the exception of true transients,

which spike up once and then go away for an indetermi-

nate period of time, if not forever, most sounds tend to

stay around for a while—at least for a few cycles. It can

be reasonably assumed that, once a signal has hit

threshold, more of it will follow; given that, there is little

point in letting the attenuation drop back just to be reas-

serted milliseconds later. The release time is a crude

memory of the size of the signal the section is having to

deal with at a given moment and by keeping the amount

of attenuation relatively stable gives less work to and

less damage to wreak for the attacking charge ramp-up.

A longer release time constant gives the attack circuitry

less to do except at the onset of material over the

threshold.

There is always the danger with long release times

(if chosen to minimize distortion) in that should a large

transient come along, the limiter will do its job and

promptly reduce the gain to prevent excess output level.

Fine, but the long release time keeps the attenuation

invoked sitting around for a long time, compressing

following program material and a large amount of

following information can be lost until the gain claws

its way back up to normal level, Fig. 25-78.

The subjective compromise between long release

times (for distortion’s sake) and rapid recovery depends

largely on the program material. More so with release

than with attack; too short a release time can really tear

up bass frequencies. Distortions due to attack and

release time constants—transient, intermodulation, and

harmonic—owe themselves to the fact that gain is

Figure 25-77. Attack time effect on waveshape.

A. Zero attack time.

B. Medium attack time.

C. Long attack time.

Unlimited

signal

Signal reducing

in amplitude

Limited

signal

Overshoot

E. Effect of medium attack time.

F. Effect of long attack time.

D. Effect of zero attack time.