What is attack?
On a compressor, attack is the time setting that determines how quickly compression takes hold once the signal crosses the threshold. It is usually shown in ms (milliseconds). Shorter values react faster, and longer ones let the front of the sound through before clamping down. The exact definition of the time and the curve differ between models.
Characteristics of a fast attack
A fast attack makes it easier to catch short peaks such as drum hits and consonants. It is useful for stabilizing level or rounding off sharp transients. Too fast, though, and the definition and drive of the performance are lost; on low frequencies the compressor can even respond to the waveform itself and sound distorted.
Characteristics of a slow attack
A slower attack preserves the transient at the front of the sound and compresses the body and sustain that follow. This can push drums forward or keep the feel of a plucked bass. Too slow, however, and short peaks are barely controlled at all, defeating the purpose of peak management.
Setting it together with release
Attack is the time going into compression; release is the time coming back out of it. Start with slightly exaggerated compression so that you can hear the difference, and set attack according to how much of the front of the sound you want to keep. Then adjust release to the tempo of the song and check that the gain reduction returns naturally before the next note.
Watch out for the same name elsewhere
The attack in a synth’s ADSR is the time it takes for the level to reach its maximum after a note starts, a different target from a compressor’s attack. Here we mean the response time of compression. For a practical example of adjusting when sidechain compression starts to act, see How to create ducking in EDM (Japanese).
