SoundBite: Does compressor attack time work above the threshold?
It's often said that the attack time of the compressor is the time taken for the compressor to react once the signal has crossed above the threshold. I've said that myself and it is true but it isn't the whole story. To cut to the chase, the attack time is the time taken for the compressor to respond when the signal level is above the threshold. Even if the level is already above the threshold the attack time still applies. Let me show you... Here I have a sine wave signal that rises suddenly from -20 dBFS to just under 0 dBFS.
And in this track the level rises suddenly from -8 dBFS to again just under 0 dBFS.
I'm going to put a compressor into both tracks with the same settings - ratio 20:1 so we can clearly hear and see what's happening, attack time 10 milliseconds, everything else default apart from the threshold at -10 dBFS. Let's listen to this...
I'll bounce them both to WAV files and I'll open them in iZotope RX. I'm using RX because it gives a good visualization - We can see what's going on. Here's my first file rising from -20 dBFS through the threshold of -10 dBFS to just under 0 dBFS. You can see the overshoot here, which is exactly as we expect because the compressor takes time to react. If I go to a point where visually the signal has settled, I get this - The attack time as seen here is around 45 milliseconds. That's different to the 10 milliseconds as set in the compressor but that is going to be another story for another day. 45 milliseconds roughly is what we see. Now let's look at my second file rising from -8 dBFS, which is already above the threshold of -10, again to just under 0 dBFS. Checking the attack time visually again I get somewhere around 45 milliseconds, so it's the same as makes hardly any difference. So there we have it - Attack time applies to the signal rising through the threshold and when the signal is already above the threshold. Thank you for listening.