Turn Down The Loudness! What?

Loudness is related to amplitude, but it isn’t the same thing – loudness is a subjective perception, while amplitude is a physical magnitude. The VCA in the compressor regulates amplitude, and compressors were originally designed to solve amplitude problems, as in: “Wouldn’t it be nice if we didn’t have to watch these stupid meters all the time and keep the music below the zero level or lose our broadcasting license? Why don’t we just design a device that automatically turns down the level whenever the music goes above zero? Hey Joe! (actually it was Morris) Hear that? Design a device that automatically turns down the level whenever the signal goes over zero, OK? Have it ready for our live broadcast schedule next week!” - and so the limiter was born. The compressor was invented a week later when the producer bitched about what the limiter did to her (actually it was his) production values and music quality! So it is useful to briefly consider the relationship between amplitude and loudness. Also, take a look at the sidebar for a more detailed look at the meaning of amplitude.

Overall, our sensation of loudness is a highly compressed version of physical reality, so that in order to change a sound so that it sounds twice as loud, we have to multiply the power of the signal by 10 (which is the same as multiplying its RMS voltage by approximately 3). Big changes in amplitude result in more modest changes in loudness (you’ve got to multiply the power by 1,000 (30 dB) to boost the loudness enough to really startle the listener). Now this is where it begins to get weird. As illustrated by the Fletcher-Munson Equal Loudness Contours, the changes in subjective loudness for a given physical change are different for different frequency ranges, so that for any given amount of physical amplitude change, the change in subjective low and high frequency loudness is much more than the change for mid-frequency loudness. Thinking about this makes my head hurt, but it means that when you compress and equalize, or when you stick an equalizer in the side-chain of the compressor, you can get into some very interesting and/or confusing situations. It is why you can, with perfectly good intentions, destroy a good sounding track: often, low frequency energy drives the compressor action while it is mid frequencies you want to control.

Sense of loudness is also affected by time and frequency. The duration of impulsive sounds have a huge effect on how loud they seem. Some years ago I was taking a tour of the Danish Acoustical Laboratories just outside of Copenhagen, and one of their set demonstrations was to fire off a .38 caliber starter’s pistol in both their humungous anechoic chamber and also in their reverberant room. In the anechoic chamber, with no reflections and a pressure shock wave from the pistol that probably lasted no more than about a millisecond, the gun sounded like a little cap pistol – pop! – even though it put out a gout of flame from the muzzle about six inches long. In the reverb room (a room with as little absorption as possible in order to permit testing of sound absorbent devices), where the sound echoes on for about two months, firing the same pistol caused the end of civilization as we know it, or so it seemed at the time: KAAAAABOOOOOOOOOMMMMMMMMMMMMMMMmmmmmmmm! What this illustrates is that you can increase apparent loudness by a whole lot without increasing the level (hell, you can even decrease it by quite a bit) just by increasing the length of time the sound is at its level. This is done by squashing the peak and then boosting the overall level.

About the frequency thing: our hearing mechanism hears across a range of about thirty so-called “critical bands,” each about a third of an octave wide. Subjective loudness is proportional to the actual number of critical bands that are invoked by a given sound. A sine wave (invoking a single critical band) will sound very soft compared to pink noise (invoking all thirty bands) of the same amplitude.