Background

Stop for a moment right here and listen to all the sounds, however faint, that are currently around you. Even in a quiet room, there are probably half a dozen different sounds you can distinguish. Each of these sounds is made up of several frequencies, and in some cases, these frequencies may overlap. They are all hitting your cochlea simultaneously, but seemingly not straining your ability to distinguish the sounds and localize them in space. This ability to distinguish the different sounds in the ambient environment has been called auditory scene analysis (Bregman, 2005).

Bregman’s (1990, 2005) view of auditory scene analysis is very much akin to the principles of gestalt psychology. That is, the auditory system uses a number of heuristic rules to determine which frequencies go with which other frequencies and which sounds are associated with which objects. These rules are not perfect—they sometimes lead to errors in auditory scene analysis—but more often than not, they allow us to correctly parse the auditory landscape. Like the gestalt rules, these processes center on the ability to group different patterns of sounds together. Auditory scene analysis rules fall into three basic types. We group by timing—that is, sounds that are produced at the same time may be grouped together (temporal segregation). We group by space—that is, those sounds that are emanating from the same place in space are grouped together (spatial segregation). We also group by frequency—that is, sounds that are of the same frequency or harmonic pattern are grouped together (spectral segregation).

On the next tab you can listen to a piece by Johann Sebatian Bach. It is a violin piece played by one violin. As you listen, try to figure out how Bach is using temporal and spectral signals to make it sound like this one instrument is playing more than one line of music.

Frequency Segregation