The Musical Control of Sound Color

Abstract

In a recent article (see Slawson 1981), I hazarded a set of claims that may be considered a theory of an aspect of musical timbre called sound color. In the present paper I shall summarize those claims briefly, develop them a bit further, and than attempt show how one could apply some of them by discussinghowmy composition Colors was structured. Invariance of Sound Color Musical sounds nearly all have such well-known properties as duration, pitch, and loudness. In addition, as we analyze what we hear a bit more closely, we realize that many of them have more arcane attributes having do with such things as the temporal course of a sound's intensity envelope, the regularity versus the randomness of its spectrum, its grain or flutter, etc. In addition all these features, according the theory I am advocating, a sound has a color that can be independently controlled and manipulated by a composer. Like pitch, loudness, etc., sound color (or simply color) is both a psychological attribute and a musical element. The term color has been used refer mixtures of instrumental sounds, but I mean by it something more abstract and more specific something that is not tied the sounds of musical instruments. I have discussed the relationship of sound color the sounds of musical instruments earlier (see ibid.) and will not dwell on it here. Suffice it say that my of the term is not inconsistent with at least certain ways that the sounds of different instruments have been discussed (e.g., see Cogan & Escot 1976). The theory is probably most directly concerned with electronic music, but it is stated ingeneral terms and may have considerably broader applicability. Now for the theory itself. Sound color is associated, not with the spectrum of a sound, but with its spectrum envelope. The first, simplest, and empirically best-established of the theory's claims is the rule: Rule 1: hold the color of a sound invariant, hold its spectrum envelope invariant. Now the concept of a spectrum envelope depends on a kind of dual process model of the production of sound. In this model, a sound is the result of a source modified by a filter. Figure 1 illustrates how a particular kind of regular (harmonic) source is modified by a filter that has two prominent pass bands, resonances, or formants - two hills that reinforce the source frequencies falling within them. The surrounding valleys mute the components of the source in their frequency regions. For all its importance in our knowledge of, for example, the acoustics of speech production (the modern classic in this field is Fant 1960), the spectrum envelope may seem a bit ephemeral. It can be said belong to a resonant object or cavity, but like that object or cavity, it can make no sound on its own (for an enlightening discussion see Huggins 1952). Some mechanical or acoustical energy must excite the object or cavity and only then does its spectrum envelope become detectable in the resulting sound. Rule 1 of the theory is equivalent saying: keep the physical properties of a sounding object or the shape of a cavity constant and the sounds you makeby exciting the object or cavity willhave constant sound color. The most common everyday use of sound color is in vowel sounds. A man, a women, or a child speaking, shouting, or whispering the same vowel sound is exploiting the invariance of sound color that is claimed by Rule 1 of the theory. Filters are used in nearly all electronic music. A particularly striking way in which Rule lis applied in that music involves the of in variant filters link, by means of the resulting invariant sound color, sections of a work having contrasting sound sources. Many electronic pieces feature a great many different sound colors drawn from a rich and varied musical My second claim represents an attempt postulate a theoretical structure forthat space. …