My recent work with the POD system (Truax 1977a; 1978a) has dealt with organizing timbre at a form-determining level. It has been concerned less with timbre as the specific acoustic properties of individual sounds than with large-scale spectral structures that define the entire composition. The relation between spectrum and timbre is complex and not entirely understood, but it is generally conceded that spectrum, particularly its temporal behavior, is a primary determinant of timbre. As discussed by Robert Erickson (1975), there is a considerable gray region in which a complex musical event can be heard as a timbre, as a composite sound, or as a chord. Modern psychoacoustics has furthered our understanding of the conditions under which frequencies are heard as individual pitches, heard as components of a spectrum, or not heard at all individually, but rather fused with others as a single percept. In general, we know that it is not only the physical characteristics of the sound that give rise to different perceptions, but also, and perhaps mainly, the context within which the frequencies are heard (McAdams and Bregman 1979). It is the ambiguity among these different modes of perception that fascinates me and that I have explored in Arras (1980), a composition for fourchannel, computer-synthesized tape. FM Timbral Construction