Apart from a few minor excitements, the 20th century has witnessed little progress in our understanding of timbre. Some fine work by researchers such as Gray and Kendall notwithstanding, the promise of multi-dimensional scaling has largely fizzled. Compared with our understanding of pitch, loudness, localization, speech, scene analysis, and other areas of audition, timbre research remains rudimentary and in disarray.
In this talk, I propose a new approach to the study of timbre. The approach has two components. The first component builds on the modest results in ecological acoustics. I will suggest that the "dimensions" of timbre relate to specific evolutionary adaptations in the auditory system. Plausible adaptive "timbre modules" might include "gender identification" (ala Li, Logan & Pastore), "size estimation" (Hogg & Huron), dominance-submissiveness (ala Ohala), food consumption (ala Arnold), energetics (e.g. Warren & Verbrugge), fear detection, and others. The second component is a generalized learning "module" whose role is the opportunistic assembling of cues that help in identifying individual sounds ("that's the fridge," "that's mom's voice," etc.).