p a p e r s
p r e s e n t a t i o n s

Kendall, R.A. and Vassilakis, P.N. (2007).  Interactions of timbre with consonance, dissonance, and roughness in multitimbral triads.  Proceedings of the 8th meeting of the Society for Music Perception and Cognition: 30-31. C. Beckett and M. Hall, editors. Montreal, Canada: Concordia University.

In a previous study done at UCLA (2005) it was hypothesized that time-variant spectral properties of natural instrument signals, and the consequent timbres, would map differentially into subject ratings of consonance and dissonance. The literature is replete with examples of rankings of intervals across the chromatic series, differing in methodological approach. Krumhansl in Cognitive Foundations of Musical Pitch (1990) thoroughly reviews and analyzes relations between probe-tone based pitch hierarchies and studies using presumed spectral properties. What has been left out of these various approaches is the use of natural instruments in orchestration. A recent experiment using multitimbral triads consisting of performed, newly recorded signals of flute, oboe and clarinet was designed to further explore the interaction of consonance and dissonance. A set of triads based on C4 for major, minor, augmented, and diminished structures yielded ten possible combinations. These were orchestrated in all possible orchestrations of the three instruments. Subjects at UCLA (music majors) and DePaul (non-majors) rated the roughness of these triads. Results were compared to an acoustical analysis based on a new roughness model and to the perceptual data from an experiment using the same triads, but digitally mixed tones from the MUMS set. ANOVA results indicated no significant differences for timbre, with differences in consonance and dissonance approaching bi-polar categorization rather than precise ordering. A second approach using training of subjects in the perception of roughness resulted in good correlations with the acoustical model in contrast to the previous approach.
The present experiment is designed to converge on hypotheses of spectral centroid and roughness correlation to timbral influences on consonance and dissonance as well as perceived roughness. It is hypothesized that separate groups of subjects for each triad quality (minor, major, diminished, and augmented) can better provide information on the interaction of timbre and consonance/dissonance. Separate groups were assigned to each of four root-based (C4) triads: Major, minor, augmented, diminished. These were orchestrated in all possible combinations of flute, oboe, and clarinet based on Kontakt emulation as well as artificial signal construction (ca. 1.5 sec each). Subjects were asked to rate randomly-presented triads on a scale of consonance and dissonance.
Perceptual data were correlated to the new roughness model and also to calculations of long-time-average-centroid (LTAC). These data were then compared to the previous series of experiments.