Abstract
The steelpan is a percussive idiophone that originated in Trinidad in the early twentieth century. Its sound is generated by striking an arrangement of notes etched into the sunken top of a steel oil drum. The arrangement of the notes on the pan appears to be critical to the timbre of the instrument, as it impacts the coupling strength between notes: If a note has a neighbor tuned an octave and 4th above, then the 3rd overtone of the note is likely to have a ratio of 2.667 rather than the harmonic 3.0. Further, the temporal envelopes of the partials depend on the coupling strength, with lower-coupled partials having audible lags relative to the fundamental. To enable psychoacoustic testing of how these idiosyncrasies give the pan its “flava,” a parameterization is needed that can morph between tone complexes that have these features, and those of a more typical instrument in which the overtones are essentially harmonic and have similar temporal envelopes. This presentation discusses a model that includes variable delay, onset, and decay times per partial (along with frequency and magnitude). It also describes methods to fit these parameters to recordings of both the steelpan and more traditional instruments.