The tritone paradox revisited: Effects of musical training, envelope peak, and response mode

Abstract

Eighteen musicians and nine nonmusicians judged the direction of tritone intervals between pairs of octave-complex (Shepard) tones. The tones in each pair had the same spectral envelope centered on either C or F♯ in one of three octaves. Nine of the musicians sang the intervals on one-half of the trial blocks instead of making the usual up/down judgments. Monte Carlo analysis revealed that 83% of the musicians and 33% of the nonmusicians showed a significant direction bias for regions of the chroma circle. This corroborates the notion of a chroma-circle template of definite orientation [Deutsch etal., Mus. Perc. 5 (1987)], but suggests that musical training may play a greater role in either its establishment or access. Biases in upward judgments were toward the C-chroma region, consistent with Deutsch’s [J. Acoust. Soc. Am. Suppl. 1 88, S139 (1990)] findings for native Californians as opposed to British. With the exception of one musician, envelope peak position had little effect on the results. In general, the two response modes, singing and up/down judgments, produced converging evidence.