Abstract:
Simultaneous pairs of 200-ms stimuli were generated in which the two sounds differed in both spectral content and pitch. In one condition, the stimulus consisted of two bandlimited noises of different center frequencies, amplitude-modulated (AM) at two different rates; in another, the two stimuli were sums of sinusoids separated by two different frequencies and had two different spectral weighting functions. In an adaptive 2AFC paradigm, subjects were asked to associate one of the spectra with one of the pitches. Within any block of trials, the two AM- or frequency-separation pitches were held constant. Results indicate that, generally speaking, there is a trade-off between pitch difference and spectral weight difference. Spectral difference thresholds were generally lower for inharmonic than for harmonic complexes. In frequency regions above 2 kHz, however, both the effect of harmonicity and the slope of the trade-off diminish, especially for low-pitch frequencies. [Work supported by NIH and the VA Medical Research.]