Robert P. Carlyon
MRC Appl. Psych. Unit, 15 Chaucer Rd., Cambridge CB2 2EF, England
Two concurrent harmonic sounds were mixed and passed through a bandpass filter, with cutoffs of either 20--1420 Hz (``low condition;'' components resolved by auditory system) or 3900--5400 Hz (``high condition;'' components unresolved). One complex (the ``masker'') had an F0 of 210 Hz in both intervals of each 2IFC trial. Sensitivity (d') was measured to F0 differences (``(Delta)F0's'') in the other (``target'') complex between the two intervals of each trial. In the low condition, listeners could hear two clear pitches in each interval, and performance on the discrimination task was reasonably good (d' approximately 0.75 for (Delta)F0=4%) regardless of whether the masker was gated synchronously with the target or started 150 ms before and ended 150 ms after it. In contrast, the mixture in the high condition was perceived as a single, aperiodic sound, similar to a crackle, and discrimination in the presence of an asynchronous masker was close to chance, even for (Delta)F0=16%. This, together with other findings, suggests that listeners are poor at extracting the F0's of two groups of overlapping unresolved harmonics. Performance in the high condition was better with a synchronous than with an asynchronous masker, and experiments will be presented which test alternative explanations for this difference.