Encoding the fundamental frequency of a complex tone in the presence of a spectrally overlapping masker

A series of experiments investigated listeners' ability to encode the fundamental frequency (FO) of a group of harmonics (the ''target'') in the presence of a second, spectrally overlapping, group (the ''masker''). Experiment 1a was a sequential FO discrimination task between two targets, whose FOs were geometrically centered on 210 Hz, in the presence of a 210-Hz masker. The target and the masker were bandpass filtered identically, either from 20 to 1420 Hz (''low-frequency'' condition) or from 3900 to 5400 Hz (''high-frequency'' condition). In the low-frequency condition the masker affected performance only moderately, regardless of whether it was gated synchronously with, or was turned on 150 ms before and off 150 ms after, each 200-ms target. In the high-frequency condition, the synchronous masker also had a moderate effect, but the asynchronous masker reduced performance dramatically. Whatever the masker gating, listeners did not hear the combination of the masker and target in this region as a mixture of two complex tones, but experienced a unitary noiselike or ''crackle'' percept. Experiment 1b showed that the large deterioration seen in the high-frequency condition of experiment la could be obtained in the low-frequency condition by reducing the FO to 62.5 Hz, suggesting that the resolvability of adjacent harmonics was important for the effect. Experiment 2 required listeners to detect a difference in FO (''Delta FO'') between two simultaneous groups of components, one filtered in the high region and the other in the low region. Performance was only slightly degraded by a continuous masker filtered in the low region, but was reduced to chance by a masker in the high region. Experiment 3 showed that, as the Delta FO between a masker and a target in the low region increased from 1% to 8%, listeners identified the mixture as sounding progressively ''less fused,'' but this was not the case in the high region. It is concluded that listeners are poor at extracting the FOs of two groups of unresolved harmonics in the same frequency region. The experiments provide no evidence that listeners can use the leading part of an asynchronous masker to identify its FO and thereby help extract the target's FO from the mixture. (C) 1996 Acoustical Society of America.