Asymmetry of masking revisited: Generalization of masker and probe bandwidth

A band of noise masked by a tone is audible at a much lower intensity increment between masker alone and masker plus just-detectable probe than is a tone masked by a band of noise. To better understand the mechanisms responsible for this asymmetry of masking, we measured thresholds of probe signals ranging in bandwidth from 0 to 256 Hz in the presence of masking signals also ranging in bandwidth from 0 to 256 Hz. Masker and probe center frequencies were 1 kHz in all cases. Two versions of the experiment are reported: (1) masker intensity constant at 70 dB SPL and (2) masker intensity roved over the range 65-75 dB SPL. The results confirm and extend previous results. For a fixed masker bandwidth, threshold intensity increments are essentially constant, and accounted for by a model based on long-term average energy, so long as probe bandwidth does not exceed masker bandwidth. If probe bandwidth exceeds masker bandwidth, threshold intensity increments decrease sharply for all masker bandwidths tested. Roved-intensity results are consistent with the hypothesis that detection is based on long-term average energy for probe bandwidth equal to or less than masker bandwidth, but that for probe bandwidth greater than masker bandwidth other cues are utilized. In addition, we investigated the possible contribution of off-frequency listening by developing and presenting results from a frequency-selective model of peripheral masking. Model results argue against detection based on spectral cues associated with off-frequency listening for signal bandwidths less than a critical band. The conclusion is that detection must utilize the temporal structure of the signal for probe bandwidth greater than masker bandwidth. (C) 1997 Acoustical Society of America.