Context effects in modulation masking

Abstract

Neal Viemeister is especially known for his contributions to our understanding of perception of amplitude modulation. Modulation masking and signal-to-noise ratio in the modulation domain have been shown to be strong predictors of speech intelligibility in noise. This study uses 40-ms amplitude-modulated tones embedded in a simultaneous 100-ms noise masker to show that for a given type of simultaneous masker, modulation masking can vary significantly depending on the type of stimulus that precedes the masked target. Three types of precursor were used to investigate these context effects, a 2-octave noise centered on the signal frequency (the same as the masker), a 7-component inharmonic complex tone, and a pure tone with a frequency equal to that of the target. All precursors had a duration of 400 ms and were presented in close temporal proximity at levels equal to those of the noise masker. The noise precursor produced on average a 7-8 dB [20log(m)] release from modulation masking. In contrast no change in the amount of masking and a small release of masking were observed for the pure-tone and complex-tone precursors, respectively. Mechanisms that may underlie the observed context effects will be discussed. [Work supported by NIH grant R01 DC015462.]