Binaural Masking of Speech by Periodically Modulated Noise

Publisher Website
Google Scholar
Abstract
The interference with intelligibility of monosyllabic words produced by continuous white noise, by modulated white noise, and by continuous speech (single talker) was studied during homophasic (N0S0) and antiphasic (NπS0) listening. Five signal-to-masker ratios, four modulation rates, and four magnitudes of modulation were used. Reception in the continuous noise was characterized by steeply sloping intelligibility functions and a 4.5-dB masking-level difference favoring antiphasic listening. Reception in modulated noise changed with the rate and depth of modulation. A 7-dB modulation yielded intelligibility functions highly comparable to those for continuous noise having the same average level. By contrast, more extreme modulation (14, 21 dB, and complete interruption) produced better intelligibility under both homophasic and antiphasic conditions than did continuous noise. This effect was particularly great when noise was completely interrupted either 4 or 20 times/sec, under which circumstances intelligibility remained above in 80% in a speech-to-noise ratio of −24 dB. The advantage of antiphasic over homophasic listening, or masking-level difference, was fairly similar for all conditions of modulated noise, averaging 3.9 dB. When the masking signal was a single competing talker, the antiphasic advantage dropped to 3.3 dB, and the intelligibility function did not duplicate any of the functions obtained in white noise, either continuous or modulated. Nonetheless, individual sets of conditions occurred where masking by speech and by modulated noise yielded equivalent performance, but the depth of modulation required for this equivalence varied with the speech-to-masker ratio being employed.

This publication has 0 references indexed in Scilit: