Auditory filter characteristics inferred from simultaneous masking: effects of procedural variables

Abstract

In a previous experiment [R. H. Margolis and A. M. Small, J. Speech Hearing Res. 18, 571–587 (1975)] the characteristics of the auditory filter were investigated by measuring the detectability (d′) of a long duration (250 msec) tonal signal as a function of the cutoff frequency of a continuous, low-pass, computer-synthesized, noise masker. In the present research, two experiments were conducted to explore the effects of certain procedural variables on the characteristics of the auditory filter inferred from low-pass masking data. In experiment I auditory filter characteristics were found to be critically dependent upon the range over which d′ is allowed to vary. Baseline detectability and the procedure for computing d′ from P(C) were found to be critical variables. In experiment II the effect of a pulsed versus continuous masker was explored. The signal was a 500-msec tone with a rise-fall time of 25 msec. The masker was gated identically to the tone or left on continuously. Tonal signals were slightly more detectable in the pulsed masker condition. The results of these experiments have led to the design of a notch-noise, pulsed-masking experiment for the investigation of auditory filter characteristics in normal and pathological ears. [Work supported by NIH and the Deafness Research Foundation.]