ASA 130th Meeting - St. Louis, MO - 1995 Nov 27 .. Dec 01

2pPP1. Noise discriminability. I. A comparison of the effects of temporal position and bandwidth.

Martin E. Rickert

Donald E. Robinson

Dept. of Psych., Indiana Univ., Bloomington, IN 47405

In previously reported work [S. F. Coble and D. E. Robinson, J. Acoust. Soc. Am. 92, 2630--2635 (1992); M. E. Rickert and D. E. Robinson, J. Acoust. Soc. Am. 93, 2386(A) (1993)] listeners discriminated among trials consisting of eiether two identical samples of noise or two nonidentical samples. Nonidentical samples were generated by replacing a segment of noise presented during the first interval with a new segment. Although the long-term power spectrum of the segments was the same, the temporal position at which segments were replaced had a significant effect on discriminability: performance was best when changes occurred at the end and was poorest when changes occurred at the beginning. In the present study the effects of temporal position were measured under two spectral conditions. In one condition, noise samples were identically filtered (100--3000 Hz or 455--655 Hz) for the entire stimulus duration (50 ms). The effect of temporal position is reduced with narrow-band stimuli but is not eliminated. In a second condition, the bandwidth was varied within each sample such that one segment was wideband (100--3000 Hz) and the other narrow band (455--655 Hz). Overall performance with mixed stimuli (1) is similar to that with pure wideband noise when the uncorrelated segment is wideband, and (2) is similar to that with pure narrow-band noise when the uncorrelated segment is narrow band. [Work supported by AFOSR.]