T. G. Forrest
Section of Neurobiol. and Behav., Mudd Hall, Cornell Univ., Ithaca, NY 14853
C. Formby
Johns Hopkins Univ. School of Medicine, Baltimore, MD 21287
The temporal envelope detector [N. Viemeister, J. Acoust. Soc. Am. 66, 1364--1380 (1979)] is a good single-channel model for auditory temporal processing. The model has three stages: (1) an initial bandpass filter, followed serially by (2) a nonlinear element, and (3) a lowpass filter. A digital version of the model using a max--min decision rule [Forrest and Green, J. Acoust. Soc. Am. 82, 1933--1943 (1987)] at the model's output has been used to simulate temporal gap detection (TGD) thresholds for silent temporal gaps in pairs of sinusoidal markers (F[sub 1] and F[sub 2]). Simulated TGD thresholds were measured as a function of marker frequency separation in the same manner as described in psychophysical experiments [Formby et al., companion paper at this meeting]. Simulated TGD thresholds depended directly on marker frequency separation. For a given frequency separation, simulated TGD thresholds were minimized by centering the symmetrical, first-stage roex filter between the two marker frequencies. This single-channel model consistently yielded lower TGD thresholds than did human listeners, but qualitatively simulated an asymmetry in the psychophysical thresholds where the relative change in the TGD threshold was smaller when the frequencies of F[sub 2]>F[sub 1] and larger for F[sub 1]>F[sub 2]. [Research supported by NIH.]