Patterns of residual masking for sounds with speech-like characteristics

Flexible instrumental facilities that permit the administration of psychoacoustic tests and the generation of stimuli under computer control have been developed for the purpose of determining poststimulatory threshold shifts in the context of speech-like sound stimulation. In an initial series of experiments, the system has been used to investigate certain of the auditory effects of brief sinusoidal ramps. The experiments have dealt with 1) matching the pitch of a steady tone to the 'pitch' of a short sweep tone, and 2) with the pattern of residual masking associated with such sweep-tone stimuli. The major result of the pitch-matching experiments is compatible with previous findings in the literature, i.e., subjects tend to place the comparison tone at a frequency close to the terminal value of the ramps. Patterns of residual masking were determined for three conditions of stimulation: 1500-Hz tone, 1000 to 1500 Hz, and 2000 to 1500 Hz sweep-tones (all 50 ms long). The sweep-tone patterns do not exhibit uniform masking over the frequency range of the ramps. They resemble the pattern obtained for the steady-state 1500-Hz tone. Moreover, they display maxima of masking at frequencies that approach the median frequencies of the pitch matches. In neither case does the sweep-tone pattern exhibit greater masking in the frequency region beyond the terminal value of the ramp. The latter result is examined in relation to the recent discussion in the literature of a perceptual overshoot or extrapolation mechanism.