5aSC24. First formant spectral properties and initial stop--consonant [voice] judgments.

Session: Friday Morning, May 17


Author: Michelle R. Molis
Author: Randy L. Diehl
Location: Dept. of Psych., Univ. of Texas, Austin, TX 78712

Abstract:

A low first formant (F1) onset frequency is an acoustic correlate of [+voice] consonants. This is predicted by the ``low-frequency hypothesis'' (LFH)---any increase in low-frequency energy near the consonantal closure will enhance the perception of [+voice] stop consonants. In contrast, Kluender, Lotto, and Jenison [J. Acoust. Soc. Am. 97, 2552--2567 (1995)] proposed that F1 onset frequency acts as a cue via an indirect effect on the temporal firing pattern of auditory-nerve fibers. Following voice onset, the synchronous firing of some mid- and high-CF fibers may become captured by energy in the region of F1. The degree of capture grows with decreased F1/F2 distance. It can also be enhanced by increasing the relative intensity of F1. In cases where F1 is more intense, the LFH predicts more [+voice] responses because the absolute amount of low-frequency energy is greater; the ``synchrony capture hypothesis'' predicts fewer [+voice] responses because synchrony capture is more likely as F1 intensity grows. In fact, manipulation of F1 intensity produced no change in subjects' responses indicating that neither hypothesis is adequate in its present formulation. Evidence for changes in synchrony capture is provided by two computer models of auditory-nerve function. [Work supported by NIH.]


from ASA 131st Meeting, Indianapolis, May 1996