4aSC2. Whistly fricatives: MRI and acoustic data.

Session: Thursday Morning, Jun 07

Author: Christine H. Shadle
Location: Haskins Labs., 300 George St., New Haven, CT 06511, shadle@haskins.yale.edu
Author: Michael Proctor
Location: Haskins Labs., 300 George St., New Haven, CT 06511,Haskins Labs., 300 George St., New Haven, CT 06511, shadle@haskins.yale.edu
Author: Khalil Iskarous
Location: Haskins Labs., 300 George St., New Haven, CT 06511,Haskins Labs., 300 George St., New Haven, CT 06511, shadle@haskins.yale.edu

Abstract:

Nonlinear effects in airflow during fricative production can lead to a whistly acoustic quality, which is used contrastively in a few languages and can occur due to dental appliances, but also often occurs in fluent speech. In a magnetic resonance imaging (MRI) study of American English fricatives produced by five subjects, it was noted that some of the subjects produced whistly fricatives. Articulatory and acoustic analyses will be presented to argue that small articulatory changes can produce significantly different acoustic results within the same segment class for a single subject. Lip rounding and presence of a sublingual cavity can contribute to whistling, but a single subject can produce whistly or normal frication if either of these factors is present or absent. Possibly different whistle mechanisms (edge versus hole tone) could account for some of these seeming anomalies. Strident fricatives are therefore an example of a class of speech sounds where small articulatory differences can lead to significant acoustic differences. [Research supported by NIH Grant No. NIDCD R01 DC 006705.]