Kinematic and Acoustic Steadiness of Sustained Vowels: Typical Production Versus Simulated Dyskinesia and Dysphonia

摘要

Kinematic and acoustic time-series measurements were employed to evaluate the utility of sustained vowel fonnant tracks as an index of vocal tract steadiness (Gerratt, 1983). Two healthy adult speakers produced sustained vowels under normal phonation conditions and while simulating articulatory dyskinesia or dysphonia. The influence of F_0 variability and abnormal voice quality on the formant tracks relative to known articu-lator movements was assessed by comparing acoustic data with kinematic parameters determined by electromagnetic articulography (EMA). Results suggest that formant variability was greater in the unsteady (simulated dyskinetic) condition, corresponding with increased kinematic variability in the tongue and jaw. In the breathy voice condition, F1 variability exceeded that explained by the kinematic data. During rough voice production, increased variability of F1and F2 was consistent with increased variability seen in the articulographic measurements. Within-vowel cross-correlations among acoustic and kinematic measures also suggest that formant variability, even during simulated dysphonia, was more highly related to kinematic variability than to variability of F_0. Findings generally support the validity of using formant variability measures, with reasonable precautions, to quantify vocal tract steadiness for modal and for rough phonation, but not in the case of Fl during breathy phonation or for the vowel /i/ with high-pitched voices.