Volume 115, Issue 4, April 2004
Index of content:
- SPEECH PRODUCTION 
115(2004); http://dx.doi.org/10.1121/1.1652033View Description Hide Description
A three-dimensional biomechanical model of tissue deformation was developed to simulate dynamic vocal fold abduction and adduction. The model was made of 1721 nearly incompressible finite elements. The cricoarytenoid joint was modeled as a rocking–sliding motion, similar to two concentric cylinders. The vocal ligament and the thyroarytenoid muscle’s fiber characteristics were implemented as a fiber–gel composite made of an isotropic ground substance imbedded with fibers. These fibers had contractile and/or passive nonlinear stress–strain characteristics. The verification of the model was made by comparing the range and speed of motion to published vocal fold kinematic data. The model simulated abduction to a maximum glottal angle of about 31°. Using the posterior-cricoarytenoid muscle, the model produced an angular abduction speed of 405° per second. The system mechanics seemed to favor abduction over adduction in both peak speed and response time, even when all intrinsic muscle properties were kept identical. The model also verified the notion that the vocalis and muscularis portions of the thyroarytenoid muscle play significantly different roles in posturing, with the muscularis portion having the larger effect on arytenoid movement. Other insights into the mechanisms of abduction/adduction were given.
On the ability of a physiologically constrained area function model of the vocal tract to produce normal formant patterns under perturbed conditions115(2004); http://dx.doi.org/10.1121/1.1689347View Description Hide Description
An area functionmodel of the vocal tract is tested for its ability to produce typical vowelformant frequencies with a perturbation at the lips. The model, which consists of a neutral shape and two weighted orthogonal shaping patterns (modes), has previously been shown to produce a nearly one-to-one mapping between formant frequencies and the weighting coefficients of the modes [Story and Titze, J. Phonetics,26, 223–260 (1998)]. In this study, a perturbation experiment was simulated by imposing a constant area “lip tube” on the model. The mapping between the mode coefficients and formant frequencies was then recomputed with the lip tube in place and showed that formant frequencies and representative of the vowels [ʊ,o,u] could no longer be produced with the model. However, when the mode coefficients were allowed to exceed their typical bounding values, the mapping between them and the formant frequencies was expanded such that the vowels [ʊ,o,u] were compensated. The area functionsgenerated by these exaggerated coefficients were shown to be similar to vocal-tract shapes reported for real speakers under similar perturbed conditions [Savariaux, Perrier, and Orliaguet, J. Acoust. Soc. Am., 98, 2428–2442 (1995)]. This suggests that the structure of this particular model captures some of the human ability to configure the vocal-tract shape under both ordinary and extraordinary conditions.
115(2004); http://dx.doi.org/10.1121/1.1652588View Description Hide Description
Magnetic resonance imaging (MRI) has served as a valuable tool for studying static postures in speech production. Now, recent improvements in temporal resolution are making it possible to examine the dynamics of vocal-tract shaping during fluent speech using MRI. The present study uses spiral k-space acquisitions with a low flip-angle gradient echo pulse sequence on a conventional GE Signa 1.5-T CV/i scanner. This strategy allows for acquisition rates of 8–9 images per second and reconstruction rates of 20–24 images per second, making veridical movies of speech production now possible. Segmental durations, positions, and interarticulator timing can all be quantitatively evaluated. Data show clear real-time movements of the lips, tongue, and velum. Sample movies and data analysis strategies are presented.