Models of frication: Advances and limitations
Christine H. Shadle · The Journal of the Acoustical Society of America · 1999
The increasing number of modeling studies of fricatives and frication in recent years is due at least partly to two major advances: increasing computational power makes it feasible to simulate the flow, and thus construct a more physically accurate model; improvements in vocal-tract imaging, particularly magnetic resonance imaging (MRI), allow capture of three-dimensional vocal-tract shape data, especially including small constrictions formed by soft tissue. However, the complexity of both techniques still tends to lead to incomplete models of frication, limited to static and/or highly simplified tract shapes. Further, the three-dimensional data are often used only to generate an area function, ignoring the cross-sectional shape data that can be crucial in defining aeroacoustic sources. Fricatives are not static, and some of their dynamic behavior, such as apparent modulation of the noise source in voiced fricatives, may be especially ill-served by quasistatic models. Dynamic MRI techniques can be used to acquire shape data with which to pursue dynamic models. Finally, the program VOAC will be discussed briefly (and in more detail elsewhere in this session) as an example of a way to use three-dimensional data and acoustic-flow interactions to enhance the model without unduly increasing computational, or conceptual, complexity.