An articulatory model
Thomas E. Baer, Philip E. Rubin · The Journal of the Acoustical Society of America · 1986
In this paper, we will review the principle features of the articulatory model implemented at Haskins Laboratories [P. Rubin, T. Baer, and P. Mermelstein, J. Acoust. Soc. Am. 70, 321–328 (1981); P. Mermelstein, J. Acoust. Soc. Am. 53, 1070–1082 (1973)], and we will discuss its design considerations. In particular, the Haskins model is intended to identify the key articulators and their linguistically relevant control dimensions, not necessarily to accurately reproduce physiological details. Two steps in the modeling process will be focused on: the transformation from articulator “positions” to midsagital shape, and the transformation from midsagital shape to area function. We will discuss the present implementation of the model and proposed refinements. Issues to be dealt with include the degrees of freedom associated with control of the articulators, and the collection of data to improve width to cross-section rules. Design of control structures for synthesizing continuous articulations will also be discussed. [Work supported in part by NIH Grants HD 01994 and NS 13617.]