Summary: EFFICIENT NUMERICAL MODELING OF VOCAL FOLD MECHANICS
PACS number: 43.70.Bk
Universit`a di Padova, Dip. di Elettronica ed Informatica.
Via Gradenigo 6/A. 35131 Padova, Italy
Tel: +39.049.827.7665 Fax: +39.049.827.7699
The literature of glottal models is reviewed and the properties of one-mass models, as opposed to
multi-mass models, are discussed. A one-mass model is then presented, in which the non-linear
interaction between fold displacement and airflow is described through a modified equation that
accounts for vertical phase differences between upper and lower fold edges. It is shown that the
system behaves qualitatively as higher-dimensional models (e.g., the two-mass IshizakaFlanagan
model). Numerical techniques for the simulation of the model are discussed, and a method is
presented which allows accurate and efficient computation of the non-linear terms.
Features of the glottal source signal (i.e. the glottal flow) carry most of the information that
characterizes voice quality and speaker identity [1, 6], and accordingly research on source models
is becoming increasingly important in speech synthesis research. Parametric models fit the glottal
signal with piecewise analytical functions, and typically use a small number of parameters. As