This paper describes the implementation of a new parametric model of the glottal geometry aimed at improving female (and male) speech synthesis in the framework of articulatory analysis!synthesis of speech. The model is controlled by parameters that are tightly coupled to physiology, such as, for example, vocal-fold abduction. It is imbedded in an articulatory analysis/synthesis system (articulatory speech mimic). To introduce naturally-occurring details in our glottal flow waveforms, we included two different kinds of glottal leakage in our model: a "linked leak" and a "parallel glottal chink". While the first is basically an incomplete glottal closure that results in, among other things, a steeper roll-off of the glottal flow spectrum, the latter models a second glottal duct that is independent of the membranous (vibrating) part of the glottis. Our simulations show that, as far as deterministic excitation (not the glottal noise) is concerned, a parallel glottal chink can actually enhance spectral levels at high frequencies relative to the no-leakage case.
Bibliographic reference. Schroeter, Juergen / Cranen, Bert (1993): "Physiologically-motivated modeling of the voice source in articulatory analysis/synthesis", In EUROSPEECH'93, 95-98.