SIMPLIFIED APPROXIMATION OF THE 4-PARAMETER LF MODEL OF VOICE SOURCE

A simplified approximation of the four-parameter, voice source model developed by Fant er al. [STL-QPSR 4, 1-12 (1985)] is described in this Letter. In our approximation, the computational complexity required to implement the four-parameter model is simplified without reducing model accuracy. The simplified approximation accommodates rapid, on-line adjustment of source parameters during speech synthesis.