Multistage coarticulation model combining articulatory, formant and cepstral features

We describe a multi-stage speech production model containing a linear, phoneme-independent coarticulation lter, followed by a nonlinear component. The latter generates two cepstra which are then additively combined: one corresponding to a relatively smooth background spectrum, and the other representing three formant-like spectral peaks. A neural net is used for both parts, but the second part also utilizes a hard-coded function that generates exactly three spectral peaks. A uni ed model of training, adaptation, and decoding is developed, each operation di ering only with respect to prior probability distributions. Prior probabilities can be introduced at each stage of the model, providing a exible framework for utilizing both speci c and general prior knowledge. We demonstrate the use of this model for speech synthesis as well as recognition.