João Paulo Teixeira

Prosody Generation Model for Tts Systems: Segmental Durations and F0 Contours with Fujisaki Model João Paulo Teixeira

This book presents the development of a prosody system for text-to-speech (TTS) applications. The prosody is responsible for a communicative intention and guarantees some naturalness in the uttered speech. The prosodic features consist in the imposition of the timing, characterized by the segmental durations and pauses, the intonation, characterized by the fundamental frequency (F0) curve, and by the intensity curve. The proposed prosody model consists of several sub-models, namely, the duration model to predict the segmental durations and the model to predict the F0 pattern. The segmental durations model consists of one ANN carefully selected concerning its architecture and type as well as the set of input features with the objective of minimizing the error between predicted and measured durations. One alternative model, is based on same considerations but uses one dedicated ANN for each phoneme. The alternative model, with dedicated ANNs, improved the final performance. The proposed model to predict the F0 contour is based on the Fujisaki model and consists of two sub-models. One predicts the Phrase Commands parameters and the other predicts the Accent Commands parameters.