Speech Prosody 2002

Aix-en-Provence, France
April 11-13, 2002

A Duration Model for Czech Text-to-Speech Synthesis

Robert Batúsek

Laboratory of Speech and Dialogue, Faculty of Informatics, Masaryk University, Brno, Czech Republic

Duration model is a standard part of current speech synthesizers. Many types of models have been used recently, e.g. multiplicative models, sum-of-products models or decision tree-based models. This paper follows a decision tree approach. It describes several versions of the duration model for Czech speech synthesis. The model presented here will be implemented in the Czech TTS system Demosthenes.

Full Paper

Bibliographic reference.  Batúsek, Robert (2002): "A duration model for Czech text-to-speech synthesis", In SP-2002, 167-170.