ISCA Archive - Robust and efficient parsing for applications such as text-to-speech conversion (original) (raw)

Robust and efficient parsing for applications such as text-to-speech conversion

Thomas Russi

In automatic synthesis of speech from text, both morphological and syntactic knowledge is necessary for correct grapheme-to-phoneme conversion, accent assignment and prosodic phrasing. In this paper, a language independent framework for syntactic and morphological analysis is presented. It relies on an extended chart parsing algorithm and handles unrestricted text in an efficient and robust manner. This approach has been implemented in the software package Syma and has been used successfully as a text analysis module in a text-to-speech system for German. It is currently being integrated into an English text-to-speech system developed at Edinburgh University.