The UPC text-to-speech system for Spanish and Catalan

This paper summarizes the text-to-speech system that has been developed in the Speech Group of the Universitat Politècnica de Catalunya (UPC). The system is composed of a core and different interfaces so that it is compatible for research, for telephone applications (either CTI boards or standard ISDN PC cards supporting CAPI), and Windows applications developed using Microsoft SAPI. The paper reviews the system making emphasis in the parts of the system which are language dependent and which allow the reading of bilingual text (Spanish and Catalan). The paper also presents new approaches in prosodic modeling (segmental duration modeling) and generation of the database of speech segments, which have been introduced last year. 1. SYSTEM OVERVIEW The UPC Text to Speech System (UPCTTS) is a bilingual system able to read text in Spanish and Catalan. Catalonia is a bilingual region of Spain, where Catalan and Spanish are spoken almost equally by a large part of the population. These two languages are similar in their set of phonemes. The system has been developed sharing as many resources as possible between the two languages, with the minimum differences in the language-dependent modules. The user can choose the desired language or, using a mark-up language, parts of text can be read in one or the other language, for instance for proper names. The system is being used in several platforms (UNIX, MacOs and Windows32) with different objectives (research, broad applications, and specific applications). To do that, the system is composed of a core, which implement the TTS system and several interfaces. The UPCTTS system can be tried via Internet [1]. 1.1. The Core of the System The core of the TTS system is composed of four modules: text normalisation, phonetic transcription, prosody generation, and speech synthesis. Text normalisation module. This module performs the normalisation of the text. It expands into full orthographic for