UPCTTS: The UPC Text-to-Speech System The UPCTTS is a system that allows text-to-speech conversion in Spanish and Catalan, using diphone concatenation and a unit-selection procedure when the database contains more than one sample per diphone. It supports SABLE mark-up language. It runs on Windows32 and UNIX computers and several interfaces have been developed for telephone applications using CTI boards or ISDN PC cards, Internet and other SAPI Windows applications. Try our On-line demo New unit-selection based TTS! Have a look on a brief system overview We will be pleased to hear your comments, either on the demo, the system's quality or the page design. Send us an e-mail. TTS On-line This is an on-line demostration of the UPC TTS system. 1. Select a configuration. METHOD LANGUAGE SPEAKER SPEECH RATE MediumSlowFastSlowestFastest Basic Diphone Spanish female male Catalan female male AUDIO FORMAT MP3WAV Unit-Selection Catalan female Spanish female 2. Type your text here OR copy one of the examples below. Then press READ NOW to listen to the result.   Un ejemplo que debe leer en voz alta el conversor texto a voz en castellano. Basic Spanish example: "Se produjo un nuevo incremento del 2 por ciento que elevó el dólar hasta las 187 pts." Basic Catalan example: "Es va produir un nou increment del 2 per cent que va elevar el dòlar fins a les 187 pts." Bilingual -Spanish / Catalan- example (using SABLE markup language): "<SPEAKER NAME="Marta"> <LANGUAGE ID="ca"> El nou director, <LANGUAGE ID="es"> Julián González, </LANGUAGE>ha promès millores. <LANGUAGE ID="es"> "Vamos a dar un giro en la gestión", </LANGUAGE>va afirmar ahir en roda de premsa. </LANGUAGE> </SPEAKER>" Dialogue application example in Spanish (using SABLE markup language): "Hay un tren que sale de <RATE SPEED="slow"> Barcelona </RATE> <BREAK> a las <BREAK> <RATE SPEED="slow"> 22h 50min </RATE> <BREAK>. Gracias por utilizar este servicio. Buen viaje." Dialogue application example in Catalan (using SABLE markup language): "Restriccions per neu a la carretera <RATE SPEED="slow"> <BREAK> N 152 <BREAK TYPE=","> entre Toses i Ribes <BREAK> </RATE> en ambdós sentits." System Overview The system is being used in UNIX and Windows32 platforms with different objectives (research, broad applications, and specific applications). The system is composed of a core, which implement the TTS system, and several interfaces. The core of the TTS system is composed of various modules: Normalize, Pause, Phonetic, F0Contour, Duration, Energy, UnitSelection and Synthesis. Current TTS system selects speech units from diphone databases populated with one sample of each possible diphone (diphone based) or from a speech corpus by a unit selection procedure (unit selection based). Different interfaces have been developped 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. More information about the system can be found in the paper: The UPC Text-to-Speech System for Spanish and Catalan Antonio Bonafonte, Ignasi Esquerra, Albert Febrer, José A. R. Fonollosa and Francesc Vallverdú 5th International Conference on Spoken Language Processing, ICSLP'98 Sydney, Australia, 30th Nov.-4th December 1998 [ Postscript ]

Send us any suggestion

This page was last updated on December 03rd, 2003