ESCA Workshop on Prosody Lund, Sweden September 27-29, 1993

Prosodic Modelling for a Text-to-Speech System in Spanish

Eduardo López-Gonzalo, Luis A. Hernández-Gómez

Dpt. SSR, E.T.S.I. Telecomunicacion, Univ. Politecnica de Madrid, Ciudad Universitaria, Madrid, Spain

In this contribution, we present the results in modeling Spanish prosody for a text-to-speech system based on units concatenation using a TD-PSOLA synthesizer. For this purpose, we have developed a new methodology to transfer to a text-to-speech system the prosody of one speaker considering both fundamental frequency and duration jointly, trying to take into account their interactions. The results on a corpus test over the system shows very good intelligibility and naturalness.

Full Paper

Bibliographic reference.  López-Gonzalo, Eduardo / Hernández-Gómez, Luis A. (1993): "Prosodic modelling for a text-to-speech system in Spanish", In Prosody-1993, 74-77.