INVESTIGADORES
GURLEKIAN Jorge Alberto
artículos
Título:
Evaluación de las descripciones analítica y perceptual de la entonación de una base de datos de oraciones declarativas de foco amplio para el español hablado en Buenos Aires
Autor/es:
GURLEKIAN, J. A., TORRES, H. Y COLANTONI, L.
Revista:
ESTUDIOS DE FONETICA EXPERIMENTAL
Editorial:
Universitat de Barcelona
Referencias:
Lugar: Barcelona; Año: 2004 p. 276 - 302
ISSN:
1575-5533
Resumen:
This paper presents an evaluation of the analytic and perceptual description of the prosodic characteristics of a an Argentine Spanish database of 741 broad focus declarative sentences. The analytic description is obtained through the application of the superpositional model proposed by Fujisaki (2003) for several languages. The perceptual description is the result of the application of a labelling method (Gurlekian, 2001) that incorporates psycho-acoustic measurements.. Parameters generated by this labelling method were used to synthesize the intonational contours. Sentences obtained in this way were used in a perception test that involved 10 native speakers of Spanish. Results indicate the validity of Fujisaki´s model for describing intonation of a large data base of Spanish broad focus declaratives, and, thus its suggests the importance of Fujisaki’s model for speech technology applications. The validity of the extended ToBI model is verified by the results of the perception test. Speakers judged the synthesized sentences obtained as highly natural with only minor differences. These results are important in two ways. First, they indicate that quantitative models can be applied to develop Spanish TTS. Second, results suggest that the extended ToBI model is adequate for describing the intonational contours of Buenos Aires Spanish declaratives, and could, then, be extended to the study of other Spanish varieties.