Application of Different Statistical Tests for Validation of Synthesized Speech Parameterized by Cepstral Coefficients and LSP

Autores/as

  • Carlos Angel Franco-Galvan Universidad Nacional Autónoma de Mexicana, Facultad de Artes BUAP, Laboratorio de Tecnologías del Lenguaje
  • José Abel Herrera-Camacho Universidad Nacional Autónoma de México, Facultad de Ingeniería UNAM
  • Boris Escalante-Ramirez Universidad Nacional Autónoma de México, Facultad de Ingeniería UNAM

DOI:

https://doi.org/10.13053/cys-23-2-2977

Palabras clave:

Speech synthesis, voice parameterization, line spectral pair

Resumen

The following document tries out different statistical norms to validate the quality of synthesized voices applied to an HTS-based Spanish synthesizer, which uses LSP and Cepstral Coefficients parameterizations. Standard MOS tests were carried out. Nevertheless, other types of quality tests were performed to reinforce the MOS results. Such as: MUSHRA, ABX and CCR. The subjective test PESQ was also applied. To validate intelligibility a SUS test was used.

Biografía del autor/a

Carlos Angel Franco-Galvan, Universidad Nacional Autónoma de Mexicana, Facultad de Artes BUAP, Laboratorio de Tecnologías del Lenguaje

PhD Student Posgrado en Ciencia e Ingeniería de la Computación

José Abel Herrera-Camacho, Universidad Nacional Autónoma de México, Facultad de Ingeniería UNAM

Profesor Titular C, Tiempo Completo

Boris Escalante-Ramirez, Universidad Nacional Autónoma de México, Facultad de Ingeniería UNAM

B.Eng. (National University of Mexico), M.E.E. (Philips International Institute, The Netherlands), Ph.D. (Technical University of Eindhoven, The Neherlands)Research interests: Computational models of human vision and its applications to digital image processing.Member of the National Research System (SNI nivel 1)

Descargas

Publicado

2019-06-27