Fourth European Conference on Speech Communication and Technology

Madrid, Spain
September 18-21, 1995

Improvement of the Quality Of Speech Synthesis by Analysis Using Segmentation and Modeling of the Excitation Signal

J. M. Gutierrez Arriola (2), F. M. Gimenez de los Galanes (1), M. H. Savoji (2)

(1) Dpto. Ingenieria Electronica, E.T.S.L Telecomunicacion, Madrid, Spain
(2) Previously with Dpto. Electronica, Universidad de Cantabria, Santander. Spain

The aim of the presented work was twofold: 1) To study whether the effect of the stochastic part of the glottal waveform can be replaced by a postprocessing of the synthesized signal when a mathematical excitation source is used and 2) To find a new way of representing the excitation signal for high quality synthesized speech with the properties of mathematical models which permits avoiding the problems related to prosody control by overlap add methods as used in PSOLA[1].

Full Paper

Bibliographic reference.  Arriola, J. M. Gutierrez / Galanes, F. M. Gimenez de los / Savoji, M. H. (1995): "Improvement of the quality of speech synthesis by analysis using segmentation and modeling of the excitation signal", In EUROSPEECH-1995, 1097-1100.