First International Conference on Spoken Language Processing (ICSLP 90)

Kobe, Japan
November 18-22, 1990

A Voice Source Model for Synthesizing Speech With Various Voice Quality Variations

Hideki Kasuya, Yuji Ando, Lu Jinlin, Osamu Komuro

Faculty of Engineering, Utsunomiya University, Utsunomiya, Japan

An acoustic voice source model is presented to generate the voicing waveform and laryngeal noise signal for synthesizing a natural-sounding speech with various voice quality variations. The model incorporates the fluctuation of fundamental periods. Special emphasis is placed on the acoustic measurements of the fluctuations and laryngeal noise from speech signals and on the perceptual salience of these quantities.

Full Paper

Bibliographic reference.  Kasuya, Hideki / Ando, Yuji / Jinlin, Lu / Komuro, Osamu (1990): "A voice source model for synthesizing speech with various voice quality variations", In ICSLP-1990, 81-84.