Abeer Alwan, Phil Bangayan
Department of Electrical Engineering 66-147E Enginering IV UCLA 405 Hilgard Av. Los Angeles, CA 90024-1594
Division of Head/Neck Surgery UCLA School of Medicine and VA Medical Center Los Angeles, CA 90024
Speech and Hearing Program Health Sciences and Technology Massachusetts Institute of Technology Cambridge, MA 02139
To be presented at the XIIIth International Congress of Phonetic Sciences, ICPhS 95, Stockholm, Sweeden, August 13-19, 1995.
This paper describes a pilot study into the mechanics of synthesizing severely pathological voices. Successful synthesis of such voices may ultimately provide a quantitative method for evaluating and documenting voice qualities. An analysis-by-synthesis approach using the formant synthesizer KLSYN was used to model the voices of 24 patients suffering from voice disorders. Results suggest a number of modifications to KLSYN that would facilitate synthesis of these voices.
This paper refers to twenty-three pairs of voices, listed below. The voices were originally sampled at 16-bit and 20kHz, but for the purposes of this page, they were downsampled to 8kHz and compressed into 8-bit mu-law format. Thus, these are not exactly the voices used in the study.
Category Natural Synthetic rough female: rf1nat rf1syn rf2nat rf2syn rough male: rm1nat rm1syn rm2nat rm2syn rough-breathy female: rbrf1nat rbrf1syn rbrf2nat rbrf2syn rough-breathy male: rbrm1nat rbrm1syn rbrm2nat rbrm2syn rbrm3nat rbrm3syn rbrm4nat rbrm4syn rbrm5nat rbrm5syn bifurcated female: bif1nat bif1syn bif2nat bif2syn bif3nat bif3syn bif4nat bif4syn bifurcated male: bim1nat bim1syn bim2nat bim2syn bim3nat bim3syn bim4nat bim4syn rough-bifurcated male: rbim1nat rbim1syn strained-breathy female: sbrf1nat sbrf1syn sbrf2nat sbrf2syn strained-rough female: srf1nat srf1syn strained-rough male: srm1nat srm1syn
Back to SPAPL Publications page.
Back to Phil's home page
Philbert Bangayan (email@example.com)