Skip to Main Content
A technique has been developed to locate, define the boundaries of, and appropriately label steady-state voiced non-fricative areas of a continuous speech utterance. It uses advanced methods of pitch and formant tracking to provide the necessary acoustic parameters. The labeling algorithm compares the analyses with vowel-sonorant tables that incorporate some coarticulation phonemena based on current research. The sounds are discussed with respect to the stress of the vowels and the position of the sonorant, i.e., initial or final form. Results for four male subjects are presented.
Acoustics, Speech, and Signal Processing, IEEE International Conference on ICASSP '76. (Volume:1 )
Date of Conference: Apr 1976