Speaker normalization of static and dynamic vowel spectral features

Abstract

Two methods are described for speaker normalizing vowel spectral features: one is a multivariable linear transformation of the features and the other is a polynomial warping of the frequency scale. Both normalization algorithms minimize the mean-square error between the transformed data of each speaker and vowel target values obtained from a "typical speaker." These normalization techniques were evaluated both for formants and a form of cepstral coefficients (DCTCs) as spectral parameters, for both static and dynamic features, and with and without fundamental frequency (F0) as an additional feature. The normalizations were tested with a series of automatic classification experiments for vowels. For all conditions, automatic vowel classification rates increased for speaker-normalized data compared to rates obtained for nonnormalized parameters. Typical classification rates for vowel test data for nonnormalized and normalized features respectively are as follows: static formants--69%/79%; formant trajectories--76%/84%; static DCTCs 75%/84%; DCTC trajectories--84%/91%. The linear transformation methods increased the classification rates slightly more than the polynomial frequency warping. The addition of F0 improved the automatic recognition results for nonnormalized vowel spectral features as much as 5.8%. However, the addition of F0 to speaker-normalized spectral features resulted in much smaller increases in automatic recognition rates.

Vowel Acoustic Space Development in Children: A Synthesis of Acoustic and Anatomic Data

“…Selection of the ideal transform for normalization purposes is beyond the scope of this paper. Several issues arise in selecting the ideal transform, including whether the procedure should be vowelintrinsic or vowel-extrinsic (Adank, Smits, & van Hout, 2004), nature of the vowel system (Disner, 1980), and transformation algorithm (Hermansky, 1990;Hillenbrand & Houde, 2003;Miller, Engebretson, & Vemula, 1980;Syrdal & Gopal, 1986;Zahorian & Jagharghi, 1991).…”

Spectral-shape features versus formants as acoustic correlates for vowels

“…This database is also described in both Nossair and Zahorian ( 1991 ) and Zahorian and Jagharghi ( 1991 ). Summarizing briefly, ten of the speakers were men (M), ten were women (W), and ten were children (C) between the ages of 7 and 11 (five male, five female).…”

Caractéristiques acoustiques des voyelles fermées tendues, relâchées et allongées en français québécois /

“…Enfin, d'autres chercheurs (Zahorian et Jagharghi, 1991Watson et Harrington, 1999;Hillenbrand et coll., 2001) …”