European Conference on Speech Technology
Edinburgh, Scotland, UK
A speaker-independent method of recognising vowels by means of matching sets of formants against templates is introduced. The formants are derived by solving for the roots of linear prediction polynomials. There is a single template for each vowel. The method works well for male speakers but is not so successful with females.
Bibliographic reference. Deterding, D. H. (1987): "Speaker normalisation for a vowel identifier based on robust formant extraction", In ECST-1987, 1313-1316.