ESCA Workshop on Automatic Speaker Recognition, Identification, and Verification

Martigny, Switzerland
April 7-9, 1994

Extraction of Formants of Oral Vowels and Critical Analysis for Speaker Characterization

Odile Mella

CRIN/CNRS & INRIA Lorraine, Vandoeuvre, France

The aim of the present paper is to study the relative efficiency of the first three formants of different French vowels for speaker characterization. First, we evaluate reliable formant frequencies for these vowels by a method based on the knowledge of the vowel and its context. The frequencies are then used to identify an unknown speaker from a group of ten known speakers. For that purpose, a speaker is represented by a vector of one, two or three formant frequencies of his utterance of a given vowel. We discuss the results of the relevance of every vowel for each combination of formant frequencies with respect to the speech production process. We also compare our results with those of other studies and with the non-uniform female/male formant frequency ratios (ki).

Full Paper

Bibliographic reference.  Mella, Odile (1994): "Extraction of formants of oral vowels and critical analysis for speaker characterization", In ASRIV-1994, 193-196.