Speaker identification methods based on pseudostationary segments of voiced sounds

Article type: Research Article

Authors: Lipeika, Antanas | Lipeikienė, Joana

Affiliations: Institute of Mathematics and Informatics, Akademijos 4, 2600 Vilnius, Lithuania. E-mail: lipeika@ktl.mii.lt

Abstract: The problem of speaker identification is investigated. Basic segments – pseudostationary intervals of voiced sounds are used for identification. The identification is carried out, comparing average distances between an investigative and comparatives. Coefficients of the linear prediction model (LPC) of a vocal tract, cepstral coefficients and LPC coefficients of an excitation signal are used for identification as features. Three speaker identification methods are presented. Experimental investigation of their performance is discussed.

Keywords: speaker identification, likelihood ratio distance, cepstral distance, vocal tract, excitation signal, reliability reserve, vector quantization

DOI: 10.3233/INF-1996-7405

Journal: Informatica, vol. 7, no. 4, pp. 469-484, 1996