JEE_V3_N4_RP7 Piyush Lotia Kamal K. Mehta M.R. Khan Journal on Electrical Engineering 2230 – 7176 3 4 46 53 MFCC, VQ, Cepstrum, LBG Algorithm Automatic speaker recognition is a field of study attributed in identifying a person from a spoken phrase. It is widely used in biometric security system, phone banking and other similar applications. This paper presents a development of a Matlab based text dependent speaker recognition system. The model used to derive a mathematical representation of the speech signal. Mel Frequency Cepstrum Coefficient Feature extraction method is used to extract a speaker’s discriminative features from the mathematical representation of the speech signal. After that a feature matching method known as Vector Quantization is implemented using the LBG Algorithm. Feature matching is carried out in order to cluster the speech features into groups of specific sound classes. Finally analysis is carried out to identify parameter values that could be used to increase the accuracy of the system. In this paper we have studied the effect of recording environment on the accuracy of speaker identification. April - June 2010 Copyright © 2010 i-manager publications. All rights reserved. i-manager Publications http://www.imanagerpublications.com/Article.aspx?ArticleId=1190