Speaker Identification Based on Physical Variation of Speech Signal

Speaker identification for the speech signal processing request, determining speaker is a challenge due to physical variation. This paper emphasizes new algorithm based on acoustic feature analysis of text-dependent speech. In this proposed method changed by ten variation methods. Acoustic all types voice calculated its arithmetical correlation coefficients and mean value. The audio characteristic with Mel-frequency cepstrum coefficient (MFCC), derivatives double derivatives. An analysed using normal different speakers, mixed data used test training purpose. Passing through various classifiers system. efficiency results are from classifier.