Perbandingan metode transformasi wavelet sebagai praposes pada sistem identifikasi pembicara

Abstract

Wavelet transformation have become increasingly popular in signal processing as image and speech. Wavelet transformation have demonstrated good time-frequency localization properties and are appropriate tools for the analysis of non-stationary signals like speech. This research would be the comparison of three orthogonal wavelet tipes, that are daubenchies, symlet, and coiflet on fourth orde and 10,15 decomposition level. As Pattern matching has used ANN Multilayer perceptron. Our experiments and simulation results indicate that the wavelet transformation is a potensial contender to be a feature extraction tool for speaker identification and Daubenchies offer the best identification rates are 86%.