A Review on Speech Analysis and Automatic Speaker Recognition

Abstract: Speech recognition technology has been a multi-disciplinary research area since many decades and it found wide range of applications in many areas. Speech analysis can be done in time and frequency domain. In this paper, different techniques for speech analysis (feature extraction) such as short time Fourier transform, linear predictive coding, Mel Frequency cepstral analysis etc. are discussed. Also different methods for automatic speaker recognition also have been discussed. Thus this paper gives a technological review on methods involved in feature extraction and automatic speaker recognition implemented so far.

Keywords: Linear Predictive Coding, Mel Frequency, Wavelet, Vector Quantization, Dynamic time warping, Hidden Markov Model, Artificial Neural Network, Support Vector Machine, Gaussian Mixture Model.