Analysis of MMSE Speech Estimation Impact in West Sumatra's Noises

Abstract

This study aimed to estimate the original voice signal which is interrupted by noise with MMSE method based on Wiener filter. The Wiener filter is classified as the MMSE estimator studied by previous researchers. The study assessed the voice signal input count down by European woman that are distorted by two types of noises, the one is noise based on the outdoor location, sound of siren firefighter, and the other is the indoor location noise, which represented by noise in lecturer room in campus. The two process signal is estimated by MMSE estimator which approximated by Wiener filter that must have founded and counted the covariance of each signal processes are related to the system. Thus the researchers tried to estimate both types of sounds noisy European woman are due to interference noise source of fire fighter and faculty room and assess its impact in the form of graphs vote against a function of time, the pattern shape of the signal and SNR.