Applied Speech Enhancement in Mobile Communication Acoustics

Abstract: This thesis presents two solutions to the problem of mobile communications degradation due to interfering background signals. The target lies in getting an estimation of the background noise signal that accompanies a phone conversation and remove it. The proposed solutions allow mobile phone use anywhere without worrying about the background noise that exists and how it interferes with the ease of conducting the conversation, as these signals are filtered before reaching the recipient. The background noise will be reduced and the overall quality of the signal improved thus achieving a clearer conversation without noise interference depending on where it takes place. There are several different techniques used for noise cancellation but we will focus mainly on Wiener and Spectral subtraction algorithms.