Voice Isolation Based on LMS Adaptive Filter

Abstract

This project aims to build a way portable (or easily made portable) headset system that effectively filters out the undesired noise while preserving the desired voice signal, allowing two people in close proximity to be able to communicate in a noisy environment. Our system implements a Least-Mean Squares (LMS) adaptive filter to identify and remove the noise from the desired audio signal before it is sent through wireless transmitter to a receiver. Our filter is effective in removing sufficiently loud non-white noise sources from 200 Hz to 4000 Hz, which is the range of most human created noise sources in human environments