Sound enhancement system using selective binary filtering

Abstract

Our proposed system allows users to selectively focus on particular sounds by choosing from various filtering options. Our goal is to develop a sound enhancement system which increases human auditory performance, especially through selective listening. To manage multiple filtering options numerically, we propose a series of listening patterns consisting of sets of binary indicators which show whether or not signals from a particular source will be presented to the listener. We then calculate the similarity of each pattern and order these patterns on the basis of their similarity. We conducted an experiment to evaluate our proposed method. Subjects listened to a sound recording using the various patterns in the order determined using the proposed method, as well as listening in an order determined using a comparative method. Subjects were asked to evaluate the naturalness of the transitions from pattern to pattern using each method, to select a listening pattern corresponding to one which was described to ...