Fuzzy Sound Field Classification in Devices with Multiple Acoustic Sensors

Abstract:
A method for a fuzzy sound field classification in mobile devices with multiple acoustic sensors is presented in this article. The classification is suited for multi-sensor applications with at least two acoustic sensors and is based on a set of sound field indicators. At least four acoustic sensors with an adequate alignment are required for a classification that is not related to the direction of the dominant sound source. The classification works independent from the source signal and classifies the sound field based only on spatial properties that are derived from the input signals. The classification focuses on the four most common basic sound fields described in room acoustic and signal processing literature; free, diffuse, reactive and noise. By classifying these four sound fields, it is possible to determine the energy distribution of the individual components of a sound field. Simulated and measured input data are used to verify and evaluate this method. Index Terms — fuzzy sound field classification, basic sound fields, room acoustics, sound field indicators