Evaluation of Two-Channel-Based Sound Source Localization Using 3D Moving Sound Creation Tool

Abstract

We proposed the way that can repeatedly evaluate the localization methods for moving sounds in the same condition regardless of a kind of methods and a number of microphones. And, we developed two-channel-based sound source localization integrated with a cross-power spectrum phase (CSP) analysis and EM algorithm. This one can localize several moving sounds and reduce localization error. Many sound source localization methods have already been developed. However, they could not be evaluated for moving sound in the same condition because it is hard to build database for moving sounds with accurate track information whenever making experiments. Also, to localize several moving sounds, conventional methods need a lot of microphone and/or prior information such as impulse response data. In this paper, we evaluated our sound localization method using 3D moving sound creation tool and confirmed that our method with two microphones can well localize the voices of a moving talker without impulse response data.