Sound Source Location Estimation using Phase Difference of Formants

Abstract

We consider a sound source position estimation algorithm in the large-scale building space like an auditorium. The position estimation is necessary for the automatic operation of the sound image localization. In this paper, we proposed a basic characteristic for the sound source position estimation algorithm based on a phase difference detection of a formant contained in a voice signal. To extract the formant, the proposed algorithm uses the notch bandpass filter (NBPF). Then we detect the phase difference of each formant. As the method for the phase difference detection, we consider study two estimation methods, one of them is a zero-crossing method and the other is a Fourier transform one. As the results, the position estimation performance of the proposed method is clarified by the computer simulation, and we evaluate the proposed method realizing a measurement system.