A Novel Method of Binaural Sound Localization Based on Dominant Frequency Separation
LI Cong-qing, Fang Wu, Shijie Dai, Lixin Sun, He Huang, Liying Sun · 2009
A novel method of binaural sound localization based on the dominant frequency separation is proposed. In this method, the ITD values is estimated by using the property for sound signals to keep their dominant frequency components unchanged in the frequency domain. To verify the validity of the suggested method, the test of the dominant frequency separation is made. Firstly, in frequency domain, the separation of the mixed signal of two signals is made by applying this method and the separated result is satisfactory. Secondly, the test of single sound source localization and that of two simultaneous sound sources localization are performed respectively under our test condition. The experimental results show that the proposed method is able to provide a valid estimation of sound source directions and offer less computational complexity.