Virtual sound field immersions by beamforming and effective crosstalk cancellation using wavelet transform analysis

Abstract

Sound field perception and realization of real spaces in virtual environments using effective binaural technology were investigated. The objective of this study was to design and develop techniques and algorithm that enable us to record perceptual sound fields in reality, reconstruct it, and perform coherent emulation in virtual spaces. The employed technique involves binaural technology with adaptive beamforming and effective crosstalk cancellation using wavelet transformation. Before virtual sound field immersions, capturing the real sound along with the room's acoustic and psychoacoustic parameters and spatial decomposition are essential tasks. Regeneration is addressed in the following two ways: 1) headphones are fixed to the user's head, and the user moves simultaneously with them so that head and body movements do not change the coupling between sensors and the user's ears; 2) the loudspeaker arrays are positioned away from the user's body, and the user moves in accordance with the sound sources. The goal of creating a full virtual space with properties of co-existence with remote spaces was accomplished using knowledge about human sound perception, cognition of the environment of the sound, and considering external events with natural impression.