Abstract
To realize physically accurate sound field reproduction, the boundary surface of a sound field to be reproduced should be spatially discretized with intervals smaller than a half wavelength. Otherwise, spatial aliasing will occur in the reproduced field, which leads to low physical reproducibility. Therefore, accurate sound field reproduction covering the full audible range up to approximately 20 kHz requires an impractically large number of sampling points, namely, the number of microphones and loudspeakers. However, it may be possible to reduce the number of sampling points if the degradation in the physical performance due to spatial aliasing does not degrade the spatial perception of the reproduced sound field. To achieve such perceptual optimization of a sound field reproduction system, it should be clarified how spatial aliasing has negative effects on the physical and perceptual reproducibility of a sound field reproduction system. Therefore, as a first step, to investigate the physical reproducibility of sound field reproduction with spatial aliasing, we numerically simulate the reproduced sound field and binaural signals that will be reproduced when a listener is inside the reproduced sound field. The numerical results of the reproduced sound field with spatial aliasing showed that sampling intervals larger than a half wavelength yield unnecessary wave fronts that reach a listener 1 ms after the main wave fronts. Furthermore, the results of the numerical simulation of a binaural signal, employing the boundary element simulation of a human head, suggested that interaural time differences and level differences are approximately reproduced when the upper-bound frequency of physically accurate reproduction is greater than 4 and 8 kHz, respectively.
