Virtual sound source construction based on adaptive crossfade processing with electro-dynamic and parametric loudspeaker arrays

Among the various high-presence sound field reproduction technologies, we focus on wave field synthesis (WFS) using a loudspeaker array, as it can construct a virtual sound source (VSS) with few restrictions. A VSS using parametric array loudspeakers (PALs) can provide more accurate direction perception than VSS using electro-dynamic loudspeakers (EDLs) because PALs hold sharp directivity by utilizing ultrasounds. However, the distance perception tends to be degraded. We previously proposed a VSS construction based on WFS with hybrid control of an EDL and a PAL arrays. This method enhances distance perception by exploiting the difference in radiation characteristics between PALs and EDLs. However, the difference in frequency responses between PAL and EDL is noticeable to the listener when switching between them. In this paper, to enable more smooth transition of the EDL and PAL arrays, we propose a VSS construction method based on adaptive crossfade processing with EDL and PAL arrays. For an arbitrary VSS construction position, the coefficients of crossfade for each frequency are determined by an adaptive filter based on measured impulse responses of VSSs with EDL and PAL arrays. We demonstrated the effectiveness of the proposed method through evaluation experiments.