Binaural Spatial Rendering Algorithm based on VBAP Technology: Evaluating the Impact of Different Virtual Speaker Arrays

With the development of 3D immersive audio in virtual/augmented reality and other applications, it has become important to accurately render the orientation of point source using binaural in real world acoustic spaces. The method for spatial binaural rendering is to convolve the target signal with Head-related transfer function (HRTF). Therefore, VBAP method is introduced into spatial binaural rendering. The relationship between sound source and virtual loudspeaker array is constructed by vector base amplitude panning (VBAP) technology, and binaural spatial signal is obtained by convolution of virtual loudspeaker signal with HRTF. The VBAP method can flexibly select the position of virtual loudspeaker, in fact, the structure of the virtual loudspeaker array and the number of loudspeakers have a crucial role in the binaural rendering performance. This paper designs a spatial binaural rendering scheme based on VBAP technology, and the effects of different array structures on binaural rendering are analyzed. First, the rectangular room model is constructed in this paper. The first order reflection nodes of the wall are determined by the image source method. The delay and attenuation of direct sound and reflection of each order are simulated by the physical position of each reflection node of the wall. Secondly, the virtual loudspeaker array is constructed, and the signal of the virtual loudspeaker array is calculated using VBAP technology. Finally, the virtual loudspeaker signal is convolved with HRTF to obtain binaural spatial signal. In order to evaluate the rendering performance of different virtual loudspeaker arrays, this paper presents three configurations of virtual loudspeaker arrays, namely uniform sampling arrays, layered sampling arrays and Fibonacci sampling arrays. The rendering performance of different array configurations is evaluated by steady-state source experiments as well as subjective and objective evaluation.