Three-dimensional audio parametric encoding based on perceptual characteristics of spatial cue1

Abstract

Abstract. Three-dimensional sound effects require a considerable number of sound channels, causing audio-visual spatial orientation sense distortion under circumstances in which the code rate is restricted by the limitation of transmission channel bandwidth and storage capacity. As a result, existing 3D audio systems are incompatible with real-time broadcasting and home theatre applications, severely limiting the application and development of 3D audio systems. By investigating the mechanism of orientation parameters perceptual redundancy, this paper studied 3D spatial orientation cue perceptual characteristics, established an orientation cue perceptual model, developed a heterogeneous quantification table accordingly, and controlled the differences between each quantified value below the quantitative value perception threshold. Using this method, only the information perceptible to the human ear was quantified and perceptual distortions were minimized. The experimental results revealed that, compared to the SLQP method, the quantified bit of the proposed method was reduced by 8.66% in low resolution and 65.23% in high resolution. In addition, the accuracy of this method was higher than that of the SLQP method, enabling better alignment with human perceptual characteristics.