New 3D Coding Approach | Concentric Spheres Spatial Quantization (CSSQ) Method
3D audio encoding based on spatial perception features, fine-tuned in 3D space. and set with non-uniform spatial quantization points.
An interesting article recently published in the China Communications journal presents a “new three-dimensional (3D) audio coding approach” called the “concentric spheres spatial quantization (CSSQ) method”, which differentiates itself from standard directional audio coding by additionally quantifying distance side information in its parameters.
From the article
Comparing the above 3D audio coding techniques, we found that most of them can express the direction of sound source by elevation and azimuth or ICLD and ICTD. However, they all have limitations in recording and expressing the distances of the sound sources, since they only use the intensity of audio signal (the overall received sound level) to express the distances of sound sources….
…our 3D audio spatial quantization method is no longer a sphere, but multiple spheres with the same spherical center. And those different spheres have the different radius, which are used to describe distances of the sound sources. And our 3D audio spatial quantization method is a non-uniform quantization, which match well with the sensitivity of human spatial perception. The concentric spheres spatial quantization method is illustrated as figure 5 (listed below).
According to this article, in addition to improving spatial perceptual quality with help from the HRTF database, the CSSQ method also decreased the bit rate (compared with DirAC) by nearly 10% while coding compared to the “spatial squeeze surround audio coding” (S3 AC).
You can view the full article here.
C. Yang, R. Hu, X. Wang, Y. Yang, M. Zhang and W. Chen, "3D audio coding approach based on spatial perception features," in China Communications, vol. 14, no. 11, pp. 126-140, Nov. 2017.