New CLD Quantization Method for Spatial Audio Coding

In spatial audio coding, spatial parameters, such as CLDs, CPCs, ICCs, are utilized for down-mix and up-mix of the multi-channel audio signals. In the current version of MPEG Surround, an universal quantization table for CLD is applied independent to channel combinations. As intervals between adjacent channels differ in the conventional 5.1ch configuration, this universal quantization scheme causes redundancies in some combinations while insufficiencies in the other combinations. In this paper, we propose a new CLD quantization method based on well-known amplitude panning law and spatial resolution of human perception. By the proposed quantization method, CLD can be represented more efficiently, and therefore, bit reduction and quality enhancement can be achieved.