面向多声道三维音频的和差压缩编码技术

Three-channel dependent mid/side coding framwork for multichannel 3D audio

分析了传统基于单声道对的和差(M/S)编码技术,针对三维音频系统的空间特性将其拓展为基于三声道的和差编码技术(3D-M/S)。并提出一种基于3D-M/S的音频编码框架,使其能够压缩任意声道配置的三维音频系统。实验结果表明与独立声道编码相比,所提方法能够在相同码率下提高约25%的客观质量。并且在同等质量下,所提方法引起编解码复杂度的增加约为基于PCA编解码方法的1/3。

The conventional mid/side coding method is discussed and generalized to a three-channel dependent M/S cod- ing （3D-M/S） method. Moreover, a codec framework based on 3D-M/S is proposed which enables the 3D-M/S to work for three-dimensional audio systems with more than three channels. It is experimentally shown that proposed framework gets about 25% objective quality improvement compared with independent coding, and only increases about 1/3 complexity of the PCA method.