一种结合G.719编解码器的参数立体声音频编解码扩展方法

An Extension Method of Parametric Stereo Audio Coding Combined with ITU-T G.719 Codec

  • 摘要: 针对ITU-T G.719全频带音频编码标准尚不支持立体声压缩编码的问题,结合G.719编解码器提出并设计基于离散傅里叶变换(discrete Fourier transform,DFT)的参数立体声音频编解码扩展模块. 输入音频信号进行 DFT变换后在子带内提取相应的空间参数,并将左右声道下混为和声道;解码时,对和声道进行能量补充,分别运用等式法和基于Givens旋转的奇异值分解法(singular value decomposition,SVD)进行解码恢复. 测试结果表明,在G.719中合理设计参数立体声编解码可以获得较好的压缩效果,对于平缓的音乐,等式法比奇异值分解法可解码得到稍高的音质,且复杂度较低;对于变化较快的音乐,奇异值分解法可得到相对较好的音质,但复杂度较高.

     

    Abstract: Accounting for ITU-T G.719 codec standard for full-band audio without stereo compression, an extension module of parametric stereo audio codec based on the discrete Fourier transform(DFT) was proposed and designed. Several spatial parameters were extracted from the subband after DFT of the input audio, and then the left and right channels were down-mixed to the sum channel when encoding. To conserve the energy of the original audio, energy compensation for the sum channel is essential before utilizing approaches of equations and singular value decomposition (SVD) based on the rotation of Givens respectively when decoding. Test results demonstrate that the method of equation can get higher quality for modest music in low complexity and SVD sacrifices the high complexity for slightly better quality for fast-changing music.

     

/

返回文章
返回
Baidu
map