码激励线性预测快速算法

Fast Algorithms for the Code Excited Linear Prediction Speech Coding

对语音编码的两大类即波形编码和参量编码的优缺点进行了分析,重点对码激励线性预测进行了研究.首先提出了感觉加权滤波器的等效变换方法,然后提出了一种码本部分重叠分割法,计算机模拟实验证明,采用上述方法,可以得到很好的合成音质,并使运算速度显著提高。

The advantages and disadvantages of the two classes of speech coding viz. the waveform coding and parameter coding, are first analysed. Then the mingled coding technology is introduced, with an emphasis on the code excited linear predictive (CELP) coding. The equivalent transform of CELP is put forward and then an improved method of code book formation is introduced. It can reduce the amount of work of calculation in the search and cut down the storage required.