基于预测编码和矢量编码的语音信号编码算法

Speech signal encoding algorithm based on predictive coding and vector coding

为了在保证译码恢复的语音质量良好的前提下,减小编码的压缩率,以减小语音信号的存储空间,文中提出一种将线性预测编码、SOM神经网络矢量编码以及Huffman编码相结合的语音信号编码算法,并利用Matlab软件编程进行了语音信号编解码实验。实验结果表明,该算法既保留了波形编码算法具有的译码语音质量较好和算法简单的优点,又具有较一般的波形编码算法更小的压缩率,该编码算法的码率达到12.8 kbit/s,小于采用ADPCM编码算法的波形编码标准G.721的码率32 kbit/s(波形编码的最小码率)。

A speech signal encoding algorithm combining the linear predictive coding,SOM neural network vector coding and Huffman coding is proposed in this paper to reduce the coding compression rate on the premise of guaranteeing a good voice quality after decoding recovery,so as to reduce the storage space of speech signals.A speech signal codec experiment was car-ried out by means of Matlab software programming.The experimental results show that the algorithm not only retains the advan-tages of good decoding voice quality and simplicity that the waveform coding algorithm has,but also has a smaller compression rate than the common waveform coding algorithm.The code rate of the coding algorithm reaches 12.8 kbit/s,smaller than 32 kbit/s,which is the code rate of the waveform coding standard G.721 using the ADPCM coding algorithm,and also the mini-mum code rate of waveform coding.