摘要
2维PV编码算法是一种将线性预测编码、SOM神经网络2维矢量编码以及Huffman编码相结合的语音信号编码算法。为了在保证译码恢复的语音质量良好的前提下,进一步减小编码的压缩率,以减小语音信号的存储空间,提出一种增加2维PV编码中矢量量化维数的高维PV编码算法。并利用Matlab软件编程进行2维、4维和8维PV算法的语音信号编解码实验。实验结果表明,在保证译码恢复声音质量良好的条件下,增加2维PV编码算法的量化矢量维数,可以减小码率,其中8维PV编码算法的码率最小为5.94 Kb/s,小于采用ADPCM编码算法的波形编码标准G.721的码率32 Kb/s(波形编码的最小码率),甚至小于采用LD-CELP编码算法的混合编码G.728的码率16 Kb/s。文中提出的编码算法在语言压缩编码方面将具有较高的研究价值和很好的应用前景。
The 2D PV coding algorithm is a speech signal encoding algorithm,in which the linear prediction coding,SOM neural network 2D vector coding and Huffman coding are combined.In order to further reduce the compression rate of coding and save the storage space of speech signal,a high-dimensional PV coding algorithm to increase the dimension of vector quantization in 2 dimensional PV coding is proposed while guaranteeing the speech quality of decoding recovery.The speech signal codec experiments of 2D,4D and 8D PV algorithms are carried out with Matlab software programming.The experimental results show that the code rate can be reduced by increasing the quantization vector dimension of the 2D PV coding algorithm while maintaining the perfect sound quality of the decoding recovery;the code rate(5.94 Kb/s)of the 8D PV coding algorithm is lowest,which is less than 32 Kb/s of the waveform coding standard G.721 got by ADPCM encoding algorithm,and even less than 16 Kb/s of the mixed coding G.728 got by LD-CELP encoding algorithm.The proposed coding algorithm has high research value and wide application prospect in language compression coding.
作者
杨超
刘云飞
徐向旭
朱弘
芮天予
YANG Chao;LIU Yunfei;XU Xiangxu;ZHU Hong;RUI Tianyu(Naval Aviation University,Yantai 264001,China;Naval Aviation University Qingdao Branch,Qingdao 266041,China;Unit 92635 of the PLA,Qingdao 266041,China;Unit 91602 of the PLA,Shanghai 200082,China;Unit 92877 of the PLA,Zhoushan 316200,China)
出处
《现代电子技术》
北大核心
2019年第11期32-35,共4页
Modern Electronics Technique
基金
国家自然科学基金项目(61402529):多层多域智能光网络安全路由算法与协议研究~~
关键词
语音编码
编码算法
矢量编码
预测误差编码
压缩编码
神经网络
speech coding
coding algorithm
vector coding
prediction error coding
compression coding
neural network
