16kb/s维数自适应矢量量化话音编码器

A Voice Coder Using the Method of 16 kb/s Demen sion-adoptive Vector Quantization

本文采用维数自适应矢量量化编码的方法,利用分级查表技术,以单片8086为中央处理器,在I/O接口中采用FIFO缓冲技术设计成16kb/s话音编译码器。其特点是在低复杂度的条件下使信比达到了18.5dB,主观非正式试听表明:与64 kb/s话音间的差别不易察觉。

A method of demension-adoptive hierachical vector quantization by table-lookup for the 16 kb/s speech coding is adopted in this paper. The coder-decoder is implemented on a single chip 8086 CUP and FIFO (first in first out) buffer technique is used for I/O interface.On the low complication condition, the resultant signal-to-noise ratio is 18.5dB. The informal listening lest shows that the difference between the 64 kb/s speech and this coder can not be distinguished.