资 源 简 介
文章介绍的说话人识别系统,采用能够反映人对语音的感知特性的9(:频率倒谱系数(9(: <+(=>(2/?8(’0+):
81(..-/-(20@,9<88)作为特征参数,同时考虑到特征参数各维分量对于不同说话人的区分程度,采用加权的办法进行矢
量量化。-This paper introduces the speaker recognition system used to reflect the people s perception of voice characteristics 9 (:-Frequency Cepstral Coefficients (9 (: <+(=>( 2 /? 8 ( 0+):
81 (..-/-( 20 @, 9 "88) as the characteristic parameter, taking into account the characteristic parameters for different components of each dimension of the distinction between the speaker level, using weighted approach to vector
Quantization. .