摘要:
在线性预测系数(LPC)的基础上,借鉴美尔倒谱系数(MFCC)计算方法,对LPC进行美尔倒谱计算,得到一种新的特征参数:线性预测美尔倒谱系数(LPMFCC)。在Matlab7.0平台上实现一个基于隐马尔可夫模型(HMM)的说话人识别系统,分别用LPMFCC及其一阶差分、MFCC及其一阶差分和基于小波包分析的特征参数(WPDC)及其一阶差分作为识别参数进行对比实验。结果表明,以LPMFCC作为特征参数的系统具有较高的识别率。
关键词:
线性预测,
美尔倒谱系数,
说话人识别
Abstract:
This paper obtains a new feature factor which is called linear prediction Mel Frequency Cepstral Coefficient(MFCC) based on the Linear Prediction Coefficient(LPC) and MFCC. In order to verify the validity of the LPMFCC, a speaker recognition system based on the HMM model using the Matlab7.0 is established. The speech recognition rate of LPMFCC and the other two feature factors are tested. The result suggests that the new feature factor is not only efficiency, but also better than other two feature factors. It can achieve high recognition rate.
Key words:
linear prediction,
Mel Frequency Cepstral Coefficient(MFCC),
speaker recognition
中图分类号:
张学锋, 王芳, 夏萍. 融合LPC与MFCC的特征参数[J]. 计算机工程, 2011, 37(4): 216-217.
ZHANG Hua-Feng, WANG Fang, JIA Ping. Feature Factor Fused on LPC and MFCC[J]. Computer Engineering, 2011, 37(4): 216-217.