首页|基于线性预测技术的语音基音检测

基于线性预测技术的语音基音检测

扫码查看
针对语音基音检测易受加性背景噪声和共振峰干扰的问题,提出了一种基于线性预测技术的语音信号基音检测算法。该算法在维纳滤波基础上,利用线性预测(LPC)技术得到预测残差信号,再对其做自相关(ACF)和平均幅度差(AMDF),得到基音的检测结果。其检测效果比单一自相关函数法和平均幅度差函数法有明显的改善。
Pitch Detection of Speech Based on Linear Prediction and Endpoint Detection
A kind of pitch detection algorithm based on linear prediction is proposed in this paper,Aiming at the problem of speech pitch detection is easily affected by additive background noise and resonance peak interference. The algorithm based on wiener filter,gets the predicted residual signal by using the technology of LPC,Then calculates the ACF and AMDF value of the residual signal to estimate the pitch. The results of comprehensive application with ACF and AMDF have obvious improvement than the single ACF and AMDF.

speech signalpitch detectionlinear predictionACFAMDF

王琛、姜占才

展开 >

青海师范大学计算机学院,西宁810001

语音信号 基音周期 线性预测 自相关 循环平均幅度差

2015

电脑开发与应用
中国北方自动控制技术研究所

电脑开发与应用

影响因子:0.265
ISSN:1003-5850
年,卷(期):2015.(3)
  • 1
  • 6