A Method to Estimate the Parameters of the Vocal－Tract Area Function from Formant Frequencies of Speech Signal

俞振利; 张礼和; 曾尚璀

您当前的位置：

首页 >

文章列表页 >

A Method to Estimate the Parameters of the Vocal－Tract Area Function from Formant Frequencies of Speech Signal

更新时间：2025-12-08

- A Method to Estimate the Parameters of the Vocal－Tract Area Function from Formant Frequencies of Speech Signal
- Acta Electronica Sinica Issue 7, (1997)
- 作者机构：
  
  杭州大学电子工程系
- 作者简介：
- 基金信息：
- DOI：
  CLC： TN912.3
- Published：1997
- 稿件说明：
移动端阅览
[1]俞振利,张礼和,曾尚璀.从语音信号的有限个共振峰频率估计声道面积参数的一个方法[J].电子学报,1997(07):32-37.

俞振利, 张礼和, 曾尚璀. A Method to Estimate the Parameters of the Vocal－Tract Area Function from Formant Frequencies of Speech Signal[J]. Acta Electronica Sinica, 1997, (7).
[1]俞振利,张礼和,曾尚璀.从语音信号的有限个共振峰频率估计声道面积参数的一个方法[J].电子学报,1997(07):32-37. DOI：

俞振利, 张礼和, 曾尚璀. A Method to Estimate the Parameters of the Vocal－Tract Area Function from Formant Frequencies of Speech Signal[J]. Acta Electronica Sinica, 1997, (7). DOI：

摘要

本文介绍从语音的有限个共振峰估计声道面积函数模型参数的扰动实现方法．在Ｓｃｈｒｏｅｄｅｒ的声管喇叭方程的扰动解法基础上，采用改进的声道面积函数模型；针对不同共振模式之间的交叉影响，采用了交叉敏感扰动矩阵关系式，及交叉扰动敏感矩阵的测试方法．设计了用递归式扰动法实现声道面积参数估计的算法．利用“共振峰─扰动”映射码书作为语音共振峰匹配最佳的初始面积扰动矢量，井用多级扰动目标和自适应增量控制，可有效地实现从有限个语音目标共振峰逆向估计声道面积函数．

Abstract

A method for estimating the parmeters of the vocal-tract area function from the first finite formants frequencies of speech signal is propsoed. An improved vocal-tract area function model is suggested based on the perturbation approach to the solution of the Webster＇s wave equation of the acoustic resonance tube by Schroeter. By considering the cross-influence of the different resonance modes of the perturbation

a cross-sensitivity perturbation relation in matrix form is presented. A method to measure the cross-sensitivity matrix is proposed. An iterative perturbation algorithm for determining the area function parameters from formant targets is developed. By looking up a ＇formant-to-perturbation ＇mapping codebook

an optimal initial perturbation vector for given formant targets is selected.A step-by-step sub-target perturbation procedure and adaptive factor for perturbation increment are adopted.An effective realizaion of the estmation of the vocal-tract area function is illustrated with comparison to the published data.

关键词

Keywords

references

Views

176

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

An Auditory Feature Extraction Algorithm Based on γ-Tone Filter-Banks

A New Consolidated Speech VBR Coding Method

Speech Enhancement Based on Minimum Mean-Square Error Short-Time Spectral Estimation and Residual Noise Reduction

Speech Enhancement Algorithms Using MMSE Estimation with Interframe Constraints

Chaos and Fractal Theories for Speech Signal Processing

Related Author

程光明

CHENG Guang-ming

WANG Xue

QIAN Zhi-hong

WANG Yue

ZHENG Bao-yu

Related Institution

Jilin University

State Intellectual Property Office of the PRC

南京邮电学院165信箱

Dept. of Electronic Engineering, Tsinghua Univ.

清华大学电子工程系北京 100084

⁰