电子学报 ›› 2016, Vol. 44 ›› Issue (3): 747-752.DOI: 10.3969/j.issn.0372-2112.2016.03.037

• 科研通信 • 上一篇    下一篇

基于最小控制GARCH模型的噪声估计算法

孟宪波, 鲍长春   

  1. 北京工业大学电子信息与控制工程学院语音与音频信号处理实验室, 北京 100124
  • 收稿日期:2014-10-08 修回日期:2015-04-28 出版日期:2016-03-25
    • 通讯作者:
    • 鲍长春
    • 作者简介:
    • 孟宪波 男,1987年出生,河北承德人,北京工业大学硕士研究生,主要研究方向为语音与音频信号处理. E-mail:mengxianbo@emails.bjut.edu.cn
    • 基金资助:
    • 国家自然科学基金 (No.61471014)

Noise Estimate Algorithm Based on Minima Controlled GARCH Model

MENG Xian-bo, BAO Chang-chun   

  1. Speech and Audio Signal Processing Laboratory, School of Electronic Information and Control Engineering, Beijing University of Technology, Beijing 100124, China
  • Received:2014-10-08 Revised:2015-04-28 Online:2016-03-25 Published:2016-03-25
    • Supported by:
    • National Natural Science Foundation of China (No.61471014)

摘要:

MCRA(Minima-Controlled Recursive Averaging)方法是经典的噪声估计算法,然而在语音段MCRA方法存在不能对噪声功率谱进行有效更新的问题.针对这一问题,本文利用广义自回归条件异方差(Generalized Autoregressive Conditional Heteroskedasticity,GARCH)模型在时频域对噪声信号建模,在MCRA算法原理的基础上,提出了基于最小控制GARCH模型的噪声估计算法,实验结果表明,本文所提的噪声估计算法能够更为准确估计噪声功率谱,将该算法应用到语音增强中能够获得到较好的语音增强效果.

关键词: 噪声估计, GARCH模型, MCRA算法, 语音增强

Abstract:

Considering the problem that the typical MCRA (Minima-Controlled Recursive Averaging) noise estimate algorithm fails to update the power spectrum of noise effectively when the speech is present,so this paper proposes a noise estimate algorithm based on minima controlled GARCH model.The noise signal is modeled as a GARCH process in time-frequency domain and then the proposed noise estimate algorithm is achieved combined with the basis of the framework of MCRA method.Experimental and testing results indicate that the proposed algorithm can estimate the spectrum of noise more accurately compared with the reference methods.When the proposed algorithm is applied into speech enhancement,a better performance can be achieved as well.

Key words: noise estimate, GARCH model, MCRA algorithm, speech enhancement

中图分类号: