摘要
谱减法是目前受到广泛研究的语音增强算法之一,但谱减法分辨率固定及存在音乐噪声残留的问题,限制了谱减法的应用推广。本文针对以上问题,利用小波变换的多分辨率特性及谱减法运算量小、处理速度快的特点,提出了一种基于小波变换和改进的Berouti谱减法的改进语音增强算法。实验结果表明,该算法在抑制噪声,增强语音信号的可懂度及鲁棒性方面优于传统的谱减法。
Spectral subtraction is one of the most widely studied speech enhancement algorithms,but the problem of spectral subtraction resolution and music noise residue limits the application and promotion of spectral subtraction.Aiming at the above problems,a new speech enhancement algorithm based on wavelet transform and improved Berouti spectral subtraction is proposed based on the multi-resolution characteristics of wavelet transform and small amount of spectral subtraction and fast processing speed.The experimental results show that the algorithm is superior to the traditional spectral subtraction in reducing noise and enhancing the intelligibility and robustness of speech signals.
作者
卢景
赵风海
LU Jing;ZHAO Fenghai(Tianjin Key Laboratory of photoelectric sensor and sensor network technology,College of Electronic Information and Optical Engineerin,Nankai University,Tianjin 300350,China)
出处
《电声技术》
2018年第12期8-12,69,共6页
Audio Engineering
关键词
谱减法
小波变换
语音增强
spectral subtraction
wavelet transform
speech enhancement
