应用于助听器的频域双通道语声增强算法  

Frequency-domain dual-channel speech enhancement algorithm for digital hearing aids

作　　者：陈新漪 原建平[1] 崔杰[1] 理华[1] CHEN Xinyi;YUAN Jianping;CUI Jie;LI Hua(Institute of Acoustics,Chinese Academy of Sciences,Beijing 100190,China;School of Electronic,Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]中国科学院声学研究所,北京100190 [2]中国科学院大学电子电气与通信工程学院,北京100049

出　　处：《应用声学》2024年第4期880-891,共12页Journal of Applied Acoustics

基　　金：国家重点研发计划项目(2020YFC2004003)。

摘　　要：在助听器的日常使用环境中,干扰噪声与期望语声往往来自不同的方向,因此将多通道语声增强技术应用于数字助听器可以有效地降低干扰噪声的影响。自适应的多通道语声增强技术可以在复杂环境中实现指向模式的动态调整,提高助听器的消噪性能。该文设计了一种基于频域广义旁瓣抵消的双通道语声增强系统,利用噪声参考信号与语声参考信号在频域的相关性对系统阻塞矩阵以及传统自适应部分进行改进,防止阻塞矩阵产生较大的语声泄露,且在语声段选择较小的步长而在噪声段选择较大的步长进行自适应权系数更新,保留期望语声同时抑制干扰噪声。仿真实验中,利用多种干扰环境中的两种客观评价指标验证了所提算法的有效性。In the daily use of hearing aids,interference noise and desired speech mostly come from different directions,and the application of multi-channel enhancement algorithms in digital hearing aids can effectively reduce the impact of interference noise.The adaptive multi-channel enhancement algorithm enables dynamic adjustment of directional patterns in complex environments to improve the noise cancellation performance of hearing aids.A dual-channel enhancement system based on frequency-domain generalized sidelobe cancellation is designed,which improves the blocking matrix part and traditional adaptive part by using the correlation between the noise reference signal and the speech reference signal in the frequency domain,so as to prevent the blocking matrix from generating large speech leakage,and the weight factor is updated by selecting a smaller step in the speech segment and a larger step in the noise segment to retain the required speech and suppress the interference noise.In the simulation experiment,two objective evaluation indicators in a variety of interference environments are used to verify the effectiveness of the proposed algorithm.

关 键 词：助听器 双通道语声增强 广义旁瓣抵消 自适应滤波 相关性 

分 类 号：TN912[电子电信—通信与信息系统]