摘要
基于广义旁瓣抵消器(generalized sidelobe canceller,GSC)算法的麦克风阵列语音增强技术已得到广泛研究,但由于其通常需传统的声源定位方法提供声源方位,语音信号信噪比(SNR)低时声源定位精度将明显下降并影响到语音增强效果.提出了一种新的麦克风阵列语音增强方法,该方法在GSC中引入可调波束形成器估计声源方位以抑制背景噪声影响.不同类型背景噪声下的实验室语音增强结果表明了该方法的有效性.
Generalized sidelobe canceller （GSC） has been widely investigated in microphone array speech enhancement.However,the traditional source localization methods are usually adopted to obtain the direction of source,which is needed for the GSC algorithm.Under relatively low signal noise ratio （SNR） the performance of the GSC will degrade due to the decreasing precision of the source localization.This paper presents a new method for the GSC based microphone array speech enhancement,which uses an adjustable beamformer （ABF） to estimating the direction of the speech source to suppress the background noises.Experimental speech enhancement results under different type of background noises validated the effectiveness of the proposed method.
出处
《厦门大学学报（自然科学版）》
CAS
CSCD
北大核心
2013年第2期186-189,共4页
Journal of Xiamen University：Natural Science
基金
国家自然科学基金项目(61105026)
福建省高校产学合作科技重大项目(2010H6022)
厦门市科技计划项目(23502Z20113008)
关键词
语音增强
麦克风阵列
可调波束形成器
广义旁瓣抵消器
speech enhancement
microphone array
adjustable beamformer
generalized sidelobe canceller
