多线谱干扰下的盔内语音通信降噪算法研究

Research on noise suppression for communication in helmet with multiple line-spectrum interference

航天服头盔内语音通信系统工作时,受到来自航天服自身的风机、泵、通风气流所产生的噪声影响,通话质量下降严重。传统降噪方法能有效抑制宽带噪声,但对线谱干扰的抑制较弱。文章以此为出发点,开展了多线谱干扰下的语音降噪算法研究,设计了一种自适应线谱检测方法,可有效识别线谱,定位其所处频带,在抑制宽带噪声的同时,进一步抑制线谱干扰。文章结合盔内实录噪声数据对算法进行了验证,结果显示,相对于传统降噪方式,文中方法对多线谱干扰的抑制能力更强,语音客观语音质量评估得分提高约6%。

Voice communication system in aerospace helmet is affected by the noise generated by the fan,pump,and ventilated air flow in space suit.Traditional noise suppression methods work well to suppress on broadband noise,but they cannot suppress line-spectrum interference effectively.In this paper,we carry out the research on noise suppression algorithm under multiple line-spectrum interference,and design an adaptive line-spectrum detection method,which can effectively identify the line spectrums,and further suppress the line-spectrum interference while suppressing the broadband noise.Compared with the traditional noise suppression method,the proposed method works well on suppressing line-spectrum interference,and the PESQ score is improved by about 6%.