基于图信号处理的MVDR波束形成多通道语音增强

Multi⁃channel speech enhancement of MVDR beamforming based on graph signal processing

新兴的图信号处理技术已经在不规则且复杂的信号处理上取得了许多研究成果,但基于图信号的多通道语音处理研究尚处于起步阶段。文中提出了一种基于图信号处理的最小方差无失真响应(Minimum Variance Distortionless Response,MVDR)波束形成方法来进行多通道语音增强。首先,考虑阵列之间位置关系与相位差的联系,定义了一种新的多通道语音邻接矩阵权值,在此基础上构建多通道语音图信号。然后,通过联合通道间图傅里叶变换和帧内图傅里叶变换,定义了多通道图傅里叶变换,在此基础上提出了图频域的MVDR方法来进行多通道语音增强。仿真表明,与传统数字信号处理中的MVDR波束形成方法相比,文中提出的基于图信号处理的MVDR波束形成方法明显提升了增强语音的质量和可懂性。

The emerging graph signal processing technology has achieved numerous research achievements in irregular and complex signal processing,but the processing of graph signals for multi⁃channel speech signals is still not fully developed.A minimum variance distortionless response(MVDR)beamforming method based on graph signal processing for multi⁃channel speech enhancement is proposed.Firstly,considering the connection between the positional relationships and the phase difference between the arrays,a new multi⁃channel speech adjacency matrix weight is defined,and on this basis,a multi⁃channel speech graph signal is constructed.Then,the multi⁃channel graph Fourier transform is defined,meaning the combination of the graph Fourier transform between the channel and the intra⁃frame graph Fourier transform.Based on this,the definition of minimum variance distortionless response beamforming is proposed in the frequency domain of the graph.Compared with the traditional MVDR beamforming method in digital signal processing(DSP),experimental data show that the speech enhancement significantly improves the speech quality and the intelligibility of the enhanced speech.