Single Channel Speech Enhancement Using Adaptive Soft-Thresholding with Bivariate EMD

Figure 3

Frame-wise noise variance (blue line) and speech energy (red line) of corresponding IMFs computed from imaginary and real parts, respectively.