Advances in Acoustics and Vibration

Research Article

Subband DCT and EMD Based Hybrid Soft Thresholding for Speech Enhancement

Figure 3

The illustration of EMD. A noisy speech signal at 10 dB SNR and its first 8 IMFs out of 14, plus a residue signal which can be observed to be close to a constant.
765454.fig.003