Subband DCT and EMD Based Hybrid Soft Thresholding for Speech Enhancement

Figure 7

Waveform of (a) the clean speech, (b) the noisy speech corrupted with white noise at 10 dB SNR, (c) the recovered speech after soft thresholding with subband DCT, and (d) the overall recovered speech of the 𝑈_EMD method.

(a)

(b)

(c)

(d)