Adaptive Filters and Applications - Digital Signal Processing: Fundamentals and Applications - page 466

Next, a simulation example is given to illustrate this idea and its results. The noise cancellation
system is assumed to have the following specifications:

• Sample rate ¼ 8,000 Hz
• Original speech data: wen.dat
• Speech corrupted by Gaussian noise with a power of 1 delayed by 5 samples from the noise
reference
• Noise reference containing Gaussian noise with a power of 1
• Adaptive FIR filter used to remove the noise
• Number of FIR filter taps ¼ 21
• Convergence factor for the LMS algorithm is chosen to be 0.01 ( < 1/21).

The speech waveforms and spectral plots for the original, corrupted, and reference noise and for the
cleaned speech are plotted in Figures 10.10A and Figure 10.10B . From the figures, it is observed that
the enhanced speech waveform and spectrum are very close to the original ones. The LMS algorithm
converges after approximately 400 iterations. The method is a very effective approach for noise
canceling. The MATLAB implementation is detailed in Program 10.1.
Program 10.1. MATLAB program for adaptive noise cancellation.

close all; clear all
load wen.dat % Given by the instructor

fs ¼ 8000; % Sampling rate

t ¼ 0:1:length(wen)-1; % Create index array

1

0

-1

0 200 400 600 800 1000 1200 1400 1600 1800 2000

5

0

-5

0 200 400 600 800 1000 1200 1400 1600 1800 2000

5

0

-5

0 200 400 600 800 1000 1200 1400 1600 1800 2000

1

0

-1

0 200 400 600 800 1000 1200 1400 1600 1800 2000

Number of samples

FIGURE 10.10A

Waveforms for original speech, corrupted speech, reference noise, and clean speech.