0. Ftp the data file `KTX` from the ftp site "ftp://iclab.ac.kr/DSP" to your folder.
In Matlab command window, load the data file "KTXSignal`. Check the data variable. The sampling rate of the signal is 100/sec.
>> load KTXSignal
>> load KTXSignalWithNoise
1. Plot `KTXSignal` along the actual time axis `t`.
2. Plot `KTXSignalWithNoise` along the actual time axis `t`.
3. We want to remove the noise from `SigWithNoise` by an average filter . We denote the filtered signal as . For an example, when L=10, in Matlab
1) Check for the length of y[n]. Let p be the length of y[n]. Determine the length of y[n] in terms of L and M=length(SigWithNoise).
2) When L=10,
● plot y[n] in discrete index n and also y in time scale.
>> plot(t,y); % length of t should be adjusted to the length of y.
● plot the frequency response of `h[n]`. In Matlab, you can use `freqz`
>> plot(theta,abs(H)); % for ploting magnitude of `h[n]`.
>> plot(theta,angle(H)); % for ploting angle of `h[n]`.
3) Repeat 2) for L=2,5,10,15,20. Observe the following as L changes.
● Degree of noise removing and smoothing y[n].
● The transient time.
● Determine the optimal L.
● Shape of abs(H(theta)). (sharpness of cutoff frequency, magnitude of side lobes)
=> You can overlay `abs(H(theta))` for L=2,10,20. Then you can recognize the shape change of abs(H(theta)).