Designing a fir filter and how to apply it to a signal

Hello, I have to design a bandstop fir filter for a college project using hamming window with passband frequencys of 0.15 and 0.3 and apply it to a signal x = sin(2*pi*0.2*t) + sin(2*pi*0.1*t), 0<t<128. So far I wrote this:

fs = 10;

fp1 = 0.15;

fp2 = 0.3;

f1 = 0.2;

f2 = 0.1;

t = 128;

n = 0:1/fs:t;

x = sin(2*pi*f1*n) + sin(2*pi*f2*n);

Wn=[fp1 fp2];

Ham = fir1(30, Wn, 'stop');

[H,w] = freqz(Ham);

y = filter(Ham,1,x);

dB = mag2db(abs(H));

figure(1)

plot(w/pi, dB);

xlabel('Frequency');

ylabel('Magnitude (dB)');

figure(2)

plot(n, y);

title("Output y(n)",'fontsize',14);

xlabel('n');

ylabel('Y(n)');

However, my output signal y is almost the same as my input signal x, when I believe all that should be left is sin(2*pi*0.1*t). If I use bandstop function as "bandstop(x,Wn,fs)" the results are fine, but I cant replicate them using the upper method. If anyone knows what I am doing wrong I would appreciate any tips and advices.

Answer 1

[Star Strider]

You need to normalise the stopband frequencies by the Nyquist frequency.

Try this:

Wn=[fp1 fp2]/(fs/2);

That worked when I tried it.

[Nediljko Ivisic]

Seems to work the trick and my output is fine now, but the plot of filter magnitude looks a bit weird. I cant read out the passband frequencys from it now. Am I ploting it wrong?

[Star Strider]

The freqz call needs to change a bit. You also need to increase the filter order, since normalising by the Nyquist frequency changes (decreases) the passband frequencies, and use filtfilt to do the actual filtering.

Change to these:

Wn=[fp1 fp2]/(fs/2);
Ham = fir1(90, Wn, 'stop');
[H,w] = freqz(Ham,1,2^16,fs);

and this:

figure(1)
plot(w, dB);

and this:

y = filtfilt(Ham,1,x);

Experiment with different filter orders to see those effects.

.