hi,
To efficiently use FFT, you need to know the sampling frequency of your signal
So you can proceed like the following : 1) You construct your rectangular pulse :
Fs=100; % Sampling frequency
t=-1:1/Fs:1;
y=zeros(size(t));
% Rectangular pulse
for n=1:length(t)
if t(n)>-0.5 && t(n)<0.5
y(n)=1;
end
end
plot(t,y)
axis([-1 1 -1 2])
grid on,
2) You compute the spectrum, the instructions here are available for any given signal :
% Discrete Fourier Transform
L=length(y);
N=ceil(log2(L));
fy=fft(y,2^N)/(L/2);
f=(Fs/2^N)*(0:2^(N-1)-1);
figure, plot(f,abs(fy(1:2^(N-1))));
Theoretically, you have to know what to expect : Take a look at this page : http://demonstrations.wolfram.com/RectangularPulseAndItsFourierTransform/
Let us try to see if the numeric result matches the theory :
G=sinc(pi*f);
figure, plot(f,G);
Does it look the same ?
