Zero Padding Problem with FFT Square Wave Pulse

Hi,

I'm trying to smoothen out the sinc pulse response that I get from the fft of a rectangular pulse. I thought that zero padding should help me to do this but I still end up with a sharp/pointy pulse after everything. Any help would be greatly appreciated.

Lamda1 = .003;

N = 8*pi./Lamda1;

Step1 = .05;

A1 = [-N/2:Step1:(N/2)-Step1];

R1 = zeros(size(A1));

R1(A1>(-N/4)& A1<(N/4)-Step1) = 1;

figure (1); plot(A1,R1)

FD1 =fft(R1)/N;

figure (2); plot(abs(FD1))

figure (3); plot(abs(fftshift(FD1)))

[Honglei Chen]

Could you clarify a little more? I don't know what you mean by smoothen out the sinc pulse response. The reason you see a pulse is because your sample rate is very high. You can try to increase Step1 and decrease Lamda1 and see if that's what you want.

[Matt]

I create the square pulse. No problems there.

The problem occurs when I look at the fft outputs. When I zoom in on Figure (3), the sinc pulse doesn't have the typical sinusoidal curvature. It is pointy, slightly triangular...

I think I need to pad the R1 with more zeros, I just don't know how to do that... I can't really change the limits of A1.

Do you know if the R1 = zeros ((size (A1)); function is correct?

Thanks.

Answer 1

[Honglei Chen]

R1 is correct. If all you want is to do zero padding and then FFT, you can do it by specifying the number of FFT points in fft() function. For example

FD1 = fft(R1,nextpow2(numel(R1)))