how to fix broad band in fft?

Hy guys.

I'm trying to obtain the fft of a array. however the result are coming out to be braod band, i expected to get bands with a width of 1 pixel.

any suggestion for this problem.

Thanks and Regards

code:

clear all
clc
z=50;
a=ones(z,z);
for i=1:z
    for j=2:z
        a(i,j)=a(i,j-1)+1;
    end
end
a = a.'*a;
subplot(121)
img(abs(a))
subplot(122)
img(abs(log(fftshift(fft2(a)))))

% this is a small example what is happening in my results

[Catalytic]

Please do not post unformatted code. It makes reading and running the code harder.

clear all
clc
z=50;
a=ones(z,z);
for i=1:z
    for j=2:z
        a(i,j)=a(i,j-1)+1;
    end
end
a = a.'*a;
subplot(121)
img(abs(a))
subplot(122)
img(abs(log(fftshift(fft2(a)))))

Also, the code does not run because the img() function has not been provided. However, it is not clear why you would expect a 1-pixel spectrum. The only image with a 1-pixel spectrum is a uniform image.

[Rabih Sokhen]

hy Catalytic, hope you are doing well,

i forgot to replace img with pcolor, it's a function that's i created.

i expected that's the band will not be broad , only 1 pixel as width , because the analytique equation should fit my band at 100% to extract to amplitude and the phase

any idea how to make my band finer ?

thank you in advance

[Catalytic]

"i expected that's the band will not be broad , only 1 pixel as width , because the analytique equation should fit my band at 100% to extract to amplitude and the phase"

I don't really understand that part. The fft2 and ifft2 form an invertible 1-1 pair. Because a=ones(50) produces a 1-pixel wide spectrum, no other a can do so. Otherwise fft2 would not be invertible.

[Rabih Sokhen]

this is a example of my work, where "a" is periodic array, however wen i use abs(fftshift(fft2(a))) i obtain bands with a width.

is this normal in matlab ?