Filtering in the complex frequency domain is not the best option because the Fourier transform contains ‘positive’ and ‘negative’ frequencies (visible as such after using the fftshift function). It is necessary to do the same operation on both the positive and negative frequencies to get the correct result.
The best solution is to convert your frequency domain filter:
exp(-pi*0.011.*f)
to the time domain, for example:
syms F f t
Fcnf(f) = exp(-pi*0.011.*f); % Frequency Domain
F = sym(100); % Choose A Frequency
Fcnt = int((Fcnf(-f) + Fcnf(f))*exp(-1i*2*pi*f*t), f, -F, F) % Inverse Fourier Transform
Fcnt = simplify(Fcnt, 'Steps', 500) % Collect Terms
figure
fplot(Fcnt, [-100 100])
and then filter in the time domain.
I leave the details to you.
