I would like to investigate the effects of filters on a signal. I have
tried different nyquist filters like Sinc and RRC.
I would like to work with non-nyquist filter such as guassian. However, I am
not sure of the choice of 'sigma' for the filter.
The signal is a complex gaussian signal and has a certain variance, say v1.
I normalize this variance, so that the variance of real & imag part are both
equal & the variance of the scaled signal is 2 ( 1 of real & 1 of imag ).
Should I choose a gaussian filter with sigma(std.deviation) = sqrt(2) ? Or
If I don't change the variance of the original signal , what will the new
filtered signal represent?
I hope my doubts are clear !