- #1

- 19

- 0

## Main Question or Discussion Point

I have about 40 tabs open on this right now and something important is slipping my grasp. I know this has been covered a million and a half times, but for some reason I cannot seem to find a straight answer (or more probably realize and understand it when I see it).

When I take the Continuous Fourier transform of a function, I get back a very nice magnitude and phase.

When I try the exact same thing with a FFT I get back a very nice magnitude, but my phase is highly erratic. Most of the time, it is just oscillates in between two values every point (eg -[tex]\pi[/tex] and [tex]\pi[/tex] or does this on a slope. I am using the atan2 function.

If I unwrap this, the phase will go to ridiculously huge numbers.

Can someone explain if why this happens and more importantly if and how I can make the discrete phase match the continuous phase?

When I take the Continuous Fourier transform of a function, I get back a very nice magnitude and phase.

When I try the exact same thing with a FFT I get back a very nice magnitude, but my phase is highly erratic. Most of the time, it is just oscillates in between two values every point (eg -[tex]\pi[/tex] and [tex]\pi[/tex] or does this on a slope. I am using the atan2 function.

If I unwrap this, the phase will go to ridiculously huge numbers.

Can someone explain if why this happens and more importantly if and how I can make the discrete phase match the continuous phase?