0.5 shift in spatial frequency array

[sevbogae]

Hi,

In the array where you put the spatial frequency as a function of the Nyquist frequency (x-axis for MTF plot), you put an additional 0.5 shift. I assume this is to center it in the bin? But since we are working with frequencies at this point, this seems incorrect to me. The first element in the spatial frequency is now 0.5 Nyquist frequencies instead of 0 Nyquist frequencies. Personally, I would omit the 0.5. Is there a specific reason you added it?

I am talking about this formula:
sf[i] = 2.*(i+0.5)/ft_len;

Thanks.

[emx77]

I am not sure about this. I consider the MTF plots to be histograms and work therefore with the bin centres. Also SF=0 is a difficult point to work with in some other FFT methods and normalization I tried at that time.