Hello,
I was wondering if it matters where to place the zeropad after windowing a time
signal (to prepare for fast FT).
More specific, I am use a 418 point Hann window to window a time signal and then
pad to 512-points before FT transforming. Will the FT be differ when I place:
1) 52 pads before AND after the window
2) 104 points before the window
3) 104 points after the window
--
So far I checked the radix-2 transform pairs for a 10point signal padded with 6
zeros and a 12point signal with 4 zeros. All resulting in the same pairs and
transform (assuming that a 0 transformed with signal[x] is the same as a
transfrom with signal[x] and a zero).
However, when I compare the absolute values of the different padded FT's
there is a difference in the order of 10^-16, is this something significant?