Re: Frequency domain integration and differentiation

Sorry about the triple post, my browser confused me.

The time lag is about 6 out of 256. I've only considered
whole delta-t shifts but will consider fractional delta-t shifts.

I haven't been splitting my data and I'm not using a filter. I'm just passing a perfect period of data, being careful to not duplicate end points.

I'm aware that for real world problems, the data should be
filtered according to nyquist. But for this contrived problem, no filter should be needed, or am I missing
your point.

Finite differences? I'm aware of finite differences
in the context of solving PDEs with linear algebra,
but not with FFTs or discrete fourier analysis.

Is "the filter" built with finite differences? Then
I could understand a fractional-delta-t shift, but I'm
not using a filter. Not yet anyway.

Perhaps again, I'm missing your point.

Thanks,

Rob
.