On Mar 16, 5:02 pm, Jerry Avins <j...@[EMAIL PROTECTED]
> wrote:
> Why do you see a flaw at the boundary? Consider a Hamming window applied
> to a group of samples you want to feed to an FFT. Since the FFT imposes
> periodicity, the proper way to compute the window makes the first and
> the _sample_after_the_last_ zero, matching that periodicity. It's a
> subtle point, but it can gain a few dB of sidelobe suppression.
That was not a reference to WITHIN, it was a reference to the flawed
definition of BETWEEN for generic signed or unsigned numbers as ``1+
WITHIN''. That's only valid as a BETWEEN for values that have a hole
on the number wheel between their largest and their smallest number.
So it would work for ASCII7 or Latin-1 or in a 32-bit Forth UTF-32
Unicode, or for addresses in a VIC-20, but not for, say, a 16-bit cell
address where TRUE and 0 are both valid addresses.
|
