... samples¹

Note that the normalised frequency in Python's scipy is differently defined. In scipy it is annoyingly defined as $f_{\mbox{scipy}} = 2 \frac {F}{F_s}$. So, in other words $f_{\mbox{scipy}}=1$ is the Nyquist frequency instead of $f=0.5$ as normally defined.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... spectrum²

This derivation is loosely based on Proakis and Manolakis (1996)

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... discretisation³

This derivation is loosely based on Proakis and Manolakis (1996)

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.