Contents Previous Next Subchapters Current Chapters-> fourier kalman_bucy filter mlmode_filter arcov fspec dpss wavelet iwavelet fnbut fncheb fnbpole fncpole fn2cbp fn2clp fc2dig fn2dbp fn2dlp Parent Chapters-> Omatrix6 filtering dpss Search Tools-> contents reference index search

Discrete Prolate Spheroidal Sequences
 Syntax `dpss(`N`, `m`, `w`)` See Also fspec

Description
Returns discrete prolate spheroidal sequences, where N an integer scalar specifying the number of points in each sequence, m is an integer scalar specifying the number of sequences to return, and w is a real or double-precision scalar specifying the fractional bandwidth of the sequences such that `0 < w < .5`, with a suggested value of `4./N.` The sequence of order `i` is placed in the (i+1)-th column of the return value for `i = 0, ... , m - 1`. The return value has m columns, N rows, and the same type as w. ``` ```These sequences are used as tapers for spectral estimation (see fspec ). There are two important factors corresponding to a taper: its bandwidth (how wide a frequency range it allows leakage over) and its rejection (how much energy is reduced when it is outside that frequency range). If w is small, the frequency range of the allowed leakage will be small. On the other hand, a smaller w leads to more energy leaking outside the allowed range (poorer rejection).

Example ```      N = 128      m = 1      w = 4d0 / N      S = dpss(N, m, w)      gtitle("Discrete Prolate Spheroidal Window")      gplot(seq(N)/double(N), S)       ``` returns the following plot: ``` ```