Window Functions and Spectral Leakage - Signal Processing of Random Physiological Signals - page 189

1 dB

1.2

-20

1.0

-40

0 T

-n bins 0 n bins

(a) (b)

FIGURE 16.10 : Seventy-five percent Tukey window. For alpha of 0.75, the trace (a) is the time
domain weighted coefficients, whereas the trace (b) shows the main frequency lobe with the first
side lobe at − 19 dB and a − 18 dB per octave side lobe fall-off rate

coefficients for alpha values of 1, 2, 3, and 4 are shown in Figs. 16.11(a), 16.12(a), 16.13(a),
and 16.14(a), respectively. The resulting Log-magnitude of the Fourier transform for
alpha values of 1, 2, 3, and 4 are shown in Figs. 16.11(b), 16.12(b), 16.13(b), and 16.14(b),
respectively. Notice in the figures what effect increasing the value of alpha from 1 to 4
has on the first side-lobe value and the side-lobe roll-off rate. Note that for each increase
of alpha by 1 the absolute magnitude of the side-lobe roll-off rate increases by 6 dB.

1.2

-20

1.0

-40

0 T -n bins 0 n bins

(a) (b)

FIGURE 16.11 : Hanning window with alpha of 1. The trace (a) is the time domain weighted
coefficients, whereas the trace (b) shows the main frequency lobe with the first side-lobe at − 23 dB

and a − 12 dB per octave side-lobe fall-off rate