Table D 1: Filter Characteristics Of Fft Windowing Functions - Tektronix 11801C User Manual

11801C User Manual
H Blackman
R * 1
0.42 * 0.5 cos(2p n
x(n) +
|
n + 0
H Blackman Harris
R * 1
0.35875 * 0.48829 cos(2p n
x(n) +
n + 0
where
R = the record length, which must be a power of 2.
The filter characteristics of the FFT windowing functions are summarized in
Table D 1. These numbers are taken from a table in Frederic J. Harris:
Handbook of Digital Signal Processing, edited by Douglas F. Elliot, Academ
ic Press, San Diego, 1987, pp. 254-255.

Table D 1: Filter Characteristics of FFT Windowing Functions

Windowing Highest
Function Side
lobe
(dB)
Rectangular -13
Triangular -27
Hanning -32
Hamming -43
Blackman -58
Blackman -92
Harris
where
bins refers to the frequency bins, the even numbered points in the FFT
waveform.
highest sidelobe is the minimum stopband attenuation, which indicates
how well leakage is blocked.
noise bandwidth is the equivalent noise bandwidth (the width of an
equivalent rectangular spectral response that would pass the same
noise power as this windowing function.
3 dB bandwidth and 6 dB bandwidth apply to the major lobe. These
affect the frequency resolution.
) ) 0.08 cos(4p n
R
)
R
) 0.14128 cos(4p n ) * 0.01168 cos(6p n
R
Noise 3 dB 6 dB
Band Band Band
width width width
(bins) (bins) (bins)
1.00 0.89 1.21
1.33 1.28 1.78
1.50 1.44 2.00
1.36 1.30 1.81
1.73 1.68 2.35
2.00 1.90 2.72
Appendix D: Algorithms
)
R
)
R
Scallop Worst
Loss Loss
(dB) (dB)
3.92 3.92
1.82 3.07
1.42 3.18
1.78 3.30
1.10 3.47
0.83 3.85
D 5