Windowing
Windowing is a technique used in Discrete Fourier Transforms
to reduce numerical computation errors resulting from the discrete nature of the
transform process. Windowing is most effective when the time domain
waveform is many periods long. Windowing may introduce amplitude
inaccuracies in the frequency domain spectrum if too few periods are used.
Piece-wise Linear waveforms imported from file might not
contain an integer number of periods and might not be comprised of 2N
points, two criteria that introduce numerical computation errors in the Discrete
Fourier Transform process. Windowing minimizes these computation
errors. In EMI Analyst, windowing is only beneficial for imported
Piece-wise Linear waveforms. All other waveforms in EMI Analyst
have exactly one period and exactly 2N points.
EMI Analyst applies a Hann Window function to the
time domain waveform when the Apply Windowing check box is
selected. The Window function is only available for the Piece-wise Linear
waveform and is only necessary for imported waveforms.
When windowing is applied, the time domain waveform will
appear tapered near the beginning and end of the waveform. This tapering
reduces "leakage" cause by the Discrete Fourier Transform process. For
more information on windowing the user is encouraged to reference any of the
numerous texts on Discrete Fourier Transforms.
More:
 Filters
|
