w = gausswin(N)
w = gausswin(N,Alpha)
w = gausswin(N) returns an
Gaussian window in a column vector,
a positive integer.
w = gausswin(N,Alpha) returns
N-point Gaussian window with
to the reciprocal of the standard deviation. The width of the window
is inversely related to the value of α. A
larger value of α produces a narrower window.
The value of α defaults to 2.5.
If the window appears to be clipped, increase
the number of points.
Create a 64-point Gaussian window. Display the result in
L = 64; wvtool(gausswin(L))
This example shows that the Fourier transform of the Gaussian window is also Gaussian with a reciprocal standard deviation. This is an illustration of the time-frequency uncertainty principle.
Create a Gaussian window of length 64 by using
gausswin and the defining equation. Set , which results in a standard deviation of 64/16 = 4. Accordingly, you expect that the Gaussian is essentially limited to the mean plus or minus 3 standard deviations, or an approximate support of [-12, 12].
N = 64; n = -(N-1)/2:(N-1)/2; alpha = 8; w = gausswin(N,alpha); stdev = (N-1)/(2*alpha); y = exp(-1/2*(n/stdev).^2); plot(n,w) hold on plot(n,y,'.') hold off xlabel('Samples') title('Gaussian Window, N = 64')
Obtain the Fourier transform of the Gaussian window at 256 points. Use
fftshift to center the Fourier transform at zero frequency (DC).
nfft = 4*N; freq = -pi:2*pi/nfft:pi-pi/nfft; wdft = fftshift(fft(w,nfft));
The Fourier transform of the Gaussian window is also Gaussian with a standard deviation that is the reciprocal of the time-domain standard deviation. Include the Gaussian normalization factor in your computation.
ydft = exp(-1/2*(freq/(1/stdev)).^2)*(stdev*sqrt(2*pi)); plot(freq/pi,abs(wdft)) hold on plot(freq/pi,abs(ydft),'.') hold off xlabel('Normalized frequency (\times\pi rad/sample)') title('Fourier Transform of Gaussian Window')
The coefficients of a Gaussian window are computed from the following equation:
where –(N – 1)/2 ≤ n ≤ (N – 1)/2 and α is inversely proportional to the standard deviation, σ, of a Gaussian random variable. The exact correspondence with the standard deviation of a Gaussian probability density function is σ = (N – 1)/(2α).
 Harris, Fredric J. “On the Use of Windows for Harmonic Analysis with the Discrete Fourier Transform.” Proceedings of the IEEE®. Vol. 66, January 1978, pp. 51–83.
 Roberts, Richard A., and C. T. Mullis. Digital Signal Processing. Reading, MA: Addison-Wesley, 1987, pp. 135–136.
Usage notes and limitations:
inputs must be constant. Expressions or variables are allowed if their
values do not change.