ENBW documentation is ambiguous/incorrect.

In the ENBW documentation page http://uk.mathworks.com/help/signal/ref/enbw.html#btrihsq-1, the enbw(window(N),Fs) function is described as returning the two-sided bandwidth (in Hz) of a window. As far as I can see, this is a correct statement and evaluates (for all w(i) in N samples)

ENBW * (Fs/N) = (N * sum(w(i)^2)/sum(w(i))^2) * (Fs/N)

But, the example code given in EquivalentRectangularNoiseBandwidthExample.m clearly contains the lines:

bw = enbw(hann(1000),Fs)
.
.
.
plots:plot(bw*[-1 -1 1 1],[-40 maxgain maxgain -40],'--')

Surely this is incorrect and the example should plot:

plots:plot(bw/2*[-1 -1 1 1],[-40 maxgain maxgain -40],'--')

Andrew

Answer 1

[Greg Dionne]

Looks like this is fixed now.