How do you calculate the noise equivalent bandwidth for an IIR or FIR filter?
$\begingroup$
$\endgroup$
$\begingroup$
$\endgroup$
Hm, isn't it the same as for any filter?
$$ B_\text{eq} = \frac{1}{|H(e^{j\omega_0})|^2} \int_0^\infty |H(e^{j\omega})|^2 d\omega $$
where $H(e^{j\omega})$ is the frequency response, and $\omega_0$ is the max-abs frequency. For FIR filters, you can just use e.g. freqz in MATLAB/Octave. For IIR filters, you'd have to do analysis or take measurements (or see if you can get away with truncating)
-
$\begingroup$ Shouldn't the denominator term be $|H(e^{j\omega_0})|^2$ ? onmyphd.com/?p=enbw.equivalent.noise.bandwidth $\endgroup$ – Peter K.♦ Nov 11 '13 at 22:15
-
-
$\begingroup$ Shouldn't it be in discrete frequency, not continuous? $\endgroup$ – Seth Nov 12 '13 at 0:04
-
$\begingroup$ I do not think so. The DTFT is continuous (actually a complex Fourier series). $\endgroup$ – toast Nov 12 '13 at 0:16
