[SciPy-User] Find the RMS from a PSD

[Charles R Harris]

On Mon, Jan 24, 2011 at 8:01 AM, E <pub at jesuislibre.net> wrote:

>  Hello scipy users.
>
> I'm new to signal processing and I've read that RMS could be found from a
> PSD. I'm interested in as I would further like to know energy in a signal
> through it's frequencies.
> My problem is I don't find how to calculate the RMS from the PSD output. It
> seems it's a matter of scale (frequencies bandwith is taken in account
> already).
>
>
The noise variance is the integral of the PSD, so take the square root. The
tricky part is knowing if you have a single sided or double sided PSD and
making sure you have the right units everywhere. The units of the PSD should
be something like Volts**2/Hz and you need to integrate over Hz. I like to
use the FFT of the autocorrelation, windowed to reduce the noise, multiplied
by the sample interval, to get the two sided PSD.

<snip>

Chuck
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20110124/fe6c33aa/attachment.html>