Geoff,

I am definitely not the one to be giving the real answers to this.  I'm 
just (re)learning this stuff, myself.  It is my impression that the FFT, 
properly calibrated, is itself a spectral density curve, in my case based 
on ground velocity.  Much more commonly seen is a plot of power density, 
indicating how much noise power is at each frequency (in each frequency 
band).  Such plots are based on the squares of the velocity numbers.

Randall 
Peters  suggests  http://physics.mercer.edu/hpage/CSP/cumulative.html 
that the Cumulative Spectral Power curve, which is pretty much like the 
integral you mention, has some advantages over the un-integrated curve, 
though the latter is what is now in general use for defining instrument and 
site noise.  You can find the power spectral density curves for most 
instruments/sites in the various research networks.

The above is about 20% more than I know about the subject, which covers a 
lot of math--mostly probablility theory and lots of calculus.  I suspect 
not for the faint of heart.

Regards,
Brett

At 09:20 AM 3/28/2009 -0700, you wrote:
>Hello You Math experts;
>
>By Spectral Density are you by chance
>figuring the area under the curve ?
>Which I believe to be a calculus thingy ?
>
>Regards;
>geoff


__________________________________________________________

Public Seismic Network Mailing List (PSN-L)