Question about comparing PSD's

Hi,
 How can we compare the Power Spectral Densities of two signals ? I do
not want a visual technique. One method would be to compute the CDF's
of the PSD's and compare those using a confidence test.Is there any
other technique.
Thanks,
Sarma

sarma.vangala@gmail.com (Amras) writes:

> Hi,
>  How can we compare the Power Spectral Densities of two signals ? I do
> not want a visual technique. One method would be to compute the CDF's
> of the PSD's and compare those using a confidence test.Is there any
> other technique.
> Thanks,
> Sarma

What are you trying to accomplish? Are you trying to find out if two signals
are correlated? Why not use a cross-correlation function?
-- 
%  Randy Yates                  % "Maybe one day I'll feel her cold embrace,
%% Fuquay-Varina, NC            %                    and kiss her interface, 
%%% 919-577-9882                %            til then, I'll leave her alone."
%%%% <yates@ieee.org>           %        'Yours Truly, 2095', *Time*, ELO   
http://home.earthlink.net/~yatescr

Amras wrote:
> Hi,
>  How can we compare the Power Spectral Densities of two signals ? I do
> not want a visual technique. One method would be to compute the CDF's
> of the PSD's and compare those using a confidence test.Is there any
> other technique.
> Thanks,
> Sarma

Kulback divergence.

Hi,
 Thanks for the replies. I am working on a signal processing technique
where a machine has to take a decision based on the PSD's of two
incoming signals. So I was wondering if there is a mechanism to
convert the entire PSD plot to a simple (if possible) single value.
Thanks,
Sarma




Stan Pawlukiewicz <spam@spam.mitre.org> wrote in message news:<cnd5g0$gg$1@newslocal.mitre.org>...
> Amras wrote:
> > Hi,
> >  How can we compare the Power Spectral Densities of two signals ? I do
> > not want a visual technique. One method would be to compute the CDF's
> > of the PSD's and compare those using a confidence test.Is there any
> > other technique.
> > Thanks,
> > Sarma
> 
> Kulback divergence.

On 2004-11-17 05:09:03 +0100, sarma.vangala@gmail.com (Amras) said:

> Hi,
>  Thanks for the replies. I am working on a signal processing technique
> where a machine has to take a decision based on the PSD's of two
> incoming signals. So I was wondering if there is a mechanism to
> convert the entire PSD plot to a simple (if possible) single value.
> Thanks,
> Sarma

You could probably use the higher moments of the distribution such as 
skewness and kurtosis to characterize and compare the two PSDs...
-- 
Stephan M. Bernsee
http://www.dspdimension.com

"Amras" <sarma.vangala@gmail.com> wrote in message
news:ae49a4ed.0411162009.43bdeb9b@posting.google.com...
> Hi,
>  Thanks for the replies. I am working on a signal processing technique
> where a machine has to take a decision based on the PSD's of two
> incoming signals. So I was wondering if there is a mechanism to
> convert the entire PSD plot to a simple (if possible) single value.
> Thanks,
> Sarma
>

Hello Sarma,

If the PSDs are single parameter PSDs, then the answer is a simple yes. If
the PSDs have two or more parameters, which ones do you wish to discard?

Clay

Stan Pawlukiewicz <spam@spam.mitre.org> wrote 

> 
> Kulback divergence.
>

I'll see your Kullback and raise you a Leibler.

Peter Kootsookos wrote:
> Stan Pawlukiewicz <spam@spam.mitre.org> wrote 
> 
> 
>>Kulback divergence.
>>
> 
> 
> I'll see your Kullback and raise you a Leibler.

So called.

On 18 Nov 2004 17:33:38 -0800, p.kootsookos@iolfree.ie (Peter
Kootsookos) wrote:

>Stan Pawlukiewicz <spam@spam.mitre.org> wrote 
>
>> 
>> Kulback divergence.
>>
>
>I'll see your Kullback and raise you a Leibler.

Hi Dr. K,

  hope ya' know I have the most *profound* respect 
for you, ... but I must ask: "Did you forget to take 
your medication today?   :-)

[-Rick-]

r.lyons@_BOGUS_ieee.org (Rick Lyons) wrote 

>   hope ya' know I have the most *profound* respect 
> for you, ... but I must ask: "Did you forget to take 
> your medication today?   :-)

;-)

I've just usually heard what StanP was referring to as the
"Kullback-Leibler" divergence.

See, e.g.

http://www.cis.hut.fi/aapo/papers/NCS99web/node26.html

Ciao,

Peter K.