the measure of "wavyness" of signals

Hello guys,

Happy thanksgiving!

I recently collected many signals.  I need to sort them based on their
"wavyness".  In other words, some signals are just like a straightline
-- almost with ZERO degree of "wavyness"; some signals are like sine
waves -- has A LOT OF "wavyness".

I would need a SINGLE value to measure "wavyness" for each signal.  As
a causal user of signal processing, I am not sure how to proceed.

Thank you very much for your input.

Alex

<lluum@yahoo.com> wrote in message 
news:1164339204.152703.252480@h54g2000cwb.googlegroups.com...
> Hello guys,
>
> Happy thanksgiving!
>
> I recently collected many signals.  I need to sort them based on their
> "wavyness".  In other words, some signals are just like a straightline
> -- almost with ZERO degree of "wavyness"; some signals are like sine
> waves -- has A LOT OF "wavyness".
>
> I would need a SINGLE value to measure "wavyness" for each signal.  As
> a causal user of signal processing, I am not sure how to proceed.
>
> Thank you very much for your input.
>
> Alex
>

Compute the variance.

First, compute the mean or average.
Subtract it out.
Square the result.
Integrate the result (add all the values together).
The higher the result, the higher the variability / variance / waviness....

Fred 

Fred Marshall wrote:
> <lluum@yahoo.com> wrote in message
> news:1164339204.152703.252480@h54g2000cwb.googlegroups.com...
> > Hello guys,
> >
> > Happy thanksgiving!
> >
> > I recently collected many signals.  I need to sort them based on their
> > "wavyness".  In other words, some signals are just like a straightline
> > -- almost with ZERO degree of "wavyness"; some signals are like sine
> > waves -- has A LOT OF "wavyness".
> >
> > I would need a SINGLE value to measure "wavyness" for each signal.  As
> > a causal user of signal processing, I am not sure how to proceed.
> >
> > Thank you very much for your input.
> >
> > Alex
> >
>
> Compute the variance.
>
> First, compute the mean or average.
> Subtract it out.
> Square the result.
> Integrate the result (add all the values together).
> The higher the result, the higher the variability / variance / waviness....
>
> Fred

I agree with Fred that variance is a measure of the "waviness", but
here's how I would calculate it for your signal, y:
var(y)

lluum@yahoo.com wrote:
> Hello guys,
> 
> Happy thanksgiving!
> 
> I recently collected many signals.  I need to sort them based on their
> "wavyness".  In other words, some signals are just like a straightline
> -- almost with ZERO degree of "wavyness"; some signals are like sine
> waves -- has A LOT OF "wavyness".
> 
> I would need a SINGLE value to measure "wavyness" for each signal.  As
> a causal user of signal processing, I am not sure how to proceed.

That's common with causal users.  Non-causal users on the other hand 
have the answer even before they are given the problem. :=)

Regards,
John
> 
> Thank you very much for your input.
> 
> Alex
> 

John Monro wrote:
> lluum@yahoo.com wrote:
> > Hello guys,
> >
> > Happy thanksgiving!
> >
> > I recently collected many signals.  I need to sort them based on their
> > "wavyness".  In other words, some signals are just like a straightline
> > -- almost with ZERO degree of "wavyness"; some signals are like sine
> > waves -- has A LOT OF "wavyness".
> >
> > I would need a SINGLE value to measure "wavyness" for each signal.  As
> > a causal user of signal processing, I am not sure how to proceed.
>
> That's common with causal users.  Non-causal users on the other hand
> have the answer even before they are given the problem. :=)
>
> Regards,
> John
> >
> > Thank you very much for your input.
> >
> > Alex
> >


He's not real clear about "waviness."  I could see that being
frequency, amplitude, or variance, or some combination.

> He's not real clear about "waviness."  I could see that being
> frequency, amplitude, or variance, or some combination.

Considering he's a causal person, I thought his "waviness" description
matched variance quite well, though it could be standard deviation that
he's looking for.

lluum@yahoo.com skrev:
> Hello guys,
>
> Happy thanksgiving!
>
> I recently collected many signals.  I need to sort them based on their
> "wavyness".  In other words, some signals are just like a straightline
> -- almost with ZERO degree of "wavyness"; some signals are like sine
> waves -- has A LOT OF "wavyness".
>
> I would need a SINGLE value to measure "wavyness" for each signal.

It seems to me that you are looking for a measure for a drifting mean.
If so, this is what I would try:

- Decide on a frame length N
- Split your signal in a number of frames of length N
- Compute the means of each frame
- Compute the variance for eacf frame
- Compare the ratio of maximum and minimum mean
  with the mean variance of the frames.

Rune

In article <1164339204.152703.252480@h54g2000cwb.googlegroups.com>, 
lluum@yahoo.com says...
> Hello guys,
> 
> Happy thanksgiving!
> 
> I recently collected many signals.  I need to sort them based on their
> "wavyness".  In other words, some signals are just like a straightline
> -- almost with ZERO degree of "wavyness"; some signals are like sine
> waves -- has A LOT OF "wavyness".
> 
> I would need a SINGLE value to measure "wavyness" for each signal.  As
> a causal user of signal processing, I am not sure how to proceed.
> 
> Thank you very much for your input.
> 
> Alex
> 
> 

you could look at the fft and see if there are significant contributions 
at frequencies not considered low.

-- 
Loren
http://blogs.mathworks.com/loren/

thank you soooooo much for your kind input.

I am sorry that I miss-spell the "casual" to "causal".  That lead to
some confusion.

Alex

Thank you very much for your kind input.

However, after looking into this, variance would NOT work.

Reason:
variance squares the departure from mean, a straigth line may well has
the same variance as that of a sine wave.

Maybe I will try kurtosis, or simply integrate the departure from mean
Sum(Ni - Avg)

Any other input will be highly appreciated
................................