"Carlos Moreno" <moreno_at_mochima_dot_com@mailinator.com> wrote in message
news:gJAph.12310$U04.181746@weber.videotron.net...
<snipped without loss of context I hope?>

> Maybe you could give it a second try at making me understand
> what I perceive as a contradiction --- I mean, my point is that
> white noise *does not* have infinite power (however, this is
> directly contradicting by the fact that PSD is really that:
> density *of power*, which would directly imply that the power
> has to be infinite for white noise) --- more specifically, the
> reason why white noise can not exist in practice *is not* that
> it has infinite power; the reason is so much simpler than that!
> W.N. can not exist because it simply can't --- no signal *in our
> reality* can vary with *truly infinite* rate of change; W.N. is
> *necessarily* a theoretical construct --- albeit one very useful,
> like Dirac's delta, or complex numbers.
>

Hi Carlos - If it doesn't matter how close together in time your sample
points are then they aproach 0 seconds separation and your power , even with
limited amplitude change _is_ infinite isn't it? I can't see any
discrepancy other than that between what you can think of and what can
actually be observed.
Best of Luck - Mike

Reply by Jerry Avins●January 12, 20072007-01-12

doggie wrote:

>>> "Gaussian" doesn't describe the (frequency) spectrum. It describes the
>>> PDF (probability density function) of the individual sample values.
>> To put it into the terms in which the question was asked, The
>> distribution of values -- the shape of a bar graph that bins the values
>> -- is bell shaped.
>>
>> The spectrum of sampled noise depends on how each sample relates to
>> those around it. The PDF -- probability density function -- depends on
>> the shape of that bar graph I mentioned above.
>>
>> Jerry
>> _____________________________________________
>> * The more in the sum, the closer to Gaussian the result becomes.
>> --
>> Engineering is the art of making what you want from things you can get.
>> �����������������������������������������������������������������������
>>
>
> So you mean that the values of the random numbers tend to be closer to the
> value of the mean as can be seen from the bell curve? If so, what is the
> mean in this case - the mean of all the random numbers generated?

The mean depends on the data. The expected mean -- not necessarily even
close to the measured mean for small sets -- depends on the way the data
are generated. Many processes are "zero mean"; some of those that aren't
are transformed to that by subtracting the mean from every instance.
A playing die, with its possibilities being the integers from 1 to 6 has
an expected mean of 3.5. Therefore, the sum of two dice has an expected
mean of 7. But the lowest and highest possible sums are 2 and 12, with
expectations of 1/36 each, while the expectation for 7 is 1/6.
You can make a pretty good approximation of a zero-mean Gaussian random
sequence by subtracting 35 from the sum of 10 dice and scaling the
result to suit the problem.
Jerry
--
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

Reply by doggie●January 12, 20072007-01-12

>> "Gaussian" doesn't describe the (frequency) spectrum. It describes the
>> PDF (probability density function) of the individual sample values.
>
>To put it into the terms in which the question was asked, The
>distribution of values -- the shape of a bar graph that bins the values
>-- is bell shaped.
>
>The spectrum of sampled noise depends on how each sample relates to
>those around it. The PDF -- probability density function -- depends on
>the shape of that bar graph I mentioned above.
>
>Jerry
>_____________________________________________
>* The more in the sum, the closer to Gaussian the result becomes.
>--
>Engineering is the art of making what you want from things you can get.
>�����������������������������������������������������������������������
>

So you mean that the values of the random numbers tend to be closer to the
value of the mean as can be seen from the bell curve? If so, what is the
mean in this case - the mean of all the random numbers generated?

Reply by Jerry Avins●January 12, 20072007-01-12

Ikaro wrote:

> I don't think this will work. Like other have pointed out, flat
> spectrum is a necessary, but not a sufficient condition for a white
> noise process.
>
> I'd imagine you will also need to do some tests on the phase spectrum
> to further determine if the process is noise or a determistic signal
> (but you don't have phase information on the PSD).
>
> Also, what I previously meant was a statistical test for white noise
> (some test done on a time series that will yield an alpha value
> indicating how confident you are that the data is white noise).
> Like a mentioned before, I don't think that t-test on the amplitude of
> the bins in the magnitude spectrum will work.
>
> -Ikaro
>
>
> Jerry Avins wrote:
>> Ikaro wrote:
>>> Hey,
>>>
>>> Like others pointed out, you can't determine if it is white simply by
>>> looking at the amplitude distribution.
>>>
>>> There are statistical tests to determine if a signal is random (like
>>> Runs Test and the Turning Points Methods).
>>>
>>> However I am not aware of any statistical test to determine if a
>>> sequence is white or not (Anyone here??)...

>>

>> Test for a flat spectrum.

I thought we were looking to determine if a noise signal is white, not
to determine if an arbitrary signal is white noise.
Jerry
--
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

Reply by Ikaro●January 12, 20072007-01-12

I don't think this will work. Like other have pointed out, flat
spectrum is a necessary, but not a sufficient condition for a white
noise process.
I'd imagine you will also need to do some tests on the phase spectrum
to further determine if the process is noise or a determistic signal
(but you don't have phase information on the PSD).
Also, what I previously meant was a statistical test for white noise
(some test done on a time series that will yield an alpha value
indicating how confident you are that the data is white noise).
Like a mentioned before, I don't think that t-test on the amplitude of
the bins in the magnitude spectrum will work.
-Ikaro
Jerry Avins wrote:

> Ikaro wrote:
> > Hey,
> >
> > Like others pointed out, you can't determine if it is white simply by
> > looking at the amplitude distribution.
> >
> > There are statistical tests to determine if a signal is random (like
> > Runs Test and the Turning Points Methods).
> >
> > However I am not aware of any statistical test to determine if a
> > sequence is white or not (Anyone here??)...
>
> Test for a flat spectrum.
>
> ...
>
> Jerry
> --
> Engineering is the art of making what you want from things you can get.
> =AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=

>
>
> Autocorrelation function and PSD are a Fourier transform pair.
>

Yes this was proven and published by Einstein back in 1914 - long
before Weiner or Khinchin.

Reply by Robert Scott●January 12, 20072007-01-12

On 12 Jan 2007 07:41:55 -0800, "Ikaro" <ikarosilva@hotmail.com> wrote:

>I don't think this will work. Like other have pointed out, flat
>spectrum is a necessary, but not a sufficient condition for a white
>noise process.
>
>I'd imagine you will also need to do some tests on the phase spectrum
>to further determine if the process is noise or a determistic signal
>(but you don't have phase information on the PSD).
>
>Also, what I previously meant was a statistical test for white noise
>(some test done on a time series that will yield an alpha value
>indicating how confident you are that the data is white noise).
>..

It is impossible to check absolutely if the signal is random or deterministic.
A good cryptographic pseudo random noise generator is completely deterministic.
But unless you know the key, it is indistinguishable from true randomness. Some
simple pseudo random number generators may fail some simple statistical tests,
but mere failure does not guarantee that the source is random.
Robert Scott
Ypsilanti, Michigan

Reply by Jerry Avins●January 12, 20072007-01-12

Oli Charlesworth wrote:
...

>> What is the basic (or not so basic) definition of gausian noise is? Is
>> it noise that has a bell shaped spectrum?
>
> "Gaussian" doesn't describe the (frequency) spectrum. It describes the
> PDF (probability density function) of the individual sample values.

To put it into the terms in which the question was asked, The
distribution of values -- the shape of a bar graph that bins the values
-- is bell shaped.
The spectrum of sampled noise depends on how each sample relates to
those around it. The PDF -- probability density function -- depends on
the shape of that bar graph I mentioned above.
Most random number generators provide a uniform distribution: any number
is as likely as any other. The sum of two uniformly distributed random
numbers is not uniformly distributed, but triangular*. Honest playing
dice generate a uniform distribution in the range 1 to 6. In the sum of
two dice, middle-size numbers are more common than high or low ones.
Jerry
_____________________________________________
* The more in the sum, the closer to Gaussian the result becomes.
--
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

Reply by Oli Charlesworth●January 12, 20072007-01-12

On Jan 12, 8:45 am, Chris Barrett
<chrisbar...@0123456789abcdefghijk113322.none> wrote:

> Chris Barrett wrote:
> > Let's say my audio is represented by a series of numbers and each number
> > has a random value. Is my audio white noise? I think it is, but I'm
> > having trouble proving it to my self.To go slightly off-topic:
>
> What is the basic (or not so basic) definition of gausian noise is? Is
> it noise that has a bell shaped spectrum?

"Gaussian" doesn't describe the (frequency) spectrum. It describes the
PDF (probability density function) of the individual sample values.
--
Oli

Reply by Michel Rouzic●January 12, 20072007-01-12

Fitlike Min wrote:

> "Chris Barrett" <chrisbarret@0123456789abcdefghijk113322.none> wrote in
> message news:Nwxph.1353$qA7.1219@newsfe15.lga...
> > Let's say my audio is represented by a series of numbers and each number
> > has a random value. Is my audio white noise? I think it is, but I'm
> > having trouble proving it to my self.
>
> Take the autocorrelation. If it's an impulse then it's white - though it may
> not necessarily be Guassian white etc (this has been explained).

I think it will only tell you whether or not the signal has a flat
spectrum, so while it may be a white noise it may be anything else that
has a flat spectrum.