Re: Noise Source with Limiter ?

From: Nicholas O. Lindan (see_at_sig.com)
Date: 10/17/04 

Date: Sun, 17 Oct 2004 00:53:18 GMT

"Tim Wescott" <tim@wescottnospamdesign.com> wrote

> He was talking about a histogram on the output of a limiter, which
> _does_ have peaks -- exactly two, centered on the upper and lower
> limits. And that ain't Gaussian.

Nobody is disagreeing on this point.

> > In any case, the output switches between ECL one and zero. You could look at
> > the output at fixed intervals and find it has a binary distribution, or
> > measure the duty cycle which would have a histogram similar to the amplitude
> > distribution of Gaussian noise.
> >
> > [...]
> >
> Neither of which is a Gaussian distribution on amplitude.

Nobody said it was.

I say 'who cares about amplitude distribution', all I want
is ones and zeros; and ones and zeros don't have an amplitude
distribution that is of any interest, as you have noted.

OP (or someone like him) says the _pulse widths_ of discriminated
analog noise follow a Gaussian distribution, _not the amplitudes.

A noise source does not have to have any analog component. It's just
unpredictable data and has (or should have) _no_ relation to any
physical process - viz. irrational numbers.

Most of us, being cheap, use a zener diode/radio noise/chattering circuit
as the source of choice and so there is an analog voltage/current
involved. It could just as easily be a flipped coin and have no
amplitude component anywhere.

I think there are two conversations going on here: one on the properties of
electronic noise and the other on random numbers. They have little to
do with each other except for both being ideally unpredictable. 

--
Nicholas O. Lindan, Cleveland, Ohio
Consulting Engineer:  Electronics; Informatics; Photonics.
Remove spaces etc. to reply: n o lindan at net com dot com
psst.. want to buy an f-stop timer? nolindan.com/da/fstop/

Relevant Pages

  • Re: Creating noise in the frequency domain?
    ... Gaussian white noise poorly models real image noise. ... What is the frequency-domain-equivalent distribution of complex ... distributed noise, the spatial domain noise will be white, almost ...
    (sci.image.processing)
  • Re: Creating noise in the frequency domain?
    ... Gaussian white noise poorly models real image noise. ... What is the frequency-domain-equivalent distribution of complex ... distributed noise, the spatial domain noise will be white, almost ...
    (sci.image.processing)
  • Re: Creating noise in the frequency domain?
    ... Not image noise. ... The Gaussian assumption isn't as bad because ... "It is just one way of modeling white noise with a Gaussian amplitude ... there is no reason to model with another distribution because the ...
    (sci.image.processing)
  • Re: Does this have a flaw in de-biasing an entropy stream?
    ... "flat" distribution most users want. ... That depends on the noise source. ... to Poisson; some are Gaussian; some are binary ... If we define "noise" to mean a physical randomness source that has ...
    (sci.crypt)
  • Re: Frequency retrieval from noise
    ... distribution, one performs fourier transform to identify the ... noise, which means that the the spectrum of the noise is flat. ... I'm asking what are the peaks in the power spectrum of pure ... AWGN is gaussian in amplitude distribution, ...
    (sci.physics)