[Don Pearce]

Randy Yates wrote:
Don Pearce writes:
[...]
My definition of white noise is that equal bandwidths contain equal
energy, whatever their centre frequency. Certainly over the full
extent of the audio band, and as I say a very long way beyond, that is
true for shot noise.

I agree that shot noise is constant in some bandwidth around a centre
frequency that extends "a very long way beyond" the audio band. What I
am saying is that, even though the term "white noise" is used
colloquially to mean "constant power over some large bandwidth," that
usage is wrong according to the definition.

That wouldn't be my usage - mine is that it is constant within any
selected small bandwidths, as long as they are the same. If you measure
the noise power over, say half the audio band in one hit, you have no
way of judging whether or not it is white. Far better would be to use,
say, 10Hz bandwidth measured at manyb centre frequencies from 20Hz to 20kHz.

It would be useful to coin a new phrase for such bandlimited noise;
"bandlimited white noise"?


I'm happy with banlimited white noise, provided that the limited band in
question contains everything you are trying to measure. If it doesn't
then you can't really use white to describe it.

There is a sentence in that paper that shows the limitation of
Papoulis's theory in that it cannot of itself show that white noise
has a zero mean (no DC component).

That was the point of both sections: neither Papoulis' nor Brown's
definitions imply zero mean.

I think the pragmatic definition is probably the more useful and
easily understood.

I would agree that myself and pretty much everyone else here understands
what is meant by "white" in this context.


I thing white in an audio context is easy enough to understand - my
first paragraph in my reply covers it.

What I do not agree with is when someone such as Jordan is careful with
the terminology and meets with resistance. Jordan is not wrong for using
the term "white noise" in its proper form!

Agreed

d