Randy Yates wrote:
Don Pearce writes:
Jorden Verwer wrote:
Don Pearce wrote:
Jorden Verwer wrote:
JosephKK wrote:
Noise-like phenomena, in increasing order of bandwith:
Offset == 1/f noise == white noise
OK alligator, where does shot noise fall in the spectrum?
I never claimed that the list was exhausistive, but to answer your
question:
I would put it between 1/f noise and white noise.
The spectrum of shot noise is white - why would it be otherwise?
Shot noise will always be band limited because electrons have a nonzero
transit time. Its bandwith is very high, but not infinite.
Not helpful - in audio terms and a long, long way beyond, it is white.
Don,
Perhaps your definition of white is different than mine. There are two
I've seen: one says pretty much directly that it is noise with infinite
bandwidth, and the other implies it through the relationship of a Dirac
delta function (the autocorrelation function of an uncorrelated white
noise process) and the Fourier Transform. You can find those definitions
in this short paper on the unrelated topic of the DC value of white
noise:
http://www.digitalsignallabs.com/white.pdf
--Randy
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.
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). I think the pragmatic definition is
probably the more useful and easily understood.
d