[sdiy] gaussian noise related to white noise

[Ethan Duni]

No no no...

Noise can be Gaussian and not white, or white but not Gaussian. Saying noise
is white is a statement about the dependence of the value of the noise at
some time on the value of the noise at another time. Noise is called "white"
if there is no correlation between the values of the noise process at
different times (that is, the noise is a sequence of independent, or at
least uncorrelated, random variables). As someone else mentioned, this is
equivalent to the power spectral density of the noise process being a
constant (which means double the power for each octave). This says almost
nothing about the actual distribution of the noise at any given time.

For noise to be Gaussian, it means that the value of the noise at any given
time has a Gaussian distribution (which may or may not depend on the value
of the noise at other times). White, Gaussian noise is probably the most
common noise model, hence the confusion.

Ethan