C***C
发帖数: 216 | 1 White noise is not always Gaussian and Gaussian noise is not always white. The
term white refers to the autocorrelation characteristics of the noise. That
is,
the autocorrelation of white noise is an impulse which implies it has an
infinite flat spectrum. Gaussian refers to the characteristics of the noise a
particular time across an ensemble of realization. That is, if we observe
several realization of the noise at particular time and collect the values,
then values will have Gaussian distribu | C***C
发帖数: 216 | 2 White noise is not necessarily Gaussian and Gaussian noise may not be white.
Whiteness means that the noise samples are not correlated and Gaussian means
that the amplitudes of the samples are Gaussian distributed. Hence, you can
have white non-Gaussian noise and non-white Gaussian noise.
Filtering a noise sequence introduce correlation into the sequence, but will
not alter the distribution of the amplitude due to the property of "linear
sum of Gaussians will be Gaussian". So, essentially, if yo
【在 C***C 的大作中提到】
: White noise is not always Gaussian and Gaussian noise is not always white. The
: term white refers to the autocorrelation characteristics of the noise. That
: is,
: the autocorrelation of white noise is an impulse which implies it has an
: infinite flat spectrum. Gaussian refers to the characteristics of the noise a
: particular time across an ensemble of realization. That is, if we observe
: several realization of the noise at particular time and collect the values,
: then values will have Gaussian distribu
|
|
