`Autocorrelations.acf`

— Function`acf(x [, lags]; demean=false, normalize=false)`

Evaluate the autocorrelation function of signal `x`

.

By default, the acf is evaluated at all available lags `0:size(x,1)-1`

, but arbitrary `lags`

can be optionally specified.

**Keywords**

`demean`

: whether to subtract the mean of`x`

before evaluating the acf`normalize`

: whether to normalize the acf to its lag-0 value

When `length(x) < 1024`

, a raw computation (`dotacf`

) is used, whereas a FFT-based algorithm is used for larger arrays (`fftacf`

).

`Autocorrelations.acf!`

— Method`acf!(r, x, lags; demean=false, normalize=false)`

Evaluate the autocorrelation of `x`

at time `lags`

in-place and store output in `r`

. See `acf`

. For this in-place version, `lags`

*must* be specified, and `length(r)==length(lags)`

.

`Autocorrelations.dotacf`

— Function`dotacf(x [, lags]; demean=false, normalize=false)`

Evaluate the autocorrelation of signal `x`

with a dot-product-based algorithm (see `acf`

).

`Autocorrelations.fftacf`

— Function`fftacf(x [, lags]; demean=false, normalize=false)`

Evaluate the autocorrelation of signal `x`

with a FFT-based algorithm (see `acf`

).