[CST-2] Autocorrelation

[Tom Puverle]

Alvin,

all the answers to the past questions are in the worksheet on his website.
As for the autocorrelation, the way it works it assumes that the noise is
random and ADDITIVE. It you correlate the function with itself (do all the
shifts from -inf to inf) the noise (since it's random) will cancel each
other out. However, since the pattern is not random, it will be reinforced.
The best it works (for those reasons) is with random gaussian noise and some
form of sin wave. (since it's perfectly periodic)
HTH,
Tom

> -----Original Message-----
> From: cst-2-admin@srcf.ucam.org [mailto:cst-2-admin@srcf.ucam.org]On
> Behalf Of Alvin
> Sent: 01 May 2002 16:58
> To: cst-2@srcf.ucam.org
> Subject: [CST-2] Autocorrelation
>
>
> Hi guys, I'm using the list!
>
> Yeah, doing a bit of Info Theory and coding, in 2001 Paper 9 Q 10
> there's a
> big section on autocorrelation.  Searched for ages to actually find the
> autocorrelation integral, but can't really find any details that
> would help
> in answering the question fully.
>
> This question will probably not come up again this year but still
> would like
> to know the answer,
>
> Cheers
>
> Alvin
>
>
>
> _______________________________________________
> CST-2 mailing list
> CST-2@srcf.ucam.org
> http://www.srcf.ucam.org/mailman/listinfo/cst-2
>