[Tig] new on the TIG wiki

[Bob Friesenhahn]

On Fri, 11 May 2007, Simon Burley wrote:
>
> No, it doesn't. Bob, please point me to exactly where Nyquist theorum
> deals with image detail. There I was thinking that the Nyquist-Shannon
> theorum only dealt with frequencies.

Nyquist-Shannon theorum deals with sampling theory (see 
http://en.wikipedia.org/wiki/Nyquist-Shannon_sampling_theorem).  While 
Nyquist and Shannon were not particularly interested in images, they 
were very interested in how to represent any sort of signal using 
discrete samples.  Computer images are comprised of discrete samples 
(pixels) and therefore adhere to the sampling theorum.  Note that the 
WikiPedia page I referenced does discuss application to images and why 
you should care.  It also discusses why correct low-pass filtering is 
necessary in order to avoid interference patterns.

> What I find most revealing is that NO-ONE has dealt with exactly how one
> should do to end up with the decimated, required resolution image.

I think that this lack of a conclusion is due to the nature of film. 
No one should doubt that the absolute minimum scanning resolution is 
dictated by Nyquist theory.  The open issue is how much further beyond 
that minimum (e.g. for a particular reproducible resolution) is 
necessary to capture all useful data from film?

Bob
======================================
Bob Friesenhahn
bfriesen at simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer,    http://www.GraphicsMagick.org/