On Tue, 4 Dec 2001, Rafael 'Dido' Sevilla wrote:
..
> I remember a technique called Lanczos smoothing used in the design of
> digital filters. It's used to attenuate the ripples in an FIR filter
> that happen when you truncate the (in general, infinitely long) series
> of Fourier coefficients that represent the filter (the Gibbs
> phenomenon). There are "sigma factors" computed here that are
..
There's no need for speculation. These functions are just used for
sampling the original image. For example a simple "dumb" linear
interpolation uses a "square wave" as its sampling function (or, a
low-pass filter) because it just averages adjacent pixels to get the
interpolated value. Unfortunately because of the sharp edges it is
non-ideal.
A wide variety of functions are used for interpolation. In general -- as
you said -- they all have different characteristics but the primary aim is
to minimize Gibbs phenomenon. ImageMagick implements most of them --
Hamming, Hanning, Bessel, Mitchell, Lanczos, and so forth. I thought
"Lanczos sampling" meant something fantastic, but basically it just means
you use the sampling function
L(x) = sinc(x) sinc(x/2) where |x| < 2
0 otherwise
--
Orlando Andico <[EMAIL PROTECTED]>
Mosaic Communications, Inc.
_
Philippine Linux Users Group. Web site and archives at http://plug.linux.org.ph
To leave: send "unsubscribe" in the body to [EMAIL PROTECTED]
To subscribe to the Linux Newbies' List: send "subscribe" in the body to
[EMAIL PROTECTED]
