[comp.graphics] Dithering Questions

johnm@trsvax.UUCP (12/05/88)

I'm looking for information on a specific kind of dithering.  Currently I am
familiar with and have experimented with:

	1) Conventional reduction in resolution (i.e. use four pixels to
	represent one and thus get five levels of intensity).

	2) Noise adding methods (like digital dot diffusion and ordered dithers
	which don't cut down your resolution but do add fuzziness to the image).

	3) Continuous correction methods (like Floyd-Steinberg in which you
	attempt to make up for errors on previous pixels by altering the
	intensity of pixels near it).

What I want is a method which combines the 1 & 2 (or even better, all three).
Suppose I have several intensity levels per pixel already but I want even more.
I'm willing to put up with the small amount of noise introduced by type 2
methods but unfortunately can't see how to adapt them to take advantage of
multiple color levels.  Does anyone have any references or perhaps brief
descriptions of methods that combine these or perhaps that take an entirely
new tack on the problem?

John Munsch