Path: utzoo!attcan!uunet!lll-winken!lll-lcc!ames!pasteur!agate!labrea!polya!kaufman
From: kaufman@polya.Stanford.EDU (Marc T. Kaufman)
Newsgroups: comp.graphics
Subject: Floyd-Steinberg dithering
Message-ID: <4827@polya.Stanford.EDU>
Date: 3 Nov 88 00:37:28 GMT
Reply-To: kaufman@polya.Stanford.EDU (Marc T. Kaufman)
Distribution: na
Organization: Stanford University
Lines: 13

(the retention time of news articles is strictly less than the time between
posting and needing the information)

I have seen several articles that give various error distribution algorithms
for Floyd-Steinberg dithering (e.g.: 1/2 to the right or 7/16 to the right,
some down, some down and to the right, some down and to the left, etc.).  Of
course I lost my copies of these amounts to each neighboring cell.  Could
someone send me a copy of the numbers.  How sensitive is this algorithm to
changes in the distribution, with respect to generation of artifacts
(or is it severely case dependent)?

Thanks for your help.

Marc Kaufman (kaufman@Polya.Stanford.Edu)