Path: utzoo!utgpu!water!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!rutgers!apple!bionet!agate!ucbvax!decwrl!labrea!polya!kaufman
From: kaufman@polya.Stanford.EDU (Marc T. Kaufman)
Newsgroups: comp.fonts
Subject: Re: Fuzzy fonts??
Message-ID: <4788@polya.Stanford.EDU>
Date: 31 Oct 88 16:59:07 GMT
References: <3448@pt.cs.cmu.edu> <6204@pogo.GPID.TEK.COM>
Reply-To: kaufman@polya.Stanford.EDU (Marc T. Kaufman)
Organization: Stanford University
Lines: 21

In article <6204@pogo.GPID.TEK.COM> curtj@pogo.GPID.TEK.COM (Curt (Jutz) Jutzi) writes:

>	Peter Karow from URW addresses this topic in his book on fonts.
>	I don't have the book with me or I would give you more information.

>	The concept is quite simple.  You generate a larger raster image of 
>	the character and reduce it by averaging the pixels around it.

>	ie. for a display that has 4 gray levels you would generate a 
>	character 2 times the point size and simply take groups of 2x2 
>	matrices within that image to generate none display pixel.

I have done this with 16-levels of gray (i.e. 4x scaling in x and y, then
count the number of black pixels in each 4x4 cell to get a gray value) for
the Macintosh.  This was probably overkill.  It works well enough that you
can read a font like Zapf Chancery when it is only 6 scan lines high.  The
down-side is that it is slower to display than a 1-bit font, the gray is not
a true blend-to-background, and the technique does you no good on devices that
don't have gray (such as Laserwriters).

Marc Kaufman (kaufman@polya.stanford.edu)