Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!usc!orion.oac.uci.edu!ucivax!megatek!toddh From: toddh@megatek.UUCP (Todd Heckel) Newsgroups: comp.graphics Subject: Re: checking whether 2 2D polygons overlap Message-ID: <849@portnoy.megatek.uucp> Date: 8 Dec 90 22:23:18 GMT References: Organization: Megatek Corporation, San Diego, Ca. Lines: 28 From article , by gessel@masada.cs.swarthmore.edu (Daniel Mark Gessel): > > Actually, anti-aliasing a zbuffer is easy. But the technique I would use is as > brute force and stupid as a z-buffer itself. Render the image into a buffer > with twice the resolution in each direction (or more) and have video hardware > that averages from 4 pixels (or more). This is probably the most > general anti-aliasing technique, and I'd like to see it in things like display > postscript and stuff. But it does use 4 (or more) times as much memory and > processing power. > > Dan > -- > Daniel Mark Gessel Independent Software Consultant > Internet: gessel@cs.swarthmore.edu and Developer > I do not represent Swarthmore College (thank God). Note that I said "real-time". I have designed systems that do exactly as you described, at either four or sixteen times the display resolution (two by two or four by four). While the images look great, they take four or (gasp) sixteen times as long to generate as an un-anti-aliased image. Newer techniques based on sorting and bit-masks, like the A-buffer, are currently being adapted for real-time image generation use. -- Todd Heckel Megatek Corporation uunet!megatek!toddh 9645 Scranton Road or San Diego, CA 92121 toddh@megatek.uucp