Not long ago, the only people who could afford to play around with fancy computer graphics were big-company designers, movie and video producers, and students in universities like the Massachusetts Institute of Technology (M.I.T.). Now nearly anyone can buy the tools to create high resolution color graphics on a desktop computer.
What would you buy? What would you do with it? What's this new science all about? We put those questions to computer graphics engineer Jeff Kesselman (who's also a son of ours). Here's the first in his guest articles for this column.There's been a great deal of hype in the past year or so about microcomputer graphics systems. In order to cut through the sales pitch and make an intelligent buying decision, you need to know a little about what computer graphics is. That's more than most computer advertising and sales people know.
Computer graphics can be broken down into two-dimensional (2D) systems and three-dimensional (3D) systems. Since all pictures made on a computer come out on a flat screen or piece of paper, those are always two-dimensional. The difference between 2D and 3D systems is how the pictures are created.
A 2D graphics system is much like a painter's canvas. It's a program that enables an artist to draw pictures on the screen using a mouse, graphics tablet, light pen, or cursor keys. The first such system, developed by Dr. Ivan Sutherland at M.I.T. in the sixties, was called Sketchpad. All 2D graphics programs designed since then have just improved and refined his original idea.
Modern 2D programs (sometimes called `draw' programs) simulate tools and techniques available to conventional artists, such as thin-line pens, thick-line pens, collage, airbrushing, and color blending. In addition, they add many techniques uniquely devised to take advantage of the abilities of a computer. Some can stretch, warp, or rotate images. Some can even automatically generate lines in perspective.
I use Deluxe Paint II from Electronic Arts on my Amiga computer for making 2D illustrations.
A 3D system is not like a canvas, but like a lump of clay in a phantom photography studio. In the late seventies, Professor Nicholas Negroponte at M.I.T. set out to design a machine to automate architectural design. While the machine was never successfully built, the 3D graphics systems of today are direct children of that early project.
The user of such a system can design 3D objects much as a sculptor would sculpt clay or a carpenter cut and assemble wood planks. To describe these objects to the computer, he uses what's often called `solid modeling' techniques. What he assembles is, in actuality, a mathematical model of the solid object, since the computer deals only with numbers.
Next, the artist gives the computer measurements that position the objects in relation to each other and/or a background. She then can tell the computer where she wants the lighting to be and how the lights are to be set, just as in a photography studio.
The last piece of information the computer needs is the position and orientation of the `camera' or the viewer's eye. Given all this information, the 3D design software mathematically simulates a camera taking a picture and displays the results.
3D systems are slower than 2D systems, but are more automated in their image creation. Here, if an artist wants several pictures of a given object from several different angles, he does not have to draw each picture separately. The computer software can move the camera eye and create the picture automatically.
Computer animation, which is probably the most glamorous type of computer graphics, is based primarily on 3D graphics. To make the starship movements seen in the television show Star Trek: The Next Generation, or the Stained Glass Knight in the movie The Young Sherlock Holmes, artists specified motions for the camera and/or the objects. The computer then used these motions to calculate the position of each object in 1/24 or 1/30 second intervals (1/24 for film, 1/30 for TV). It then generated one picture or `frame' for each step in the motion.
Some automated motion is possible using 2D systems. While the objects themselves are flat, some software lets you move and rotate these flat objects in the two dimensions. Some programs also allow you to vary the shape of an object over time, for instance to shrink an octagon down to a square, then enlarge it into a triangle which you can zoom off-screen. The main limitations of 2D animation systems are their lack of three-dimensional objects and 3D movement.
There is a compromise kind of animation software called 2.5D. (Honest!) Actually, they're really just 2D programs that cheat a little to try and hide that fact. They often seem to move objects `towards' or `away' from the viewer, but the objects are still basically flat - and appear that way if you rotate them so that an edge is facing the screen.
So conceptually, computer graphics is easy to understand. It's done with either sketch pads or cameras. If you need to know more, our favorite nontechnical source is S. Klein's bimonthly magazine Computer Graphics Review (617-443-4671).
As a service to readers, the columnists answer questions and send a checklist of back issues if you enclose a stamped self-addressed envelope: TBC, 4343 W Beltline Hwy, Madison WI 53711. You can read back issues of this twice-weekly column at the electronic library, NewsNet, reachable via computer plus modem over phone lines. For NewsNet information, 800-345-1301.