                          Worlds of the Imagination:
                       Virtual Reality Graphics Software
                         Copyright 1995 by Herb Chong

 It's now been several months since I have been adding a graphic into each
 issue of WindoWatch.  It's time that I explained how I created them and
 how you can go about creating some yourself. At this time, doing virtual
 reality graphics like those I've been showing is not for the impatient. It
 takes hours and sometimes weeks to get something just right. Each of the
 images may take hours or days to render on a fast machine. You need to be
 able to think and visualize in three dimensions and have an artistic
 inclination, even if you don't have a huge talent.

 Myst and Doom are the most well known of the computer games that use
 virtual reality. Myst's animations and images were all done on a Mac using
 an expensive professional software system. If you've seen Myst screens
 before, you'll see that they are hyper-realistic. Surfaces are a little
 too clean and sharp. On the other hand, it's pretty realistic how objects
 move and change shape around you when you move toward or past them. Doom
 is a different, less ambitious, virtual reality game. Absolute accuracy
 isn't the point. Notice that nothing casts shadows?  However, its
 backgrounds and surfaces do otherwise and change properly as you move
 about the dungeon.  More import-antly, they change quickly so that on
 faster machines, there is no flicker at all, even though, the differences
 in perspective is generated as needed. By contrast, Myst uses
 pre-generated animations for all its movement. It blend them together
 seamlessly so that it looks like you are just moving, not playing
 carefully pre-planned m vements of a camera.

 Virtual reality graphics is computer art, although some say that computers
 and art are mutually exclusive. With the tools like those I will describe
 in this article, you will be able to create nearly everything you can
 imagine. The better your ability to create new ideas and new ways of
 looking at things, the more amazing your imagery will be. I'll start off
 with some basic concepts of computer graphics and move onto the tools you
 need to get started. Then, I will show a gallery of some images from some
 projects I have worked on, or that other people have worked on and that I
 have captured. Finally, I will give a list of software that I have used,
 interesting places to visit on-line, and magazines to check out. Doing
 virtual reality graphics is tremendous fun and lets you turn some of the
 images in your mind into images others can admire.

 Here is a sample, just to give you a flavor of what is to come. You can
 create an image just like this, with no artistic talent at all, in about 5
 minutes. You do have to know a bit about how to use your program, but
 that's a different kind of talent. This is a cube, a cylinder, and three
 spheres. They are resting on a plane surface (the grass) and surrounded by
 a hemisphere (the sky and clouds). There are 2 lights illuminating the
 scene, and some of the objects glow a little too. The hardest part of
 doing this picture was figuring out how to get a perfect hemisphere and
 then trying to find a nice picture of sky and clouds to put on it. I
 created this picture using trueSpace 2.0, a program at the high end of the
 price range at about $450 to $500 in stores.

 How does one go about creating an image like Figure 1. Let's start at the
 beginning. What do I mean by virtual reality graphics software? It means
 many things to many people. Some people insist that it's not virtual
 reality unless you are wearing goggles, gloves, and experienc-ing the
 sights and sounds, and maybe even the smells, of a world as you move
 through it. That's pretty high tech stuff and, frankly, not only very
 expensive, but not nearly as realistic as the advertising would have you
 believe. Some people call Doom a virtual reality game. Well, it is, but
 only in the sense that you can move about and interact with a world
 someone else has created for you. The Doom construc-tion sets make worlds
 or levels of your own, but there are only certain things you can do.

 The kinds of virtual reality graphics software I am going to tell you
 about are the kinds that an artist would use to create images and movies
 of objects and scenes. The art might be for its own sake, as part of
 another image, or it might be used to visualize the sequence of events in
 an accident investigation  called forensic graphics animation. There might
 be only a single image or a sequence made into a movie. Another limitation
 on the software I'm going to talk about is that it can't be too expensive.
 The prices of the software I'm going to describe range from free to about
 $500. This covers the range from very simplistic software to complex
 software used by professionals. The truly professional software used to
 make movies in Hollywood start in the several thousand dollar range and go
 up. Normally, commercial virtual reality movie studios have specialized
 hardware to help them make their movies faster. I'm not going to look at
 any of that software or hardware.

 Modeling

 The first thing that virtual reality graphics software has to do for you
 is to be able to model or represent the world that you want to create as
 an image or movie. The modeling tool can be an integral part of a program,
 a separate program part of a virtual reality graphics package, or a
 completely separate program that can create objects you can load into a
 scene builder. The modeler might be so simple that it only allows you to
 import 2D or 3D clipart to a full-blown CAD program that can completely
 design an office building. No matter what it's capabilities, the modeling
 program's job is to create and manage objects for you.

 Most modeling program have a set of primitives that they can create if you
 don't import any clipart. Usually, they are simple shapes like sphere,
 circle, cube, cone, flat plane, and cylinder. A good modeling program lets
 you use these objects to build almost anything you can imagine. All
 modeling programs let you load pre-created objects and modify them to some
 degree. Some programs allow only one object format, usually a proprietary
 one. Most allow you to use files from CAD programs too. The higher end
 tools also understand some of their competitor's object formats and import
 them.

 What do you do with objects in a modeling program? For most such tools
 used with virtual reality graphics software, one of the most important
 things they do, is let you alter an object's geometry. This means
 reshaping the object. If a lamp is just a little too big, you might want
 to shrink it. If you're framing a picture, the frame might be a little too
 thin. Modeling programs also let you deform objects, sometimes
 arbitrarily, sometimes only in certain ways, and they can let you stretch
 and shape and twist parts to your heart's content. If you want to bend a
 cone to touch its toes, you should be able to do it. You'd be amazed at
 what shapes four spheres and a cube can be end up like with about two
 minutes manipulation. I'll show you in one of the images in the art
 gallery at the end of this article. The remaining manipulation a good
 modeling program can do is to break apart groups or join together objects
 into groups.

 The next important thing that a modeling program does for you is that it
 allows you to apply textures or materials to the surfaces of an object.
 Some programs will let you create your own textures. Others give you only
 a fixed palette to work from. Whatever the modeling program's texture
 manipulation capabilities, they all allow you to pick a texture and apply
 it to an object.

 Texture in virtual reality graphics programs is a more general concept
 than common English use of the word implies. Before a texture is applied
 to an object, the object's not visible, just like some ghosts are supposed
 to be. Once a texture is applied, light can interact with it. These are
 the lights that you work with in the virtual reality graphics software,
 not the lights in your office around your computer! What are some examples
 of textures? Well, water is a texture. An entire object, the contents of a
 swimming pool, is made of water. It has certain properties and certain
 interactions with light. Concrete is another texture. The important thing
 to remember here is that to virtual reality graphics software, a texture
 is a solid and not just a surface property. Yes, you can apply a texture
 only to the surface of an object, but usually that is not sufficient.
 Materials that allow light to pass through them like water and glass need
 to be made entirely of a material, not just painted on the surface with a
 coating.

 The technique used to generate the scene governs just how much of a
 material's bulk properties need to be modeled. There are two major ways of
 rendering images in software of the target price range, shading, and ray
 tracing.  I won't get into the details of either until I finish talking
 about scenes, but suffice it to say for now that shading requires less
 modeling of a material's properties than ray tracing. More details later.

 The remaining features in modeling programs vary so much that I'm not
 going to describe them. Suffice it to say that the kinds of things you can
 do to text with a text editor or word processor, you can do to objects
 with a good object modeling program, like cut, paste, rearrange, move, and
 so on. Time for some more pictures. Figures 2 and 3 are images I created
 using Visual Reality 2.0. It's a program that typically retails for under
 $200. It comes on 7 CD-ROMs with a huge 3D clipart library and pre-built
 scenes. Figure 2 is from the Simply Scenes CD Jurassic Adventure. It's a
 scene full of plants and landscaping and all set for populating with any
 one of ten dinosaurs (included but not in this image). The volcano looks
 pretty realistic, doesn't it. Figure 3 is from the Orbit City Simply
 Scenes CD. It's a city floating in space. This is a view looking out from
 the center of the city past the hydrodome and the landing pad to the
 planet surface below. You can take objects out of and put your ow objects
 into the scenes.

 Visual Reality comes with two other Simply Scenes collections, the Starter
 Pack, and Northern Castle. Imagine the computer games you could create
 scenes for in a dark and gloomy castle!

 Scenes
 Once you have a bunch of objects, you still don't have an image, and
 certainly not a movie. Scenes are where everything starts to come
 to-gether. Although terminology varies in the manuals, a scene is a
 collection of objects, their positions, and orientations, lights,
 back-ground, and a camera. The lights have intensities, possibly direction
 and scatter, and sometimes shadows. The camera has a position,
 orientation, and a focal length. Varying the focal length is zooming, just
 like you would with a zoom lens on a camera. The scene editor might be a
 separate program, or it might be the same program as the modeling program.
 It's useful to be part of the modeling program be-cause one of the things
 you do when building scenes is move objects around, perhaps rotating and
 scaling them, so that they are arranged just the way you want them to be.
 You move lights around in a similar way to get the right effect too.

 Most of the scene editors work in wire frame mode. This means that you
 don't see the object as a solid object. Instead, you see a skeleton of
 each object. When there are many objects in a scene, it can sometimes
 become very hard to tell apart one object from another. Newer pro-grams
 have a solid viewing option. Objects are drawn with surfaces visible so
 that you can tell what the relationships between objects are. You can see
 if two objects are touching or one is being blocked more easily. However,
 this costs lots of CPU power to do. Don't even think about it without a
 fast Pentium system and a very fast video card.

 Building good scenes is easier if you have an eye for composition, but
 since this is all done on a computer, you can render a test image whenever
 you want to see if things look good or not. Most modeling programs will
 allow you to manipulate the camera and move it around to get the right
 view. Unlike real life photography, if the mountain is too big, you can
 shrink it a little. Talk about power! Don't like the clouds, get new ones.
 It's easier than hiring and firing extras in a real movie.

 Speaking of mountains, it's time for another picture, of mountains, of
 course! Figure 4 has appeared in a prior issue of WindoWatch. It's a
 computer generated picture of Yosemite Valley looking from west to east,
 near El Capitan. Although it is quite realistic, you can tell that it's
 computer generated. The trees have no shadows under them, the leaves are
 too sharply visible, and there are "creases" in the rocky cliffs in the
 background. I created this image using Vista Pro for Windows 3.0. It's a
 program that takes US Geological Service Digital Elevation Model (DEM)
 data of sections of land, and allows you to model the surface and populate
 it with plant life. You can also get a DOS version. It's slightly faster
 than the Windows version, although the Windows version has some additional
 features. If you buy the CD-ROM version, you get maps of most of the US at
 low resolution and some selected areas at higher resolution. If you don't
 like pictures of the Earth, on the CD-ROM are compatible maps covering
 about 1/2 of the surface of Mars. If that's still no enough, there are
 simple ways to generate your own landscapes.

 Rendering

 Leave it to computer types to come up with a complicated name for a simple
 idea. When you render the image, you do the equivalent of clicking the
 shutter button on a camera. In this case, the camera is a virtual one in
 the virtual world you have created. How fitting! Under the covers though,
 a tremendous amount is happening. Because many people spent many years
 figuring out how to do something they thought was simple, they couldn't
 just call it something ordinary, so everyone in the know now calls it
 rendering.

 There are two types of rendering used in virtual reality graphics software
 of the target price range, shading and ray-tracing. They take two opposite
 approaches to how to figure out what appears in your images. Both have
 their advantages and disadvantages. Expensive programs sometimes use a
 lighting model called radiosity to generate realistic images. It takes
 tremendous amounts of CPU power to use and for some situations, generates
 very unrealistic images. Shading and raytracing have been around for a
 while and, compared to radiosity, much faster.

 Shading first figures out from the viewpoint, view direction, and view
 focal length, what objects are visible. If an object is partly visible,
 the program conceptually cuts the object off so that only the visible
 portion is left. Next, it examines every visible surface and looks to see
 what lights are shining on it. If there are any, it uses a mathematical
 formula to figure out, given the color of the light and the surface
 material, what the apparent color of the part of the surface should be. If
 there is no light, then the program might assume a default ambient light
 and calculates the resulting color. Then it figures out, for each dot
 (pixel) in your image, what is visible from the dot, and colors the dot
 appropriately. This was the first solid view of objects generated for
 computers. Prior to shading, only wire frame models were used. There was
 neither the theory nor the computing power available to do any better.
 Shading was invented in the late 1960's and early 1970's.

 Ray tracing takes a completely different approach. It goes from the view
 point, in the direction of the view, out in the exact direction of each dot
 that will appear in your final image. This is a backwards ray of light
 tracing back from where it would have hit the view point, out through the
 target dot. It keeps going backwards until it hits an object. When it does,
 the program figures out how much of the ray would have bounced, any change
 in color caused by the texture, splits into two and refracts as necessary,
 and keeps going back until it hits a light, doesn't run into any more
 objects, or has been dimmed so much by its reflections that nothing could
 have come from that ray and made it to the camera. The last case is easy.

 The program paints the dot black. The first case is almost as easy. Now
 that the program knows that it has hit a light, it knows the color and
 brightness of the light. It retraces the path from the light to the
 viewpoint, altering the brightness and color as appropriat , until it
 reaches the dot. Whatever the color the light last was set to is the new
 color of the dot. If the ray has split into two because of refraction and
 reflection, the two contri-butions are added together and the blended ray
 ends up at the dot. This can happen many times to build up the color of a
 single dot if the scene has plenty of partly transparent, refractive, or
 reflective objects. What does this really mean? Time for more pictures.

 Figures 5 and 6 are of the same scene. Figure 5 was shaded while Figure 6
 was raytraced. There are two important differences you should note: the
 shiny cone isn't reflecting anything except light in the shaded version,
 and the colored glass cube is rendered completely differently. Look at the
 base of the shine cone in the raytraced version. You should be able to see
 on the bottom right and tiny part of the sphere reflected back. Similarly,
 looking through the purple glass cube, you can see that the cone is
 reflecting both the picture on the cube and the back of the purple cube.

 Finally, you can see on the faces of the cube both reflected and refracted
 images of the cube and the sphere. That's the difference you get with
 raytracing. So why doesn't everybody do raytracing by default? It took
 about 30 seconds on my system to render the shaded image at 640x480. It
 took about 15 minutes to do the raytraced version. This is for a very
 simple scene. In the scene I'm working on for my current animation, I am
 working with several thousand objects at once. It takes just over 100
 megabytes of RAM to load the scene and another 100 megabytes of RAM to ray
 trace it. It takes about 6 to 10 hours to raytraced a single 640x480 frame.

 There is one refinement left that is commonly used in raytracing
 pro-grams. Notice that there are no shadows in both images. Although one
 can generate realistic shadows for shaded images, ray traced images are
 where the best details are from . Figure 7 is the same scene raytraced
 with shadows enabled on the lights. Notice that now you can tell that the
 glass cube is floating in the air. There is a shadow under the entire
 cube.

 Animation
 If you have a fast enough computer, the really exciting stuff happens when
 you begin doing animations.  Animations are nothing more than a sequence
 of images played back sufficiently fast that you can't see the individual
 frames changing. Instead, you just get the impression of a continuous
 image that is changing. This is how both television and motion pictures
 both work.

 With a good rendering program, you can create single frames or whole
 bunches of frames. However, if you try to use a plain rendering program to
 make a movie, you're going to be disappointed. Just like trying to use an
 ordinary camera to make movies by slowly moving and taking a picture and
 playing back in a movie projector, you get too rapid and jerky changes in
 the picture. To get smooth animation, you have to move the camera or
 objects on a smooth path and with small enough movements to make the
 motions not too abrupt. For normal amateur computer video work, this means
 you need to gener-ate frames at 15 per second.

 Time to do some arithmetic. A twenty second clip is 600 frames. If you run
 a small image and stick with simple scenes and rendering techniques, the
 frames might take about a minute each. Six hundred minutes is ten hours of
 rendering time. Don't plan to do animations unless you have a fast
 computer or a lot of patience. I normally generate video animations at a
 size of 320x200 pixels. This translates into a 192K output BMP at 24-bit
 color (no animation program I mention in this article outputs less). Six
 hundred such frames occupy at least 112 megabytes of disk space. You need
 lots of disk space to do this too. I reserve a 1GB drive just for
 outputting my animations. I have created ones with up to 5000 frames. That
 almost filled the output drive.

 To get the smooth paths you need for smooth animations, you need a program
 that can generate good camera movement paths. The more sophisticated tools
 will have animation tools designed just to edit paths. These days, almost
 all program work with something called B-splines. Corel Draw is an example
 of a program that generates and works with B-spine curves. However, Corel
 Draw's drawing package only does two dimensional B-splines. For animation
 work, you need three dimensional B-splines. Programs like VistaPro and
 Visual Reality have separate modules that animate parts of a scene.
 TrueSpace has the animation builder all included. Besides animating the
 camera, you can animate other objects in the scene. Objects can move on
 their own spine path. So can lights. Fancier animation programs like
 trueSpace can animate textures too, so that an object's surface can change
 as you go by it. Objects can even deform from frame to frame of an
 animation.

 If you had to specify every state of every object in every frame of an
 animation, you would never get anything interesting done. Imagine trying
 to co-ordinate six hundred individual movements of only one object for the
 twenty second hypothetical example. Instead, animation building tools work
 with something called key frames. The term is borrowed from the cartoon
 motion picture industry. The main artists drawing an animation sequence
 don't draw every frame. Instead, they draw just the important ones. At a
 certain time (frame number) they know that something has to be a certain
 way. For instance, the hammer head starts at frame 3 and hits the glass at
 frame 9. A bunch of apprentice artists, called 'tweeners, fill in the
 missing frames so that everything looks smooth when all the frames are put
 together. They're called 'tweeners because the fill the "in between"
 frames.  Fortunately for you and I, we own computers. Yes, you still have
 to figure out the keyframes and get them right, but then you can tell the
 computer to fill everything in.

 If all you want to do is to fly past something for the six hundred frames
 on a straight line, you just tell the program where it is supposed to be
 on frame 0 and where it is supposed to be on frame six hundred. The
 computer takes care of everything else. If you want a curved path, you
 need to do more work. If the line between two places doesn't have to be
 straight, all kinds of things can happen. This is where B-splines come in.
 They are a way of specifying smoothly curved paths with a few "control
 points" and some "tension" adjustments. The control points are exactly
 where you want to have a keyframe. By adjusting the tension of the path on
 either side of a key frame, you can come up with amazingly complex paths
 with only two key points. If you want to get fancy, you might want to use
 several dozen control points, but you don't need to. The computer knows
 the path and just divides it up smoothly based on how much time (number of
 frames) happens between the points.

 Unfortunately, animations are huge. Figure 8 is the first frame of an
 animation I have been working on for a while. I created a ten second clip
 of it and compressed it as much as I could and still preserve enough
 detail to see what the camera is flying past. It is just over 800Kb in
 size. This is in Video for Windows AVI format with the Cinepak compressor.
 If you have any version of Video for Windows 1.1c or higher, you can play
 back the file. Video for Windows has other compressors which make
 different tradeoffs of quality versus file size. Most Windows rendering
 programs allow you to choose what output format to use. However, none of
 them explain which animation format makes which tradeoffs. MPEG does the
 best compression for a given quality or file size, but there are as of yet
 not that many good MPEG players available.

 Fancier rendering programs allow much more than path, object, or texture
 animation. Figure 9 is similar to one that has appeared before in
 WindoWatch and was used on the WindoWatch home page for a while. It was
 created using Typestry for Windows 2.0. This is a prog-ram specialized
 just for raytracing and animating text. You would probably use it to do a
 logo for TV.

 One thing you can see in this picture is that there is fog surrounding the
 letters. With Typestry, it's possible to have the fog flow by, rippling
 and swirling as it passes. Talk about neat effects! You can also put the
 type on a flag and have the flag blowing in the wind, and some other
 effects too.

 Summary

 So, here we are. I've shown you some of the art I have done using virtual
 reality graphics software and given you some insight into what goes into
 creating some of the images I have done. I've also mentioned some of the
 software I have used and how much it takes to use it. This article is a
 bare overview of what is available and what can be done in virtual reality
 graphics. You should have an idea now of what terms are used by people and
 what to look for when trying to decide if you should jump in or not. If
 you intend to do more than dabble in virtual reality graphics, be prepared
 to spend the time to learn to use your software well, and be prepared to
 spend lots of time at it. Some of the images I did for this article took
 more than a twenty-four hours to render on a 486/66. I have done images
 which have taken 60 or more hours to render. The 5000 frame animation
 took, with interruptions, more than 11 hours to do the low resolution
 preview and 10 days to render the final version on my Pentiu . Raytracing
 also takes plenty of RAM. Several of the images in the gallery took more
 than 80Mb of RAM to load and another 60MB of RAM to raytrace. Shading
 would have been a bit less expensive at about 25MB more. If you really
 want to design and build worlds of your own, you need to invest plenty in
 it, but how often do you get to design a world just the way you like it?

 Abbreviated List of Resources

 3D Design - a magazine devoted to virtual reality graphics software
 published by Miller Freeman. Contact them on Compuserve in the DDDMAG
 forum, 600 Harrison St, San Francisco, CA 94107, or 1-415-905-2200.

 Caligari - trueSpace 2.0 and trueSpace/SE - modeling, rendering, and
 animation packages. The full version of trueSpace has a street price of
 under $500. The SE version is basically just trueSpace 1.0. It has a
 street price of under $100. Both versions allow you to pick shading or ray
 tracing rendering techniques. I like trueSpace because everything is in
 one module. It is really good at rendering and modeling changes on the fly
 so that you can try things as they occur to you. You can check out their
 web page at http://www.caligari.com. They are also on Compuserve in the
 Graphics Vendor C forum. Otherwise, you can contact Caligari at 1933
 Landings Drive, Mountain View, CA 94043 or 1-415-390-9600.

 Corel Corporation and Ray Dream Software - Corel Draw 6.0 includes a
 package called Corel Dream. This is a repackaging of Ray Dream Designer by
 Ray Dream Software. Both do modeling and rendering using raytracing. If you
 want to do animation, you need to look at Ray Dream Studio. It is Ray Dream
 Designer with added animation, clipart, and textures. Corel Draw 6.0 has a
 street price under $500. Ray Dream Designer 4.0 is available for under
 $100, while Ray Dream Studio 4.0 is under $300. Contact Corel at 1600
 Carling Avenue, Ottawa, Ontario, Canada K1Z 8R7 or 1-613-728-8200.

 Pixar - Typestry for Windows - it's a modeling, rendering and animation
 package for typefaces only. The other drawback is that it only raytraces.
 This means that it can take a long time to do images. You can get it for
 under $200 street price. If you need to do logos or anything that involves
 fancy text, Typestry is your program. Contact Pixar at 1001 West Cutting
 Boulevard, Richmond, CA 94804 or 1-510-236-4000.

 Virtual Reality Labs - VistaPro and Makepath Flight Director - Vistapro is
 a modeling and rendering package only. If you want to do animation, you
 need to use Makepath Flight Director. Vistapro costs under $90 street
 price, but you'll most likely have to pay list price for Makepath Flight
 Director as very few places carry it. You'll have to get it from VRLI
 directly. List price is $69. You can contact VRLI at 2341 Ganador Court,
 San Luis Obispo, CA 93401 or 1-805-545-8515.

 Visual Software - Visual Reality 2.0 and Simply 3D 2.0 - these two
 packages also do modeling, rendering, and animation. Visual Reality is the
 full featured package and can be had for under $200 street price. Simply
 3D is under $40 street price. Both are powerful rendering and animation
 tools. Simply 3D is much less capable at modeling, but it's really cheap.
 Both come with a big collection of scenes and clipart for you to use. The
 only thing I don't like about either of these is that you need to work in
 separate programs for each phase of the rendering process. You need to get
 models in the modeler and then assemble the scenes in the rendered and
 scene editor. This means more planning to get the right effects. Programs
 that are all in one like trueSpace allow more experimentation. However,
 Visual Reality has more control over its animation. You can contact Visual
 Software on the Animation Vendor A forum on Compuserve, at 21731 Ventura
 Boulevard, Suite 310, Woodland Hills, CA 91364, or 1-818-8 3-7900.

 POV-Ray Team - POVRAY is a freeware ray tracing program. You can do
 tremendous things with it and you have huge amounts of control over the
 objects you work with. There are two problems however. First, it's only a
 DOS program. Second, it uses an ASCII scene description file as input. You
 have to work out the coordinates of everything yourself. However, free is
 free. You can obtain POVRAY from the Graphics Developers forum on
 Compuserve, many BBSs, and from Walnut Creek CD-ROM. Walnut Creek
 publishes "The Official Poverty CD-ROM". It contains the software, all the
 documentation, hundreds of scenes for you to render, hundreds of images to
 look at, and dozens of animations too. You can check out Walnut Creek on
 the Internet at http://www.cdrom.com. Otherwise, you can contact them at
 4041 Pike Lane, Sta D-902, Concord, CA 94520 or 1-510-674-0783.


 Figures 15 through 20 are a project I am working on right now. These are
 all frames taken from a scene that I am animating. Each pair of frames are
 shaded and raytraced versions of the same scene. See if you can tell the
 difference. I did these images in trueSpace 2.0. The pic-tures on the wall
 are all photographs of mine that I have scanned in and converted into
 textures. The models of people from Noumenon Labs, KETIV Technologies, and
 Modern Medium. You can contact KETIV at 6601 NE 78th. Court, Suite A-8,
 Portland, OR 97218 or 1-503-252-3230. You can contact Modern Medium at 580
 West 8th. Avenue, Eugene, OR 97401 or 1-503-343-4281.


 As anyone can plainly see, this article of Herb Chong's  represents both a
 major effort and a major contribution.  We are very proud that he allowed
 WindoWatch to publish his work.  In the last weeks, Herb has become a
 member of IBM's research team. He is the Contributing Editor of WindoWatch
 and will be the Guest Editor of The WindoWatch Anniversary Issue
 or Vol.2 No.1 mid-January of 1996

                                  ww


