DETAILS


COLUMNS


CONTRIBUTIONS

a a a

GAMING & GRAPHICS

Vol.32 No.1 February 1999
ACM SIGGRAPH



Looking for Some Art Amidst the Technology



Richard Rouse III
Paranoid Productions


February 99 Columns
Entertaining the Future Visfiles


Richard Rouse III
Previous article by Richard Rouse III Next article by Richard Rouse III

Just why is Myst so popular? This seems to be a question that has stymied the computer gaming industry for the last couple of years. Cyanís adventure game Myst, with its finely rendered 3D appearance and ďslideshowĒ style gameplay has continued to sell well to this day, while games which on the surface appear to be of quite similar content and quality have sold dismally, not even making blips on best seller charts. The dominant industry opinion at this point seems to have chalked the whole Myst phenomena up to timing and dumb luck.

Iím certainly not one to be overly concerned with sales of products as an indicator of their actual quality. I can name lots of gorgeous, brilliantly conceived games that have tanked horribly in the sales department. But at the same time, I think that when industry professionals sneer at Myst and call it a bad game that ďgot lucky,Ē itís just a case of sour grapes. Granted, good timing and marketing were certainly involved in the gameís financial success, at least initially. But it seems that for Myst to have been that popular for so long, there must be something there which draws people in, keeps them interested and which has in turn led to the phenomenal success of the gameís sequel-of-sorts, Riven.

One of the major complaints leveled against Myst by those in the industry as well as by hard-core gamers is that Myst is so simple in its gameplay and technology. With 1997ís release of Riven some four years after the release of Myst, many of these hard-core gamers were aghast to discover that nary a technology improvement had been made. One couldnít even ďfree-lookĒ around the environments in Riven ó something QuickTime VR made somewhat possible ó let alone fully roam around the countryside, as one could in games like Quake. But, when one thinks about it, is it the games offering the greatest technological improvements that tend to stick with people the longest? Who plays Wolfenstein 3D anymore, even though it was one of the first games to offer fully scrolling, first-person, texture-mapped environments, and which was tremendously popular when it was released? As I think back on my favorite games ó Prince of Persia, Lemmings, Tetris, Millipede, Loom, Railroad Tycoon ó almost none of them offered any sort of ground breaking technology. Yet still they fill my gaming memories, completely influence the way I design games and I am still happy to go back and play them when I get the chance.

Artist and animator Steve Ogden, it seems, has thought a great deal about Mystís success. His answers are not quite so simple, however, and in the following article he delves into the tricky interaction between technology and art. Could it be that Myst has something which its imitators ó and in fact most computer games ó do not? Read on to find out.

ó Richard Rouse III



ÖBut is it Art?



Steve Ogden
Cyan, Inc.

As technology has become more tightly entwined with art, we have seen some great examples of a balanced marriage between the two. In this article, Iím going to discuss this tricky balance as it relates to the special effects in George Lucasí 1977 blockbuster movie Star Wars, and to the graphics in Cyanís 1993 blockbuster computer game Myst. I have intentionally picked older products for my two main examples so that Iíd be talking about classic members of each productís genre, and not some flash in the pan. Star Wars has had more than 20 years to prove its longevity. Myst is markedly younger, but in terms of the extremely young computer game market, itís very old, and continues to be extremely popular. But Iím also going to talk a little about the other, unfortunately more common pairing of art and technology, in which one discipline is overindulged at the expense of the other. As for this latter phenomenon, I believe it is embodied in the current stampede toward better and faster real-time 3D (RT3D) engines in computer games.

Iíve got nothing against RT3D, mind you. I think itís incredible. It is, in fact, a step on the road to the game of the future. Itís just that it appeals to that primitive part of your brain, the one that slumps you slack-jawed in front of the TV muttering, ďOoohhhÖpretty lightsÖĒ

Thereís nothing inherently wrong with appealing to that part of human nature. It gets the bell ringing. Weíre attracted to flash and dazzle, the way we may be physically attracted to a visually appealing mate, and only later discover what a fine person awaits beneath that outward beauty. Base level attraction can attract us to something greater, but there had better be something greater there, otherwise all youíve got is a beautiful airhead. So itís important not to appeal to the primitive brain at the expense of substance. And since youíve gotta have substance, why not go all the way and produce art?

In the past, I have worked with artists who were fond of pointing out that as commercial artists, what we do is not art. I would just point out that it does not necessarily have to be so. As a professional artist, you use the tools of art daily. You have to have color and contrast, light and shadow, style and substance, text and subtext. While youíre at it, why not try to create a little art? You are a hack only as much as you allow yourself to be.

Art makes the difference between Star Wars and its imitators, and likewise between Myst and its clones. I donít think the differences between these original works of art and their wannabe cousins can simply be ascribed to generational degradation, in the way that even a laser copier degrades copies. I think the would-be copycats missed the point of the originals. I believe that even though Star Wars and Myst both used technology that was state of the art when each was created, each was more than the sum of its technology. At the heart of each, there was an idea, and the technology served that idea. The copycats went wrong when they copied the technology, but did not force it to conform to any idea greater than a desire to replicate someone elseís financial success.

At the heart of any creative endeavor, if it is to have any sort of value beyond novelty, there must be an idea. The product, all other components of the creative endeavor, are simply attached to that idea. But those components all must support the idea, or the entire structure breaks down.




Figure 1
Figure 1: Star Wars had incredible special effects, but George Lucas was smart enough to know movies are not made through effects alone.

Star Wars

Letís look at Star Wars. Itís not my favorite movie, but I think that its stunning popularity over time allows its blend of art and technology to stand as a positive role model. I would suggest to anyone sublimating art to technology for short term reward that they rethink their stand in favor of the more balanced approach Star Wars used. The long term reward may be far greater, and the product certainly will be.

The heart of Star Wars, despite its more melodramatic overtones, is comprised of classic themes. Itís a fairly simple tale of knights and empire and rebellion. Once that idea was nailed down, the characters were defined in terms of classic literature, and the story pretty much could have written itself. So when it came to the special effects, all the guys at ILM had to do was just not screw it up. And they didnít. They created special effects the likes of which the world had never seen ó cameras in motion, explosions, aliens, lasers, spaceships zipping around like gnats. But even though Star Wars contained special effects and technology which for their day were incredible, they really only served to support George Lucasí idea. And yet, he wanted to do more. Scenes hit the cutting room floor because the technology he needed hadnít been invented yet. Thatís one reason why in 1997, he rereleased the Star Wars movies with a bunch of new special effects.

But what George Lucas didnít do, even in the 20th anniversary rerelease of that film, is as important as what he did. For instance, he did not release Star Wars in 3D, using some sort of special glasses to communicate a 3D experience to the audience. Itís not beyond the realm of possibility that the prospect came up. The technology certainly existed. Had the movie been released that way, Iím sure it would have been a sensation. It would have been cool to see all those ships buzzing around like that in 3D. But in reality, it would not have been the same film. Attention would have been diverted from the development process to the delivery medium, and the technology would likely not have been used as much in support of the idea as in an attempt to maximize the 3D effect. The film would likely have ended up as little more than a gimmick. I doubt that it would be the icon it is today. It is possible that the production costs of doing a film like that in 3D would have tipped the balance against it. But I prefer to subscribe to the notion that Lucas knew what he was doing, and understood the distinction between technology used in the development of a piece of art, and technology used in the delivery of a piece of art. In the development process, technology can be tossed around with abandon, with little effect on the audience. But in the delivery, technology must be exercised with caution.




Figure 2
Figure 2: Mystís world was its interface. If the player wanted to go somewhere or investigate an object, he simply had to click on that location or object.

Figure 3

Figure 4
Figure 3 and 4: Rivenís skies had clouds, and its water was animated. Myst had neither, but the fact that many people mistakenly remember that it did is testament to the gameís incredible immersive power despite its alleged technological weaknesses.

Figure 5
Figure 5: A detail from Riven. Clues can be subtle parts of small details in a prerendered world. Subtlety, however, is not the strong suit of RT3D. A shot like this in RT3D would likely be a smear of pixels.

Myst

Perhaps the computer game genre is in that half-dreaming state that the science fiction movie biz inhabited back in those days prior to Star Warsí release. Perhaps the computer game equivalent of the cinematic Star Wars is just around the bend. But when a new game comes out now, we donít hear so much about how the gameplay will be, or how fun it will be, but we get this intense sales pitch about the fantastic technology with which the game will be delivered. All this technology is brilliant, but what does it mean when a game comes out, just a simple little game with no fancy engine at all, and it blows the doors off of the entire biz? It becomes the King of Computer Game Sales and itís just aÖaÖslideshow?

Ah, we must be talking about Myst.

How could this low-tech game attract such an incredible audience with no RT3D? By its numbers alone (according to Richard Watson at Cyan, Inc., Myst has sold more than 5 million copies worldwide since 1993, and its successor Riven has sold more than 2 million copies in just over a year), it actually is the Star Wars of the computer game genre. See? It already happened, and without the great stride forward in technology.

For the record, Iím not one of those misguided souls that believes that movies and games are almost the same thing. I only draw the parallel between Star Wars and Myst because they both used art and technology in support of their ideas. The problem I see is that certain sectors of the computer game industry seem to be obsessed with technology to the exclusion of the gameplay just as Hollywood appears to be so enamored with glitzy special effects at the expense of story and/or character development.

Yes, Myst was a point-and-click slideshow format game. But look at whatís on the screen in Myst. The user was given everything he needed to play the game through that interface. There were no radar screens, or health bars or buttons or any other ďinterfaceĒ nonsense. In this game, you didnít need it. Cyanís artificial world was its own interface. If you wanted to go somewhere, you clicked there. If you were presented with a puzzle, you looked at how the world behaved in that universe, and solved the puzzle accordingly. In Myst, the puzzle solutions made sense in the same way its interface made sense, at least in the context of its universe. It was internally, logically, consistent and in this way its design supported its central idea.

This is where the horrible Myst clones of the early 90s went astray. For some reason, many of those infernal games wanted you to practically solve a crossword puzzle in order to do something as simple as open a door. In those cases, the gamesí design was actually getting in the way of their ideas. If the puzzles had been more subtly laid into the environment, they would have helped form an experience worth having, in a more or less believable environment, not just a collection of puzzles loosely strung together. Itís fine for a game to be so loosely glued, but wouldnít it be worth it just to go that little bit further?

The artwork in Myst is interesting to consider. In terms of supporting the idea, Mystís graphics are perfect. There is an odd air about them, as if from some sort of half-remembered dream. There is a strangeness, yet a familiarity about the environment in Myst. Thatís fitting, because the best puzzles in that game are at the same time strange and familiar. The core of Myst (the Díni, who write books that link to innumerable worlds) is surreal, and its graphics are, too. Itís a perfect match between subject and artwork. In fact, the less detailed graphics of Myst may have suited the Díni idea better than those of Mystís hyper-realistic successor Riven (as painful as it is for me as an artist to admit that).

Itís funny the little ways our brains trick us. In Riven, the water rippled, and the sky was dotted with fluffy white clouds. In Myst, there were no clouds in the sky, and the water didnít move. If you were to compare images from the two worlds side by side, the differences would be painfully obvious. Yet there are people who swear they remember clouds and moving water in Myst. So, is it that Riven gave us more than we needed, or is it that Myst gave us just enough? It seems the human mind will fill in some details.

Figure 6
Figure 6: Oddworld: Abeís Oddysee used sprites over prerendered backgrounds to deliver the gameís depth of detail.

Figure 7
Figure 7: Grim Fandango also used prerendered backgrounds to deliver its detailed world, but it also used RT3D elements over them to give the player the impression he had more interaction with the environment.

Figure 8
Figure 8: Wheel of Time is one of a new breed of game that uses RT3D in such a way that it is approaching the quality of prerendered graphics, at least for interior shots.



Steve Ogden is an Artist/Animator who until recently worked at Leaping Lizard Software, where he always talked about how much he loved Myst. You might not know it from reading the preceding column, but Ogden didnít start working for Cyan, Inc., makers of Myst and Riven, until after he wrote the article.

Steve Ogden
Cyan, Inc.
14617 N. Newport Highway
Mead, WA 99021
Website

Richard Rouse III is Lead Designer and President of Paranoid Productions and has published two games to date: Odyssey - The Legend of Nemesis and Damage Incorporated. He is currently serving as Lead Designer and AI Programmer at Leaping Lizard Software on the forthcoming Centipede 3D, to be published by Hasbro Interactive and due out ó you guessed it ó by Christmas. Your feedback to this column is encouraged at the address below.

Richard Rouse III
2124 I St. N.W. #306
Washington, D.C. 20037

Tel: +1-202-861-5513

Website

The copyright of articles and images printed remains with the author unless otherwise indicated.

Which brings me to my other point about Mystís artwork. In Mystís Channelwood Age, there were about 500,000 polygons. At this time, RT3D technology is just about to the point where it can comfortably handle that many polygons. Where RT3D fails currently is in the texture department. We effectively have a graphic size limit of 256 pixels square for a texture, where the images in Myst, because it used prerendered graphics, were in many cases much larger. (And in Riven, they were ridiculously large ó 60 MB or more!) So we can effectively set the user down in a RT3D game full of the same quality of geometry used in Myst, but we canít texture it as extensively, which breaks down the gameís environment into blurry pixels at close range.

Itís a tradeoff. Some people want the extra freedom of movement, and are willing to accept the image degradation in return. In a graphic adventure, however, we canít accept that degradation. Itís not that the graphic adventure audience is a bunch of morons who can only immerse themselves in a painstakingly detailed environment, although I do believe the more detailed environment, even without the freedom of movement, is easier to believe in, and certainly nicer to look at.

But the real reason for clarity in the images of a graphic adventure game is that clues can be lurking at any turn, and puzzles can be hiding in the details of the scene. If the environment is to be a believable one in support of the idea, the way these clues and puzzles need to be delivered must not violate it. Therefore if something is to be hidden in the details of a scene, there must be an appropriate level of detail available in the scene in which to hide it. If not, clues and puzzles wind up as metaphor or iconography, neither of which fades very gracefully into the surroundings. Once a piece of entertainment violates its internal logic, it breaks the audienceís suspension of disbelief. Internal contradictions force a creation to collapse under its own weight.

Riven was created with the same attention to internal logic that helped make Myst such a success. But when Riven was released, using that same old slideshow interface, gamers and critics screamed aloud about its lack of technological innovation. Where was the flash? Where was the next great stride forward?

Use What Works Best

Well, to state the obvious, youíve got to use what interface works best for the nature of your beast. The comments I heard upon the release of Riven have indicated to me a lack of reason prevalent in the game industry currently. Perhaps it needs to be said: RT3D wonít work for everything, and doesnít need to, any more than Iíd recommend the slideshow interface for every game. All delivery systems should be viewed as valid, but of varying degrees of appropriateness. The finished product must be viewed in terms of how well it supports its central idea, and delivery can be a huge part of that.

And the slideshow interface is not necessarily dead, folks. It lurks there, just under the surface, in disguise. Look at Oddworld Inhabitantsí Oddworld: Abeís Oddysee. What a strange hybrid that game is. Itís part Mario Brothers-style side-scroller, part screen troller, part twitch game. And yet, the environments in the game are served up slideshow-style. The user maneuvers animated sprites of the main character, Abe, through this environment, jumping, rolling, running, sneaking, searching, and yes, flatulating on cue. But even so, when Abe leaves the scene, the ďcameraĒ does not track him. He vanishes off screen. Then the next background loads, ready to bestow the next page of the adventure upon the user, at which point, Abe reappears at the opposite side of the screen, ready for action.

Someone made the comment to me the other day that he was frustrated that Abeís Oddysee did not scroll, but I think that Oddworld Inhabitants actually used this function to their benefit. There are many cases where the slideshow delivery actually helps the gameplay. For instance, Abe can lose guards who are chasing him just by running off screen in the early, tutorial-style levels. The limited use of technology doesnít make for an inferior game, just a different one. And as far as the quality of the art involved in Abeís Oddysee, I have heard rumors that Abeís Exoddus, the follow-up to Abeís Oddysee, might just snag an Oscar.

Lucas Artsí game Grim Fandango is another interesting hybrid, which uses prerendered graphic backgrounds served up slideshow-style, and renders real-time characters over it ala Westwood Studiosí game Blade Runner or Capcomís game Resident Evil, among others. Even though some of its puzzles seem a bit arbitrary, I think that Grim Fandango is an example of art and RT3D technology being used properly in a game. The real-time elements present the user with enhanced interaction, while the prerendered backgrounds present the environment with the necessary detail. Itís a nice mix. So, I just canít see the harm of prerendered graphics in a slideshow interface in the context of a game that needs the user to explore the environment in detail. But thatís why I donít have a problem with Myst and Riven, and why a lot of my peers do.

So, it bears repeating: Iím not against RT3D, really Iím not. Itís just that, when I set out as an artist to convince the user of the reality of my artificial world, I want my work to get to the screen as realistically as possible. I want to be able to present the elements of a convincing world: light and shadow, reflection, textures that do not obviously repeat and surfaces that do not break down into pixels upon close inspection. I do not want polygons to be the obvious building blocks of this world.

Right now, unfortunately, that list of items that help convey a convincing picture of a world reads like a list of things that can be achieved best through the dated, unfashionable slideshow interface. Itís also a list of things that currently are very difficult, and thus rare, in RT3D work. As I said to some people, itís an acceptable tradeoff. But not to me. I know we can do better, and that we will, but unfortunately we must first pass through this rather unsightly and uncomfortable stage.

RT3D is getting better. Look at Legend Entertainmentís forthcoming action-adventure game The Wheel of Time compared to the RT3D games from a couple of years ago. There are cathedral shots in The Wheel of Time that are remarkably realistic. MaybeÖjust maybeÖRT3D is almost where it needs to be in order for the end of the slideshow interface to be a realistic proposition. Maybe.

Time is such a funny thing. In 1998, I can sit here and say that RT3D isnít quite up to snuff, but in 2000, who knows? After years of defending and explaining the slideshow interface, I find myself for the first time seeing that soon the two different approaches will be just about even in what they can deliver. Rest assured at that point, I will want to be in the thick of it. But even then, in the midst of glorious RT3D with full freedom of motion, I hope that we all will continue to use art to serve the idea and avoid being seduced by technology.