DETAILS


COLUMNS


CONTRIBUTIONS


ARCHIVE



GAMING & GRAPHICS

Vol.33 No.4 November 1999
ACM SIGGRAPH


Advancing Game Graphics: A War of Escalation



Richard Rouse III
Paranoid Productions

November 99 Columns
About the Cover Visfiles

Gaming & Graphics
Previous Gaming & Graphics Previous Gaming & Graphics

Itís confession time: Iíve always feared the onward march of computer game technology. Not feared it enough to ignore it, pretend it doesnít exist or not try to learn everything I could about it. But as a game-design-centric game developer, I have always feared that as each and every game genre has to suddenly be in 3D (regardless of whether the game in question benefits from it), and as every game must include every bit of visual razzle dazzle that every other game has, inevitably more time is spent worrying about the technology than the core gameplay. All of a sudden, developing the game Iím interested in creating is now complicated and encumbered with a lot of special effects it doesnít need. If the game is to have any chance of finding a publisher, it will need to promise visual effects technology that is superior to all the other games available.

I donít know how many times Iíve seen game reviewers make the following comment: "Well, Game X would have been entertaining enough three years ago, but now it just looks hopelessly dated. Itís just not fun anymore." In my dictionary, fun is a timeless thing. If a game is fun five years ago, it will still be fun today, as it will be fun five years from now. More likely, games that "stop being fun" were never fun in the first place, and simply wowed then-current gamers with their special effects. Once the effects are dated, theyíre no longer so special, and the game stops entertaining the players, since it was little more than a hat-full of visual trickery in the first place. But good gameplay doesnít date the same way 3D game engines do; people who liked Tetris back in the late Ď80s will find it just as fun today.

When will it ever end? I believe that, at some point, as games start to have the visual qualities of modern movies, the rush to improve visual effects will start to ebb, and the design and content of the games will become a more central concern for game developers. But until then, the race for visually groundbreaking work will continue. Steve Ogden has been kind enough to write this issueís column, where he poses a troubling question: how good is good enough?

- Richard Rouse III

How Good is Good Enough?

Steve Ogden
Cyan, Inc.

In June of 1922 in New Yorkís Hotel McAlpine, the members of the American Society of Magicians were treated to a secret screening of perhaps the worldís first dinosaur movie. The film was brought to them by Sir Arthur Conan Doyle. Doyle informed the society that the film contained footage he had recorded on his recent dinosaur expedition to a far off mysterious land.

The magicians, not an easy audience to fool mind you, were reportedly fairly well convinced by the sight that greeted them on the screen. For several minutes, as Doyleís film unfolded, dinosaurs cavorted among palm trees and fought each other amid dense jungle vegetation. When the film was over, the men spoke in excited tones about the fascinating display they had just witnessed on-screen.

And now, the punch line. Doyle informed the magicians that what they had just witnessed had been a trick of sorts: trick photography. The film had actually contained early special effects tests for Doyleís movie version of The Lost World, based on his book of the same name. A young special effects genius named Willis OíBrien had built dinosaur models and a few little tabletop sets, and made the dinosaurs move via good old-fashioned stop motion animation, which at the time was ground-breaking technology.

That viewing served as a demonstration to Doyle that OíBrienís technique would do what Doyle needed it to do: convince the audience that the dinosaurs they were seeing were real. A special effects star was born. The Lost World was released three years later, and OíBrien would go on to perfect his craft on such films as King Kong and Mighty Joe Young, among others.




Figure 1a: The Lost World (1925) and Figure 1b: The Lost World (1997): Itís amazing anyone ever thought that the dinosaurs in the original The Lost World (on the left) were real. But by the time the most recent The Lost World (on the right) came out, the viewing public was far too jaded to think the dinosaurs might be real.

Now, look at Figure 1a. Itís a still of one of the effects scenes from Doyleís The Lost World.†A rather menacing looking dinosaur (a tyrannosaurus, Iím guessing) is looking to make a snack of the young damsel in distress. In 1999, it is hard to believe anyone could be convinced by a scene like that. And yet, the American Society of Magicians believed.

Flash forward to 1997 and Stephen Spielbergís The Lost World (Iím given to understand that Crichton used the title to pay homage to Doyleís earlier work). This time, there are very convincing computer-generated dinosaurs complete with motion blur, eyelids that actually flex and incredibly lifelike skin (see Figure 1b). The dinosaurs cast shadows on the landscape, each other and the human actors in the scenes with them. They splash in the rain late in the film. Intellectually, as a modern audience, we know those dinosaurs canít be real, but there really isnít much evidence on the screen to support that fact. They are integrated almost flawlessly into the film. Still, I donít think anyone in the film-going public of the late Ď90s would have believed that the dinosaurs were real.

Cutting-Edge Techniques

The different production values of the two versions of The Lost World notwithstanding, they are nevertheless united in the fact that each film used filmmaking techniques that were out on the cutting edge. The results were drastically different, but think about their incredibly different audiences. In 1925, who had seen footage of moving dinosaurs? Nobody. But by 1997, who hadnít? The difference here is in the visual sophistication of the audience.

Iím aware that I use movies as examples fairly often, perhaps to excess, in discussing computer game graphics. But I think movies provide a good frame of reference because games and movies use some of the same tools of art, and both audiences are increasingly visually sophisticated. I particularly find The Lost World case study to be useful in this instance because of the huge chunk of time that went by between versions. Their differences serve to highlight the mind-blowing advances in SFX for movies that have occurred over the past 75 years. I think it provides a parallel example with which to compare the tremendous leaps in technology and visual arts achievement that have taken place in the computer games industry over the past 10 to 20 years.

Technology and art have been chasing each other in a war of escalation in the computer game business for quite a while now. Products improve, more is expected, products improve again in the next revision, still more is expected, and so on. But how good do the graphics in a game have to be? In some ways the answer is as varied as the types of games available. I will explore this question in depth in a moment. But first, letís think about the different ways in which we can represent a gaming environment. Currently, we have an astonishing arsenal at our disposal. We can present characters and environments in stunningly rich detail, right down to the stitches on a characterís clothing, or pebbles on the ground. We can imbue the game environment with believability, and narrative-appropriate frame of reference if need be. We can use lighting, stereo music and 3D sound effects to provide the appropriate mood. Our characters can move, fight and emote. (Granted most of us are still working that last one...) We can do anything from lavishly hand drawn art to pre-rendered orthographic views with sprites to slide shows to virtual reality or any combination thereof. We can do anything, that is, but go back to the way games looked before the latest improvement.

It All Began with Icons

The farther back in time we go, the farther we move away from the glossy high tech games we now enjoy. Resolution decreases. And as we reduce the resolution of the characters and environments, we move closer to the point that eventually they become mere icons. And it is with icons that computer games really began.


Figure 2: Pong: The ultimate in icon gaming, Pongís minimalist representation of a ping-pong game was amazingly popular.




Figure 3: Centipede (1980) and Centipede (1998): The graphics in the new version of Centipede (right) are immensely more sophisticated than in the original. But do they make the game any better?
Think about Pong.

Pong, for you young Ďuns out there, was a ping pong simulation. It consisted of a black screen with short, thick white lines on each side (the paddles) which could be manipulated vertically by the player or players, in an attempt to control a bouncing 10-pixel square (the ball) that beeped when struck (see Figure 2). Thatís it. But it was astoundingly popular back in the early Ď70s. It was not a ping-pong game; it was a series of icons that stood in for a ping-pong game. The conversion took place in the playerís head.

Icons went a long way in those days. I can remember long hours spent in front of my Atari home system playing Demon Attack against inscrutable icons. There was nothing convincing about that gaming environment, and yet I know my heart pounded as the action intensified. I understood the threat that those icons represented to my high score, and within that context, I was involved.

It is in that context, the age of the icon, that I wish you to view my next now and then case study. Consider the game Centipede that was released as an arcade game by Atari in the early Ď80s, and revamped into a real-time 3D (RT3D) shooter for the PC and Sony PlayStation by Hasbro Interactive in 1998 (see Figure 3). I have to say that both are fun, just fun in different ways.

The gameplay in the Atari version is redundant but frenetic such that its pacing presents a continual challenge. Limited by the technology available at the time, the game consists entirely of icons moving on a black-background "environment," but itís still a very enjoyable game. It really is a nice piece of game design - its genius is in its simplicity. Think how difficult it would be to envision a game like the original Centipede without having seen it first. My hat is off to designer Ed Logg.

The Hasbro version (on which I did a little work while at Leaping Lizard Software) is not as relentless, but it makes up for that by varying the complexity and style of the gameplay across several diverse locales. Courtesy of its RT3D nature, it operates in a more believable environment, which is to say, I suppose, that it had an environment at all. The introduction of exploration and puzzle solving pushes the Hasbro version in a whole different direction than that of its predecessor. But since the icon-based version worked so well, I wonder if a push in that direction was necessary?

Raising the Bar

Commercially it was, because if one were to release a game that looked like the original Centipede today, aside from its limited nostalgic appeal, itíd be laughed off the shelves. Yes, there are classic game emulators out there and emulated classics have been sold in collections to modest retail success. But essentially, the reign of icon-based arcade games is over, and has been for quite a while. Every time we improve our products, the results become more impressive, and the bar raises a little bit. But the important thing to note is that when the bar raises, it raises permanently, and our expectations are higher from that point on. The question of whether the game needed to be remade (as opposed to simply rereleased as an emulator) is another matter entirely, but once that decision was made, the game absolutely had to be updated with the latest technology.

Now, I donít want to veer too far off topic here, but I think it warrants a paragraph or two: I believe that struggling to overcome the limitations of the technology may actually have made the Atari version of Centipede stronger than it might otherwise have been. The truth is, I find that many artists do their best work when some limitation is placed on them. Oh, sure, we whine and complain about being restricted, but thereís something about confining our creative energy to a degree - it seems to concentrate that energy in an attempt to transcend the limitation. By contrast, creativity without limits is, I think, more power than most people can handle. The work produced under such conditions tends to be unfocused.

In the modern Centipede, many more options were open to the Hasbro/Leaping Lizard team. We were not as limited by technology, so we didnít have to work so hard to overcome the limitations of our medium. We still had limitations, certainly, but compared to what was available in 1980 for the original game, we had the universe at our command. And in fact, we put a lot of the universe in that game. There are things in that game that couldnít have been done with icons; complex, multi-tiered mission objectives that involve changing the landscape over time. That itís fun to play is a testament to the hard work and long hours designers Richard Rouse and Mark Bullock put in. That it hasnít become more popular may be a sign that the original had some appeal we werenít able to recreate even though our tools were ridiculously more powerful.

Using Technology to Improve Gameplay

You can see from the Centipede case study how progress can drive us to remake games. I think, however, technological improvement is the weakest reason to update a game. But I can see a good reason to update a game. Think about the old text-based adventure games, among which Zork is probably the most famous. Here was a faction of computer games that had no art or even icon whatsoever. They simply described in detailed text what the player needed to see in order to play. This moved the experience into the playerís head, where I think true involvement begins and ends. Icon-based gameplay can be involving, but description goes so much deeper.

However, text wonít work for everything. Consider the situation where the player needs to see a cave out of the corner of his eye. The text would read, "Out of the corner of your eye, you see an entrance to a cave," or something along those lines. The problem is, the phrase "out of the corner of your eye" draws attention to itself. Rather than being a seemingly insignificant detail in the gameís environment, it stands out as a clue or a destination, which breaks the spell of the game to a degree.


Figure 4: Zork: Nemesis: The sophisticated computer graphics in†Zork:†Nemesis†allow for a what-you-see-is-what-you-see style of gameplay, in sharp contrast to the text descriptions of locations found in the original Zork games.


Figure 5:†Thief: The Dark Project:†The sophisticated 3D engine used in Thief (the shadows in particular) was an essential component of†Thiefís gameplay.

The fact is, neither text nor icon would be adequate to present this situation. The best, most involving way to present a situation where the player sees something out of the corner of his eye, is to present the scene to the player, and actually let him see the item out of the corner of his eye. Consider Zork: Nemesis, which is a full-on Myst-style graphics adventure set in the original Zork universe (see Figure 4). I can see how Zork: Nemesis might actually work better in some instances than the original text-based Zork games. This situation is a good reason for upgrading a game graphically.

But for whatever reasons, as time has worn on, we have found other ways to present ever more complex puzzles and situations, ways that have moved farther away from either icons or text-based description. Look at Looking Glass Studiosí Thief: The Dark Project (see Figure 5). Itís an incredible RT3D game with amazingly detailed textures and moody lighting. You need to hide your character in the shadows of this game to avoid discovery. Quite seriously, without the technological advance of lighting effects that allow the character to hide in the shadows, this game would not exist, at least not in its current state. Itís eerie how involving it is, and Thief presents a very complex set of rules in a way that the player can easily understand. As far as presentation of the idea and environment, I think Thief has found the best way to go for the type of game it is, at least until the next jump in technology.

How Good Do Game Graphics Need to Be?

The technology is moving forward. Soon weíll be able to put on-screen just about whatever we want to, and the difference between what we can do today versus just a few short years from now will be the same as the difference between the two versions of†The Lost World. And in that glorious future, what will be our aim? Do we want to recreate reality, or do we want to improve on it, put our own spin on it? And, more to the point, exactly how good do game graphics need to be? We are well past the mid-point on the continuum between icons and thoroughly convincing visual representation. So how good is good enough?

Iíd like to answer that question in three ways: as an artist, as a game developer and as a gamer.

As an artist, I think that the art should be as great as it can be; not just great within context, as in "Hey, this looks great for a game," but great compared to any other medium. If Iím going to present virtual reality, I want it to be as convincing as possible. I wonít be satisfied until the level of visual representation I can provide through RT3D, for instance, looks as good as a big budget movie. Itís a high bar, but arenít we heading there anyway?

As a developer, Iím quick to point out that the artist in me has no respect for reality. Right now, there are technical limitations. There are budgetary constraints. There are concerns about the consumer target machine. There is a time limit; eventually a product must ship. Otherwise, itís not a product but a hobby, an expensive habit. So as a developer, I want a game that looks good, runs well and ships on time. If it doesnít completely recreate reality, weíll live with it. The next game will be better.

As a gamer, I want the artist and developer to hurry up and decide what they are going to do and get on with it. I can see a set-point between how long it takes to do the art, and how long it will take to get that product to me, the anxiously awaiting audience. I want the game as soon as possible, but itíd better be worth the money. It doesnít have to be the most incredible art that will ever be made, just better than anything thatís out now. And if the gameís fun enough to play, I might be willing to overlook some superficial visual shortcomings.

Subjective, no? Obviously, thereís a tradeoff - the more time we spend on the art, the fewer games we are able to put out. So, in the end, itís a question of quality over quantity. I am willing to live with fewer games if it means they will be of a higher caliber. Iíd rather have that than the current situation where the market is flooded with mediocrity.

Whatís Ahead?

The artís going to continue to get better. Audiences demand it, and thereís no stopping it now. Spielbergís The Lost World is visually superior to Doyleís version, but even the dinosaurs in Spielbergís own Jurassic Park pale in comparison to those in his The Lost World, though not much time had passed. It seems that every time the bar raises, what came before seems sad and inadequate. The sad thing about progress? Artistic innovation is so often not backwardly compatible. So, after the next wave of games rolls through this year, and next year, and the next, who will be interested in playing last yearís crop of games anymore? Iím sure you know someone who refuses to watch black and white movies, on principle alone. In a few years, wonít these games from the Ď90s look just as dated and unfashionable as a black and white movie looks to some people now?

But Iím not worried about the graphics any more. There are lots of great programs out there that allow us to create incredible visuals. And the bar has been raised to an appropriate level - regardless of the target youíre aiming at, everyone knows that graphics need to at least come up to a certain level. What I have begun to worry about recently is content. So much attention has been lavished on artistic and technical achievement that I think it is distracting us from the process of actually making better games. Only content can compel us to play a game that is past its prime, for if a game is only the sum of its technological advances, it has nothing to offer when its technology is trumped by the inevitable newcomer.

Trust me, the art will be good enough. Now letís work on the rest of the game.



Steve "Og" Ogden is an artist at Cyan, Inc., where he is working on a new top secret project. Prior to his work at Cyan, he was employed by Leaping Lizard Software, where he worked on Game Magic (a casino gambling cabinet released by Bally Gaming) and Centipede (the 3D remake released by Hasbro Interactive). His comments on the rising bar of art aside, he is proud to be indulging the ever-increasing visual sophistication of Cyanís audiences.

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 recently functioned as Lead Designer and AI Programmer at Leaping Lizard Software on Centipede 3D, published by Hasbro Interactive. Having at long last completed work on the PlayStation version of that game, Rouse has opted for a change of view and has moved on to a new project at Surreal Software.

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

Website


Richard Rouse III
716 11th Ave E.
Seattle, WA 98102


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