Computer Graphics Volume 32, Number 2, May 1998

Do Computer Games Need to be 3D?
by Richard Rouse III

If you ask a random computer game professional what the most importantinnovation in computer games has been in the last couple of years, morelikely than not they will tell you that it has been the industry's movetoward 3D graphics and environments.  Take, for instance, the followingquote from the January, 1998 issue of Next Generation magazine:
"If there is such a thing as a videogame Holy Grail, it isnot greater interactivity or improved design, despite what certain game-designgurus would have gamers believe.  Instead, for better or worse, theperennial object of desire for developers and gamers is the creation ofrealistic 3D worlds.  The crucial extra dimension is the key to awhole new breed of game."
Almost everyone involved in making games seems to be thinking of how totake old game designs and transform them into hot new 3D properties. Some of these efforts have resulted in exciting new game "breeds" thatengage the player in an experience that is all but impossible without 3Dtechnology.  Witness Mario's jump to the 3D world in Shigeru Miyamoto'sSuper Mario 64.  But then again, witness the dismal failure of gamessuch as Frogger 3D, a title which most everyone will agree is less entertainingthan its early Eighties predecessor.  So this brings us to the questionat hand:  are all games better in 3D?

Before proceeding, clarification of just what we are talking about isin order.  When I refer to a 3D game, I mean one in which the playeris able to move "anywhere" in a given space.  The computer then dynamicallyrecalculates what the player's view of the world should be, and redrawsit accordingly.  A game such as Super Mario 64 or Quake is a 3D game. A game such as Rand and Robyn Miller's Myst is not a 3D game:  thoughits graphics are rendered using 3D models, they are all static images and,as such, the player cannot move freely through the world.  Games ofthis ilk are frequently referred to as being "on rails" due to the waythey restrict the player's movement to a specific path.  The staticimages in Myst were created using a 3D modeling tool, but they're staticimages nonetheless, and hence the game cannot be considered truly "3D."

As I mentioned before, 3D technology allows computer game designersto create gaming experiences for the player which could not exist withoutsaid technology.  Take, for instance John Carmack and John Romero'sQuake.  Design purists might argue that Quake's gameplay is littlemore than, say, Ed Logg's Gauntlet, released in 1985.  But the 3Dtechnology in Quake allows for a level of immersion and excitement thatwas not found in 2D games.  Spinning around a corner at high speed,not knowing what threat one might find there provides a visceral thrillthat is altogether missing in Gauntlet (though that beloved game certainlyhas its own appeal).  In Quake's case, the 3D technology created anentirely new game.  One could make similar arguments for the originalityof games such as Tomb Raider and Super Mario 64 which take 2D games (Princeof Persia and Super Mario Bros. respectively), add 3D technology, and createa distinct and unique experience for the player.

I certainly enjoy a lot of 3D games, and from my praising of Quake,Tomb Raider, and Super Mario 64 one might conclude that I am thoroughlyconvinced we should do away with 2D games right quick.  But what ofa game like Jonathan Ackley and Larry Ahern's The Curse of Monkey Island,or, for that matter, either of its predecessors, The Secret of Monkey Islandand Monkey Island 2:  LeChuck's Revenge, both designed by Ron Gilbert. These are games where the hand-drawn, cartoony nature of the game's characters- including the player's surrogate, Guybrush Threepwood - is as inherentto the game's charm and appeal as the 3D engine is to Quake's.  Whatif you replaced these characters with polygon based creatures - like thosefound in Quake - and rendered the world with assorted texture-mapped surfaces. Certainly the game's flavor - reminiscent of classic Warner Bros. cartoons- would be destroyed and a much less appealing game would remain.

Or how about a game such as Sid Meier and Bruce Shelly's Civilization,where the appeal of the gameplay is based on careful strategic planningand viewing an entire world at once.  How could 3D graphics add anythingto the game's appeal beyond fancy window-dressing?  For example, BrianReynold's Civilization II added the ability to "view city," whereby a playercan take a look at his favorite city from a close-up, aerial view. This provides a nice looking glimpse of what a given city's layout lookslike, with its coliseums, harbors, Wonders of the World, and so forth. Though it is visually pleasing, does anyone really use this feature morethan a few times?  Certainly not, for it provides no data that isat all helpful in playing the game, and serves only as a pointless distractionfrom the player's main goal.  So too would a 3D engine distract fromwhat the game's about, without adding anything of value to the playingexperience.

To recap, my argument against 3D graphics as a technology all computergames should embrace has two fronts; the first is an aesthetic one, maintainingthat rendered 3D graphics will never be able to create the look some designersmay want for the games, such as the cartoon based look of the Monkey Islandseries; and the second argues that a 3D engine will not necessarily enhancethe play of every game it is applied to, sometimes making the game worseby adding distracting fluff on top of an otherwise solid game, such asa hypothetical Civilization 3D.

Recently I tried pitching a game design to a number of publishers. Though I would rather not go into the details of the design itself, let'ssay that it combined the gameplay elements of Will Wright's Sim City crossedwith Meier and Shelly's Railroad Tycoon in a unique setting.  It wasdefinitely a game where the gameplay was in no way dependent on a 3D engine,nor would such technology have improved the player's experience in anymeaningful way.  As such, in the design treatment I described an isometricAge of Empires type engine as the one the game would use.  I showedthe document to one industry-savvy friend, and he pointed out that I mightwant to discuss what 3D accelerator boards the game would be supporting. When I countered that the game was not intended to be 3D at all, he said"Well, you should see what you can do about making it 3D."  Later,a potential publisher I showed the document to said that he liked the design,but wondered if the game's engine couldn't be 3D, so that the game wouldbe able to compete technologically with other games that would be availablewhen my game was released.  When asked how that would improve gameplay,he was hard-pressed to think of a way, but insisted a 3D engine would benecessary nonetheless.

And these two people were, from their frame of reference as businessmen,correct.  For it is hard to communicate "compelling storyline" or"highly refined gameplay" on the back of a box, while a hot 3D engine communicateseasily via glitzy screen-shots.  Having your game be 3D is a "moneyin the bank" sort of guarantee that makes investors feel more comfortable,just as a film's financial success seems more likely if it has a big-namestar attached to it, or a new book's potential profitability greater ifJohn Grisham were to praise it on the dust jacket.  As businessmen,publishers need to show profit for the money they invest in getting gamesdeveloped, and if having them all be "hot 3D" games seems to make thatmore likely, then some would argue it is their responsibility to make sureall their games are 3D, regardless of the needs of the gameplay.

But I am not a businessman; I am more interested in creating good gameswhich appeal to not only established, hard-core game fans but also to awide spectrum of the general public.  I am one of those "game designgurus" (or perhaps just a guru-in-training) that the Next Generation quoteabove refers to so derisively.  Without doubt the most popular gameof the last several years is the aforementioned Myst, which, as I explained,is nothing like a 3D game.  Myst has a more elusive quality to it,which has escaped the designers of its countless imitators, none of whichhas come close to its level of financial success.  Though the gamewas universally praised by critics on its release, it is almost as universallydespised by the same critics now, who contend that it is some sort of massmarketing fluke.  No, I think Myst has something to it which "real"people (non-gamers, if you will) find appealing, and the game community'sderision of it is just so many sour grapes.  As these non-gamers playMyst, do they think "Wow, if only this was true 3D!"  Hardly. The game survives on the strength of its beautiful graphics, strong storyline,and compelling gameplay; a fully 3D engine would do nothing to improvethe storyline or gameplay, and would almost certainly make the graphicsuglier.

Right now, some readers may be saying "What do you mean, make the graphicsuglier?  It'd be real 3D, dag-nabbit."  To the hard-core gamer- call him what you will, aficionado, hobbyist, or fanboy - the game mightbe more impressive, as they would probably agree that Quake's graphicsare much better than Myst's.  But to the lay-person, the graphicsin Quake look, well, horrible.  To the non-gamer Myst's are obviouslysuperior.  Take a minute and look at a Quake or even Quake II monster: look at those angular polygons sticking out all over the place, with atexture mapped octagon posing for its head.  I hate to burst anyone'sbubble, but that is not what the general public considers visually appealing. Myst's graphics were 3D rendered as well, true, but notice that the fewtimes characters appear they are actually filmed humans.  The Millerbrothers were smart enough to know that 3D models cannot create a goodlooking human being, so they all but left living creatures out of theirgame.

Take the success of the Disney distributed, Pixar created movie ToyStory, which used 3D rendered animation and was a massive financial and,in my opinion, artistic success.  "The public certainly liked those3D characters, didn't they!" the 3D graphics pundit will say.  Butagain, Pixar was very smart.  They must have asked themselves "Whatcharacters can we create using 3D models that we can make look like theydo in real life?"  The answer, of course, was toys, which are notreal in the first-place, so one can create a very compelling and realisticMr. Potato-Head using rendered 3D graphics techniques.  The humansin the film, instead of trying to look realistic, appear as very stylizedcreations, and many would agree that they are the least-appealing aspectof the movie; Pixar was smart enough to keep their screen-time down toa minimum.

I have a friend who is a pretty typical non-gamer; she has no bias againstgames, and will enjoy some titles such as Civilization or Pac-Man. I recently showed her some screen-shots from the recently released BladeRunner game, in particular those from the 3D-rendered cut-scenes, featuring3D models of humans.  "What do you think of these graphics?" I asked her.  "They look ugly" she replied, not a doubt in her mind. Though critical response to Blade Runner's graphics has been near-universallypositive, and everyone marvels at how nice the 3D rendered cut-scenes are,to the non-gamer they still look sub-standard and unattractive.  Andthese are the pre-rendered cut-scenes I'm talking about here, not the actualgame-play graphics.  One day, when we have the processing power atour disposal to create 3D graphics as nice as Blade Runner's cut-sceneson the fly, in real-time, they'll still look ugly to the vast majorityof the population.  So if we want to put characters in our games,and we want our games to appeal to as many people as possible, is it reallythe best option to make these people out of polygons?

Of course, clever designers will be able to turn a disadvantage intoan advantage, and will use the blocky look of 3D rendered characters tocreate a specifically stylized appearance.  For instance, Tim Schafer,designer of the forthcoming LucasArts adventure game Grim Fandango, saidthe following in the November 1997 issue of Computer Gaming World:

"Mexican folk art has primitive depictions of skeletons, andit would be great to see them come to life.  I thought it was thekind of thing that would look really good in 3D, unlike a realistic humanfigure with the low poly count that you are sometimes restricted to. Realistic art usually just doesn't look that good.  Instead of fightingthe tech limitations of 3D, you have to embrace [them] and turn them intoa style."
By setting the game in "the land of the dead" and evoking a Mexican folkart look, Schafer made a similar decision to Pixar's, exploiting the limitationsof the technology to his advantage.  But not every designer will wantto adopt a stylized folk art look for their game.  So must every designerhave to use 3D models of humans for their characters, or make their gamescharacter-less as the Millers did?

To make another comparison to movies, think of a 3D engine as similarto the use of special effects in films.  Consider it this way: a game's 3D engine acts as a central and constant special effect throughoutthe player's gaming experience.  Special effects in movies have comea long way in the last twenty years, and a film such as Starship Trooperswould certainly have been impossible without them.  But does thisexciting new computer imaging technology mean that absolutely every moviehas to use it?  Will other movies that do not have such special effectsbe criticized for their lack of computer generated insects?  Hardly. It seems like as many people, if not more, are interested in seeing a no-techmovie like As Good As It Gets as want to see Starship Troopers.  Itis patently absurd that any critic would say "As Good As It Gets is animpeccably written and deftly acted film, but its special effects justcannot compare with those found in Starship Troopers."  Why is itany different for computer games?  Why does every computer game needto have the 3D engine special effect?

I asked a game-industry friend of mine recently if he had gotten a chanceto play Final Fantasy VII, a game whose sales numbers indicate it has significantmainstream appeal, certainly in Japan.  My friend replied, no, hewas not interested, since Final Fantasy VII was not true 3D.  Thisway of thinking is absurd to me.  Consider this:  the archetypalforms of computer games are the games that people have been playing forcenturies.  Take, for instance, Chess.  You can buy a Chess setat your local drug store for three dollars, and it comes with cheap littleplastic pieces and a crummy cardboard playing board.  Or you can goto a Chess specialty shop in beautiful New York City, buy lovely marbleChess pieces and a suitably fine leather board to accompany them. In the latter case you will have a significantly more visually pleasingChess set than in the former.  But when you are over at your nephew'shouse and he wants to play Chess on his drug-store-purchased set, shouldyou refuse?  "Sorry lad, I only play with marble pieces."  Yournephew may play as well as Bobby Fischer, but why bother?  His piecesare made of plastic, and they're ugly.

Of course it is foolhardy to make a one-to-one comparison between computergames and traditional games, or computer games and films.  But I thinkthe above analogies do hold some truth, and I think the industry's growingmentality that if a game is not 3D it will not appeal to consumers is afoolish one.  I am personally glad that every movie coming out ofHollywood is not loaded down special effects, though I can still be delightedby Starship Troopers.  Gamers and game developers should not limitthemselves to high-tech thrill-rides, when the subtleties of a Myst, MonkeyIsland, or Civilization can be as much, if nor more, rewarding.