This is an article I wrote many years ago, when GTA IV had just come out and I played the game almost to completion. [1] I never got around to publishing it, though. Now, with GTA V out for the PlayStation and X-Box, I dug up this post and figured I’d clean it up and post it anyway.

From 5 Reasons GTA IV Is The Worst Great Game Ever Made (Cracked):

“…there’s simply no way to accurately tell everybody that this is the most jaw dropping game you’ve ever played, and at the same time you fucking hate it so much it’s like a knife in your eye.”

It’s mind-boggling to think what the testers for this game went through. But which one thought I would ever be interested in seeing the front of my helicopter during a mission? Which one thought it was OK for an enemy to be able to shoot me through stairs that are all but impenetrable for me? Who here thinks it should ever be possible to drop a grenade on your foot? Why do you make me re-spawn a city block away from the helicopter, forcing me to sprint to it for 30 seconds? Which part of the fun is that?

First, the good stuff

On the other hand, there are also these cool little side missions, like stopping on a little, green person icon and then chatting with a person who gives you a hundred bucks because he’s coked off his ass.

At 95% completion, I’ve spent over 72 hours in the game over two months of play. It’s an embarrassing amount of time, I know. But it is a fun game and there is a tremendous amount to do. But even with that huge time budget, I didn’t waste a lot of time; I had fun, but I was more-or-less constantly pursuing game goals rather than enjoying the so-called sandbox world. [2] Clearly there’s no danger of there not being enough game to play; players whining that the game is finished too easily or quickly have no idea what they’re talking about.

Spotty realism

A lot of video-game culture has long been based on a notion that fun includes suffering. That is, you have to earn your fun; you can’t just have fun all the time … where’s the fun in that? So, sometimes you have to do things that aren’t really fun in the interest of the story or continuity or whatever. Though there’s usually a reason why you have to do more tedious things—for the sake of realism, usually—there is no rhyme or reason as to when you have to stick to realism and when you don’t.

For example, you can carry around what must be a half-ton of military hardware, but you have to “jump” to clear a knee-high wall. You have a million dollars at the end of the game, but you can’t buy anything except hot dogs and cab rides. You can’t buy extra apartments or parking spots. You can’t buy the cars or bikes you want—you have to steal them, but they’re often not available until you’re already driving the one that you want. I’m not sure whether this is some sort of existential statement on the part of the game designers or not.

Stunt jumps and pigeons

The auto-save feature is great, but it inexplicably doesn’t save after stunt jumps or pigeons, both of which are needed to achieve 100%. The designers add all these goals and know that, over the course of the last several versions, the 100%-achievement has become the goal of many games. However, they don’t offer much help in getting there. You have no idea which jumps you’ve done or which pigeons you’ve killed.

This is where the suffering comes in; if you want to do all 50 stunt jumps, you have to search high and low and probably do them all twice in order to be really sure you did them all. That’s a shame because the bloody things are in the game already—Rockstar has already paid for them—but they make them so hard to find. Why do I have to go to the Internet—to the notoriously unreliable world of documentation created by 14-year-olds—in order to finish the game? While I’m prepared to spend 72 hours playing, I’m not going to spend two times that amount of time just driving around looking for stuff. That’s not fun.

Rockstar came so close in so many places. There’s an Internet where you can look up the pigeon or stunt jump maps and see where they all are. But they don’t show you which ones you’ve already done. A pity. The vigilante missions are documented for you in the statistics and the “Most Wanted” missions are also clear. You can put almost anything else on your GPS, but not the jumps or pigeons. It’s OK to make some things a bit more difficult, but don’t make them tedious.

Grinding missions

It’s similar with the main missions. If you die, you wake up in the hospital and pay a lot of money. That’s fine because you generally have a lot of money. At least you don’t lose your weapons. Seconds after getting out of the hospital, you get a text message asking if you want to re-try the mission. You can accept and you can try again. Without armor.

Since you failed before with armor, the odds that you’re going to succeed without it are slim. So, you have to ignore the redo-this-mission message and go shopping. There is some armor laying around the city, but it’s really hard to find and you only get a map for it when you’ve completed all the missions, after which you don’t really care that much about armor anymore.

The magical Internet comes to the rescue with maps of armor locations—and instructions of varying reliability for finding said armor—but it’s often still quite far away from you. And you don’t have a vehicle when you get out of the hospital. Are we having fun yet?

So, step one is to find a cab, get into it without stealing it (also not that easy, see below) and direct the driver to the weapons store, where you can buy armor. The cab usually leaves immediately and you’ll often be hard-pressed to find another. Of course, you can usually steal another car and drive yourself to the start of the missions but, when you’re grinding missions, you’re not in the mood to do five minutes of driving, gawking at the beautifully rendered city. Your fun is currently doing the next mission. The rest of the stuff is shopping, which is not much fun.

Stealing cars

The game is called “Grand Theft Auto” for a reason. You’re going to be stealing a lot of cars. But, in contrast to previous incarnations, you don’t have to drive yourself everywhere. You can take a cab instead, which is a relatively realistic evolution of the “Trip-Skip” feature in GTA III. It’s not as quick, but it’s more realistic.

To steal a car, you press the triangle button. To hire a cab, you press and hold the triangle button. Do you see the potential for trouble here? The heuristic for detecting a press versus a press-and-hold is not as reliable as one would hope. I don’t have enough fingers to count the number of times I tried to get into the only cab I’d seen in minutes and ended up watching the game make me steal it instead. There was invariably a cop car immediately behind me to make the extracurricular (remember, grinding missions) field-trip even longer.

I think they could have made the default action for a cab be to get in the back-seat and make you press and hold to steal it (something I never once wanted to do). Sure, it’s kind of cute to have the game be full of surprises, which often leads to completely believable consequences, but it shouldn’t find some way to make your game longer even if you’re a good player. The game shouldn’t be artificially lengthened by fucking with you.

The problem with Grand Theft Auto is what’s so awesome about it: the immersion. When you’re immersed in the game and you’re yanked back out of that immersion by a reminder that you’re playing a game, it’s frustrating—to put it mildly. As a player, you’re fooled into thinking that Nico—your character—is actually looking at the car you want to steal because he’s facing it. However, when you press the triangle button, he moves into auto-sprint mode and heads for another car off-camera. You mash desperately on the buttons to get him to head back to the other car, but you’ve wasted precious seconds and, sometimes, been killed for the glitch.

Internally, the data and algorithms the game uses to determine which things are around you and what exactly you’re seeing came to a different conclusion than the one you came to based on your interpretation of what the graphics system rendered. It’s wholly understandable that this should happen because what GTA does is hard. Really hard. But they do it so well that they pay the price—with frustrated users—when they only do it 98% of the time.

Motorcycles and helicopters

Motorcycles are also notoriously difficult to pick up or straddle. Nico has to be facing just right in order to get on and the bike can’t be laying in a strange position on the street or it’s impossible for him to get it. Immersion is dead. And so, most likely, is Nico. The final mission involves a motorcycle. After fighting through an entire army of flunkies, you have to jump on and continue pursuit on a motorcycle. I stood next to the thing, pressed the triangle button and watched Nico sprint across the dock on which the bike was parked and start to climb down a ladder into the water. Not exactly what I had in mind. By my third try, I’d figured out exactly how I had to be positioned in order to reliably get on the bike.

Getting killed in a hail of bullets because you stormed out of a bank into a 5-star police presence is expected—it was a twenty-to-one shot that you’d get out of there alive, but it’s still fun to try. Getting killed because your avatar is unable to get on a motorcycle or unable to climb a knee-high wall in an expeditious manner is not acceptable.

Actually, a casual, relaxed gamer would probably be able to laugh off failure due solely to game mechanics if that gamer didn’t have to invest so much time into getting to that point. Sometimes it’s north of five minutes of preparation and shopping to get to a mission, a further ten minutes of creeping through the mission, making few mistakes, then dying in a hail of bullets because the helicopter door won’t open when you’re standing an inch to the left of the invisible game-world trigger. [3]

Rinse, lather, repeat

As a reward for having played so well, you get to repeat the exact same thing again, which is very late 80s—early 90s. Whereas it’s nice of Rockstar to let you skip in-game cinematics, some of the missions include longer drives wherein you get to hear the same conversation over and over and over again. To their credit, Rockstar has actually included two separate conversations for many of the missions, for those of us that don’t reload a saved game and instead replay the mission.

The game’s hyper-realism is its downfall; when something doesn’t work as expected, you’re not only disappointed, you’re screaming at the television. Case in point, the mission called “Final Destination” involves hunting down a dealer, then icing him. Several things get in the way of this being an easy mission, though.

The dealer has a friend when you confront him; this friend opens fire as soon as the cut-scene is done and the game helpfully auto-aims on the other guy, who’s running away from you. Naturally, the other guy (who you have to kill to fulfill the mission) can take an insane amount of damage at this phase in the mission, because you’re not really supposed to be able to kill him so easily. On subsequent runs through the mission, you’ll learn to move the auto-aim reticle to the enemy that’s actually shooting you, then taking care of the other guy.

…who’s gone at this point because he hopped the tracks. Fair enough, you have to follow him, through the station and down some stairs. This is all pretty tricky navigation only because the console controller is as useless as tits on a man for this. Give me a mouse and keyboard back for first-person stuff, then we can dispense with the hilarity of auto-aim as well.

So, the guy invariably gets to the bottom of the stairs and invariably gets into his car. Will auto-aim let you take out his tires? His engine? His smiling face that you can clearly see through the windshield just five feet away? No. Once he’s in the car, he’s pretty much invulnerable. At this point, you’re left to working with manual aiming for the very first time ever in the game or getting into a car and chasing him.

Let’s tackle the chase scene first. [1] He’s a terrible driver, so he’s easy to spin out and push into oncoming traffic, etc., etc. You also have a weapon, which you can fire out the window. You can only fire straight ahead or out the driver’s side window, so you can’t pull up on his left side and ice him that way. In fact, shooting him from the side with innumerable bullets seemed to have no effect. Shooting his vehicle from behind was similarly effective. Pinning his vehicle sideways in between tool booths, seeing his head and shoulders silhouetted in the daylight framed by his window and firing numerous times had no effect on his state of being alive. [2]

Other times, his spectacular driving ability pins him into a secluded area without cops and no escape. He constantly drives into a wall while he’s running through his given lines. You are standing by the driver’s side window and auto-aim ignores you. You take over with manual-aim and shakily move the reticle to his head, squeezing off round after round, to no avail. You are not allowed to kill him this way. He must either die in a car crash or you must take him down outside of a vehicle.

In this particular attempt, he shimmied his vehicle up an embankment and out, but could only repeatedly ram his car head-on into a wall outside in seeming frustration. By the time I jumped a wall and got up there, he was blessedly on foot and could be hunted down. The “mission accomplished” music just seconds later was sweet bliss, but my gunfire had attracted the attention of local law enforcement. There were no vehicles around, so I hightailed it up the street, but was forced to a walk as the game forced me to call my employer at that very moment.

Your character can’t run and talk on the cell phone and your character always phones in immediately after a mission. That means the cops had a pretty easy time picking me up after this mission, as I was incapable of doing anything but moseying up the street and chatting with my boss.

The game is incredibly well-done and usually lots of fun; it’s that much more irritating when you’re thrown out of the illusion and realize that, clever as you are for having thought of shooting the guy’s tires, you can’t do that. You probably can at other points in the game, where the script calls for it. At this point, putting the reticle on Lenny’s tires and pulling the trigger made a bunch of sparks, but no flat tires. Shooting through the windshield and open side windows was not effective. You still have to follow the rules of the game.

LollipopGTA lets you do a lot of naughty things: hire prostitutes, gamble, run rackets, kill people, commit vehicular manslaughter and on and on. The game is named after a crime and features thugs, gangsters and women of ill repute all over it. Yet, once again, the white knights of morality are assailing it as being bad for kids. No kidding; that’s why it’s rated M for mature. According to the ESRB, no one under 17 should be buying it or even playing it. Problem solved, right?

Wrong. There are plenty of people who aren’t happy with the ratings system because it doesn’t prevent people from buying it—and thus, requires actual parenting and monitoring of children to prevent them from getting their grubby, little, budding-criminal hands on it. Case in point: the article MADD Protests GTA 4 Over Drunk Driving (Shacknews) covers a complaint by MADD (Mothers Against Drunk Driving) about a feature in GTA IV “where players have to drive drunk”. Not only do you not have to drive drunk, “the game [actually] recommends that players make use of a taxi once drunk”. What more could you want? The game puts the player in a real-life position where they could make the wrong choice, then advises them to make the right one. I repeat: what more could MADD want? Isn’t that pretty much exactly what a MADD video game would want to do? And here comes Rockstar investing dozens of millions of dollars to do their jobs for them. They couldn’t have planned it any better themselves and it didn’t cost them a dime.

Instead of rejoicing in the free publicity for their cause and patting Rockstar on the back for a job well done, they’ve instead asked for sanctions against Rockstar Games. “MADD is calling on the Electronic Software Ratings Board to re-rate the game as an Adults-Only title”; however, they apparently realize that upping the age limit by one whole year isn’t going to do a damned thing, so they’ve also asked that “Rockstar […] consider a stop in distribution.”

Side. Splitting. Laughter.

Broker Bridge HillEscape on a Crotch RocketGTA IV hits the ground running with a rain of reviews preceding the official release date (just in time to fuel midnight lines around the Best Buy) which, according to GTA IV Reviews: An Exercise In Hyperbolism by Luke Plunkett (Kotaku) are unilaterally gushy. Welcome to the land of opportunity by Crispin Boyer (1Up.com) is, apparently, one of the better ones. It gushes on some about the new multiplayer mode (which is apparently as good as advertised), then actually offers some criticism:

“I do wish that Rockstar had added checkpoints in the multistage missions to cut down on the grunt work of frustrating retries. GTA4 also suffers from those little things that have always plagued the series, such as sudden pop-in of objects in the environment and the occasional repetition of car models in your immediate vicinity.”

But, on the other hand, the game improves on its predecessors in its virtual reality simulation (which is no mean trick, considering how integrated San Andreas was). You don’t even have to do missions—you can just interact with the world by walking around:

“…Liberty City’s breathtaking vistas, incredibly varied scenery, and lived-in look … [t]he city just feels alive. Mosey on foot for just a few minutes and you’ll eavesdrop on the cellphone conversations of nearby pedestrians, witness cops arresting other criminals for a change, and even run into the random man on the street who will give you a mission or interact with Niko in some other special way. But more than just feeling alive, everything in this world is so integrated. [1]”

There’s also the article, Grand Theft Auto IV Setting (GTAIV.net), which goes into loving detail about the locations within the game and how they stack up against the real NYC.

From the look of the screenshots, the art direction makes it clearly a GTA game; the videos show animation that is familiar, just better. Video Teases GTA 4 Animation Technology (Shacknews) covers the GTA IV animation system in more detail. The video shows figures being struck by projectiles and falling down in real-time and with far more believability than traditional animation. Where most mainstream games these days use a bone animation system, they breathe life into the skeletons with motion capture data or “rag doll” physics. The combination, however, leads to less-than-realistic results because, whereas the motion-capture data is very life-like, it can only animate expected behavior. Anything else is rendered with rag-doll physics, which takes only the bones and their weights into account. [2] The GTA IV engine uses NaturalMotion’s euphoria animation technology to imbue its characters with life-like behavior through the spectrum; in fact, it requires no motion-capture data at all.

Middle of AlgonquinPanorama of the CityLiberty City at Sunset

That’s not to say that traditional motion-capture is dead. Another game, coming out in June, is Metal Gear Solid 4. They recently released an April Fool’s video, Konami/Ubisoft April Fools’ Joke: Metal Gear Creed (Shacknews) which, though it goes on waaaaay too long and isn’t a funny April Fool’s joke (unless you’re 15 and a total game nerd), features some amazing animation in the initial street fight scenes. It’s in fact quite difficult to tell whether you’re watching Al Jazeera or playing a game.

GTA IV also isn’t alone in using filtering to give its rendered scenes a more “artistic” look. Many recent games are moving away from the once-desired photo-realism and moving in a more pleasing—and fun—direction. One game available now is Team Fortress 2, which is a multi-player game with a whole host of cartoon characters instead of the traditional super-rendered avatars to which we’ve grown accustomed over the last few years. The game comes from Valve Software, purveyors of the hyper-realistic Half-Life 2 series; it uses the same engine. The Team Fortress 2 Trailer (Fileshack) is pretty huge, but lots of fun.

Team Fortress II: Sniper and ScientistTeam Fortress II: EngineersTeam Fortress II: Cast of Characters

Another game following in this vein is Battlefield Heroes. The original Battlefield was a breakout multi-player strategy and team-fighting game and its first sequel followed in its footsteps with the same thing in Vietnam—where it boasted better and more realistic graphics. The next sequel leaves all that behind and combines the same gameplay with a more fun, cartoony style. The two movies, Battlefield: BC Tutorial Teaches Strategy (Shacknews) and New Battlefield Heroes Trailer (Shacknews) should give you an idea of what you can expect.

Battlefield Heroes: the Royal ArmyBattlefield Heroes: Attack from Above

And finally, there’s Valkyria, a new game from Japan that uses filtering to render the game as if it was hand-drawn anime. The screenshots are quite good, though the illusion breaks up a bit when examined too closely; the trailer available at Valkyria Chronicles Trailer Illustrates Fine Filter Use (Shacknews) is fantastic, though. These days, video almost always trumps still screenshots—even those that have been touched up.

Valkyria: Cast of Characters

Video Gaming has come full circle.

It began with kids feeding quarters into machines at the local arcade, honing their skills with an endless stream of silver. The occasional talent would rise above the rest and gain fame in the neighborhood for his (or her) mad skills. There were masters of Pac-man and Space Invaders. Then came the home versions of these games, which allowed you to train at home, for free. People got better, but they left the arcades, taking the show out of video gaming. In the mid- to late-90's, the show was back, as multi-player gaming took hold and allowed players to strut their mad skills once again—this time by directly “pwning” other players to prove superiority.

Consoles got better and better, games attracted more and more players, with some games, like World of Warcraft, boasting tens of millions of players worldwide. Starcraft is nearly a state religion in South Korea, with at least two or three television channels devoted to showing games of the country’s top players. In the States, things haven’t come that far (yet), but the best players are crawling out of their living rooms and basements to once again strut their eerily well-honed skills in public. They have found a new home in the big-box electronics stores, like Best Buy, which set up little cathedrals for exactly this purpose: ostensibly to let people try out the console, but realistically to allow star players to showcase their talent.

The game that best lends itself to this type of thing is Guitar Hero or Rock Star, where players cradle a miniature guitar controller with five buttons on it and match a stream of colored lights on an animated fretboard with furious taps of the appropriately colored button. It’s exactly the same things as Dance, Dance, Revolution but without all the exercise and with the goal of becoming a rock god instead. To play well takes a lot of practice and no small amount of skill. As with all video games, the creators anticipated that people would invest dozens of hours in the mastery of their game and appropriately added songs that are so phenomenally hard that their inclusion looks like a joke to the casual player.

Not only can a player showcase his or her talents [1] in the local neighborhood again but, thanks to YouTube, can also upload agonizingly long and boring tributes to their prowess for the internet community as well. Whereas their prowess at playing fake guitar is clear, editing a video of what is essentially an unblinking person smashing five colored buttons for 8.5 minutes to a speed metal song that most people wouldn’t ordinarily listen to if paid to do so is something that is clearly not within their power. Watching a somewhat blurry video (thanks YouTube and handheld phone cameras) tends to focus the viewer on just what a clacky racket the controller makes when pounded on. [2]

Watching a tremendous display of talent live is something else altogether. Articles like Best Buy Bodhisattva by Julian Murdoch (Gamers with Jobs) convey wonderfully the ass-clenching excitement of watching someone just master something difficult, live and in person.

“…3, 4, 5 minutes into the song. Kyle slips deeper into what is clearly a state of Samadhi; He no longer perceives a space between himself and the game. There is no him. There is no song. There is no guitar. … At just over 6 minutes, the song becomes even more ludicrous. While actually playing it will ever remain for me an uncrossable gap, I am enough a student of the form to recognize the crux. He is Lance Armstrong approaching the bottom of Alpe D’Huez: Will he attack?”

The Lance Armstrong reference is pure gold.

It must have been really something to watch this youth arrive with his posse, dominate with a burst of talent, then leave without a trace. It would be easy to say that it’s nice to see that “kids these days” are good at something, but they always have been. It used to be skateboarding and its more physically demanding brethren—now it’s Guitar Hero. The things they’re interested in and spend ridiculous amounts of time on and becoming mind-bogglingly proficient at are just not guaranteed to be lucrative or make them successful or help them move out of your home, like, ever. It is this lack of applicability of nurtured talent that adult society has always had a problem with, as it watches sour-pussed as “kids waste their lives”. The instinct to channel energy into societally useful work is an acquired taste; it takes years of brainwashing to eradicate the innately human desire to chase rainbows without regard to time or goals. On the other hand, the man who wrote the article is 40 and is part of perhaps the first generation that manages to somehow appreciate and respect things that the next generation does.

As with so many time investments that seem futile when you’re not deep within the obsession yourself, one wonders why these people don’t just go learn how to play a real guitar instead. It’s much harder, but they clearly have the interest and time. It’s probably just too hard and most of the guys online look like they’re just trying to avoid doing their history papers for whichever large state university they attend rather than trying to acquire a true skill or exercise a true passion.

On a final note, South Park has already lampooned the genre in their usual exquisite manner, with an “audition” consisting of a kid tapping away accapella on the controller to show his chops as people gaze in wonder. Guitar Queer-o (South Park Zone) is available for free online.

Another game that’s been in development for a long time is Crysis. Earlier this year there was a pretty good review of it, Crytek’s Jack Mamais on Crysis by Chris Remo (Shack News), where the reviewer noted that “running on a beefy DirectX 10-capable NVIDIA 8800, the game was never short of gorgeous.” Granted, an NVidia 8800 is no shrinking violet of a graphics card—with the top-of-the-line version weighing in at $600+ and 177 Watts—but still, it was running the game “at a fluctuating framerate in the 20-30fps range” at the “ridiculous resolution of 2048x1536”. Naturally, the developer claims that the game runsjust fine on an older card, and that even a “[a] three year old graphics card should be pretty good. … It won’t be cutting edge with the graphics but it’ll be super good.”

The game apparently has a story because Crytek has “hired writers” and everything, including “a couple of Hollywood guys working on the story with us”. Awesome. A couple of Hollywood guys are a sure-fire way of getting quality adult-themed entertainment into a game. Just look at the quality writing in your average Hollywood action blockbuster. It’s a plan that can’t fail. Here’s a brief summary of plot points, player capabilities, etc.: nano-suit gives super-human powers, player can drive/fly all sorts of vehicles and “Attack and Defend” multiplayer mode that is completely different because it’s called “Power Struggle” instead. Oh yeah, and it’s non-linear in completely new and exciting ways.

Despite whining from the developer that “[i]t kills us when people call it a tech demo”, the thing that really screams “buy me!” are the screenshots. Behold:

Guarding the MothershipDeath from AboveFirefight

Hightailing ItSurrounded!The Tropics

All sarcasm aside, it looks amazing. It may be the same old type of game (super soldier clears out world full of whackos with heavy weaponry single-handedly), but looks to be one of the first that takes advantage of the kind of power a $600 graphics card can offer. So if you’ve got one of those lying around or you’ve got a wad of cash burning a hole in your pocket, put this one on your Christmas list … if it’s out by then.

For even more screenshots, check out Crysis Screenshots (Shack News).

The first screenshots of Halo 3 are floating around, showing it to be a next-generation game ready to squeeze every drop of performance and visual effects out of the graphics horsepower found in the X-Box. One in particular (shown below), reveals an impressively cinematic quality, with good material detail, reflections and very natural depth-of-field effects. Notice how the focus is on the right knee, with everything else appropriately softened. It’s not clear how well this level of detail translates to playability (it usually doesn’t add much, to be honest). The surroundings—other than the master chief, his rifle, his busted truck and a clump of grass—are sparse to non-existent, which is also not very encouraging.

Closeup of Master Chief

Still, it looks pretty sharp; for more screenshots, check out Halo 3 Screenshots (ShackNews).

Alan Wake is a psychological action thriller coming to the X-Box and PC in sometime in 2007. The game’s namesake is a writer, living in the woods somewhere, presumably along a coastline. It takes the realism of a Grand Theft Auto world to new heights, with forested lanscapes as well as small American towns rendered with an incredible level of detail. Throw in weather effects, a day/night cycle and realistic physics and this game has real-world environments like we’ve never seen before. The game takes place primarily outdoors, in absolutely huge environments—all with real-time lighting. The composite screenshots below show two scenes from the world of Alan Wake rendered at different times of the day, with different weather.

Light and Weather Sunrise / Sunset

As with most video games engines these days, screenshots, while pretty, don’t nearly do the immersive experience justice anymore. The video below shows the game engine in action, showing the world of Alan Wake, with a developer putting it through its paces on a quad-core machine. Naturally, the quad-core machine makes things move a lot more quickly than your older machine is likely to. Nonetheless, the attention to detail, cinematics and sheer immersiveness is impressive: even narrated by a robot and videotaped off a screen, the water, shadows, reflections and lighting are incredibly realistic. Make sure to stick around until the end, when a simulated tornado tears up several buildings and vehicles—pulling them apart piece by realistic piece.

And it looks like it has a story. No kidding. The movies available of the home page for the game are very impressive and worth watching if you’re sick of all of the best-looking games involving weapon selection and exquisitely rendered crates. I saw a lot of trees, houses, roads and sunsets, but no crates in any of the movies. Looking good. This game and engine come from the makers of the excellent Max Payne series, which also set new standards for both graphics and storytelling, not to mention introducing one of the most innovative gaming techniques in the last decade: bullet-time.

The character of Alan Wake is also rendered quite well and seems to integrate into the game world seamlessly—so as not to startle an immersed gamer back into reality.

Mountain SunriseTweed!No Lens Flare

As noted in the caption, the Remedy engine has opted for a more realistic approach to lighting than one was likely to find in previous engines. The “Mountain Sunrise” shot shows just how much better this “real world” will look than GTA (which is the current record holder right now, as far as this author is concerned).

Megatextured Pine ForestA recent article about Quake Wars: Enemy Territory mentioned that it was using the latest and greatest of rendering technologies from id Software, called Megatexturing. This Q&A with John Carmack includes more details on the development, timeline and features of this technology. He sees Megatexturing as a natural extension of texture memory management into the virtual space, as is already done for graphics memory and system memory. He’s quick to note that there is nothing groundbreaking about the idea, but that for him the solution popped out when he realized that “texture tiling and repeating is really just a very, very specialized form of data compression”. This data compression is necessary because “you generally don’t have enough memory to be able to have the exact texture that you’d like everywhere”.

The Doom3 engine’s approach to texturing now allows a game to act as if it has an unlimited texture space, using a “more complicated fragment program … [to] pick out exactly what should be on [the mesh]” at any given point on the terrain. Carmack sees this as a major breakthrough in realism for games, because artists can generate these large surfaces and then “do hand touch ups in a lot of different places to accent around features”. This patina of realism on top of the randomness is enough to convince the player that the landscape is real, something that is more difficult using only random terrain. It’s one of those “key elements that you start looking at in games that look really dated”; at a certain point, one realizes that the hardware has come far enough to toss out the tradeoff and just do it right, in the same way that the original Doom3 engine did away with pre-rendered lightmaps several years ago.

It’s important to realize that just because an artist can hand-paint the entire several square kilometers of a map doesn’t mean she has to. The engine runs just fine using repeated textures instead of one large one, but “[i]t’s taken [texturing] from being a resource constraint to something that becomes a design trade off.” The upcoming QW:ET will premiere this feature for games on the X-Box 360 and PC. But, as is usual with Carmack, it’s “essentially already abandoned” and has been replaced with something better. The latest version takes this “texture space” concept and generalizes it to all rendering in the engine:

“[A] similar technology that creates this unique mapping of everything, and use it in a more general sense so that we could have it on architectural models, and arbitrary characters … [this] is what we’re using in our current title that’s under development”

Carmack’s convinced that he’s got the defining technology for the next generation, and that, “just by itself, even with no newer presentation technologies, allowing unique texturing on lots and lots of surfaces, I think, is the key enabler for this generation.” He’s been right several times before when he bet on BSPs, 3D hardware, bezier surfaces and—most recently—bump-mapping and dynamic lighting. It’s a safe bet that id Software will once again emerge at the top of the heap in the next big showdown.

The Electronic Entertainment Expo—E3 for short—is taking place right now and is producing the expected wave of hype, “in-game” movies and “screenshots”. Check out the E3 insider or the IGN site for all the latest effusive marketing/reporting. Games that want to sell these days have to look good; to that end, publishers use high-quality screenshots and ridiculously good-looking in-game movies to draw in their audience. Recent years have seen an escalation in these types of tricks, with Sony releasing an “in-game” movie last year for hardware that doesn’t even exist yet. The movie was rendered on simulated hardware—most likely at an abysmal frame-rate—and sold as “what you will likely see in over a year’s time when the Playstation 3 finally comes out”.

Bullshot |bool’shät| – n. A screenshot fabricated by a company to misrepresent the graphics of a game; a combination of the words bullshit and screenshot.

What is a bullshot and what isn’t? What are the parameters for a believable screenshot? How fast was the game running when the shot was taken? On what kind of hardware? For whom are these detail levels recommended or even enabled? At what point do we stop calling them screenshots and start calling them movie stills? And finally, at what point is the level of detail so high that it’s a waste of processor power considering the pace of the game?

Hi-Def Humans

In the olden days, games used mip-mapping and level-of-detail (LOD) to ease the memory and rendering burdens, respectively, of a scene. Mip-mapping uses lower detail textures and LOD algorithms switches out high-polygon models for low-polygon ones for smaller, more distant objects. For all I know, this is still happening in game engines today, but, with a standard 128MB of video memory and lord knows how many pixel pipelines, texture units and sheer brawny transistor goodness available, the stakes are higher. In particular, the maximum level of detail is much higher, so objects seen from a short range have an incredible amount of detail. Take a look at the screenshots [1] from Brothers In Arms: Hell’s Highway below.

Up Close and Personal Stop Crying Soldier!

Each of the faces on these soldiers has more polygons than an entire character from any game made just a few years ago. The nearly photographic detail isn’t just in the closest soldier’s face, but in the clothes of the other soldiers behind him. Even the pieces of street rubble are unique and appear individually rendered. In the second scene, the soldier’s face is hidden in shadow he throws on himself, but his uniform actually looks like wool and look at the hands! Hands are never rendered with that amount of detail: this looks like Toy Story 2. Is it even conceivable that this is a game screenshot? Not on any hardware you’ll be owning anytime soon.

Geek Porn

This isn’t the first time screenshots have been released showcasing the Unreal 2007 engine [2]. The Hell’s Highway Movie #2 (FileShack) “shows the differences between the game’s previous engine and the new one.” It’s clear they’ve upped the ante on content considerably, but it’s such a tech-y video: the feature list reads like they’re trying to sell you the Unreal engine, not the game. Throughout the movie, the following callouts appear, to draw attention to a particularly cool rendering feature:

• High Definition; 16x Texture Resolution
• realistic, Non-Greasy Normal Mapped and Specular Lighting
• Real Time Self Shadows with Fuzzy Attenuation
• 16 Bit Color with High Dynamic Range
• High Fidelity Motion Capture “Anim Tree” Driven Animation
• Advanced Pixel Shaders with Subsurface Scattering

The focus here is so far away from what makes a game playable: it’s as if a salesman spent his entire sales pitch talking about the heated cupholders in the Porsche he’s trying to sell you. In a way, he can; you both know the car’s fast, but so are all other sports cars. They’re probably all good enough for you to cruise for chicks in—the deciding factor then becomes the small detail. For first person shooters, you know you’re going to be running around shooting shit with the same guns you’ve been shooting shit with for the last decade. But now you’re doing it with fuzzy attenuation and non-greasy normal maps. Woohoo!

Hi-Def Plants

No discussion of mind-blowing detail today is complete without mentioning the excellent work being done by Crytek with their CryEngine 2 and new game, Crysis [3]. The focus of the new engine and game is to render nature in all of its splendor, opening up a whole new vista of experience for gamers who haven’t left their parent’s basements in years. Though most prefer the dank, cramped hallways of Doom3, there are those who want to assuage an inner desire for the blue sky and green plants of nature. Crytek aims to fill this void:

Crysis − Shoreline Crysis − Sunrise Crysis − Jungle Warfare

A few years ago, publishers were spending a lot of money on server farms to render scenes like these for a few seconds of cinematics. Today, high-powered game engines deliver the same thing in real-time. The jungle shot, in particular, has a staggering amount of crisp detail provided by another litany of features: “Volumetric Clouds, Realtime Ambient Maps, Soft Shadows, Depth of Field [and] Motion Blur”. You can take a look for yourself in the Crysis GDC 2006 Movie (File Shack). The attention to detail is impressive, with plants reacting realistically to a soldier running through them and soldiers shooting them.

First Person Shooters Über Alles

It looks like a damned movie—but it’s still a movie of soldiers and tanks. This time they seem to be at war with nature as well as with each other. Though Crysis focuses on integrating full-blown nature, it still makes sure to dot the i’s and cross the t’s of the genre: guns? check. soldiers? check. Nice plants: brrrraaaaap!!!! mow ‘em down, they’re blocking the sight line to the enemy base.

How cool can we make soldiers look? Pretty cool as evidenced by the next set of screenshots:

Crysis − Attack the Shack Quake Wars − Group Combat Gears of War − Group

Crysis shows off its integration of nature with the man-made (with painstakingly rendered hands, of course). Interestingly enough, both Quake Wars [4] and Gears of War [5] have about the same style of ensemble. A closer look shows a level of detail higher than even that of fancy comic book covers—with shadowing and highlighting that are taking games closer and closer to photorealism. But it’s all about soldiers and fighting and blowing shit up. And looking cool while doing it. Bestowing cool on the little (or grotesquely fat, depending on your mental image of a stereotypical gaming geek) dwellar in the cellar is what sells these games.

The soldiers are all male and beefed up (at least in Gears of War) and the games even make sure to put the word “War” in their title, just in case you were too thick to get the point. I’m not complaining—I play these games too. It would be cool, though, if these amazing real-time interactive near-video technologies could be used for something more uplifting than the glorification of war [6]. Look at these heroic poses from Gears of War:

Gears of War − Single Gears of War with highlights! Quake Wars − Men in Sunlight

Huge freakin’ arms with a little depth-of-field to put the focus on your grimly determined self. Or there you are, crouched behind a wall, muscular neck straining to draw a bead on the ugly alien. [7] Sure, this looks fantastic, but make sure you know what the pretty, pretty, pretty effects are selling you. Real-time lighting? Yup. Specular highlighting? Yup. Overbright? Yup. Wholesale slaughter of the other? Bought and sold.

A Difference in Style

The third shot above, Men in Sunlight is included to juxtapose, once again, Gears of War and Quake Wars. They may both feature soldiers, guns, the word “War” in their title, a story about repulsion of an alien attack and an emphasis on post-apocalyptic scenery, but they still look different. The art direction in Gears of War is distinctly more cartoonish than that in Quake Wars. This is not a judgement of the Unreal engine, but games made with the Doom engine just look more like photos of the real world than those made with the Unreal engine. Unreal games tend to have characters that look like they dropped out of a Japanese cartoon whereas those in Quake/Doom games tend to be proportioned like people from Earth. The two group shots above show this difference quite clearly as well.

So what’s to conclude from this long, rambling half–love-affair with awesome graphics and half–diatribe against the unrelenting focus on war in video games? The upcoming games look amazing; the emphasis on look leaves you wondering whether they’ll play about the same as everything else you’ve played, but this time with realistic squad behavior! If you want the games to look like the shots above when you play them, you probably need to invest heavily in your PC or just buy a damned X-Box 360 [8].

Fighting Off the Strogg 'Bugs'The next installment in the Quake world is being designed not by id Software, but by Splash Damage [1]. It’s a multiplayer-only game set in massive outdoor environments using the Doom3 engine. Quake Wars: Enemy Territory Q&A (Shack News) is an interview with the lead developer. That’s right, you just read “massive outdoor environments” and “Doom3 engine” in the same sentence. How can this be?

Those familiar with game engines know that each has its strengths and weaknesses drawn from the type of game for which they were designed. The Doom3 engine is exceedingly good at moodily-lit, shadowy, cramped interiors with its “real-time lighting everywhere” paradigm. The original game had almost no outdoor sequences and the few tiny ones it had made noticeably larger demands on the hardware. The Quake3 engine was similarly built to the specifications of the game design: a fast, good-looking, multiplayer game with battles occurring in small arenas. Carmack [2]—being Carmack—threw in true bezier curves just to keep things interesting for himself, but otherwise pretty standard for a shooter today (though quite revolutionary at the time). Map designers that attempted larger outdoor environments were either frustrated or learned a lot of mapping tricks to speed up the rendering. However, with the release of Q3: Team Arena, id Software had bolted a terrain-renderer onto the main engine in order to allow much larger maps and relatively-realistic and curvy terrain.

Lush Landscapes

Pine ForestsTroops in a ForestMegaTexture technology is a similar advancement for the Doom3 engine. This technology has two components: a renderer and a toolset to generate content. The renderer—written by John Carmack—lets the game access a “five gigabyte source texture” to draw the visible terrain, while using “only around ten megabytes of video memory and twenty megabytes of system memory”. Content designers use a tool called MegaGen to build and maintain this enormous texture:

“automatically distribute materials such as grass, sand and rocks across the landscape, based on altitudes and the angles of incline. MegaGen makes moss grow up the steeper slopes and cling to rocks, grass grow in the flatter areas, and sand and snow gather appropriately in the crevices between rocks. Our Artists are then able to paint additional fine unique detail such as cracks in road surfaces, or they can texture modeled elements such as shell and plasma blast craters in the terrain.”

As the screenshots to the right show, this id engine has no trouble rendering extremely convincing, lush and natural environments on a grand scale. As befits an id engine, the colors are much more muted than in other outdoor games [3]. Though graphics are what sells the game in many cases, a multiplayer game lives and dies by its immersiveness, which the Doom3 engine had in spades when it was rendering indoor environments.

But, what makes a game immersive in multiplayer, outdoor environments?

Ensuring Immersion

Physics and networking, naturally. If the environment reacts as expected and the player doesn’t notice the network, there is no loss of immersion. id Software has a very good reputation in the networking department as well, being known for delivering the tightest networking code around—a fact which accounted for the popularity of Quake 3 and its spawn over a number of years. They’ve added refinements to networking for large maps, adding an “‘Area of Relevance’, which works somewhat like ‘Level of Detail’ for graphics”. That is, the player is sent only the information that is useful in a situation, optimizing away irrelevant details that would be important much closer. “[W]hich way [a player’s] head is facing or how many grenades he has” isn’t important when that player is a mile away. Easing network load in this way allows more relevant information to be transmitted (like physics data) and allows more players to take part.

On the physics front, id Software has worked with Splash Damage to improve the home-grown physics engine they employ to support “suspension, propulsion and friction” and to “derive gameplay-affecting properties from the texture”, which, as mentioned above, is a huge, unique, painted tapestry rather than a collection of large tiles.

“This lets us have great off-road vehicles that can climb rocks, boats that have buoyancy and flying vehicles that react the way you’d expect to lift, drag, thrust and friction.”

This interview, Enemy Mine: Todd Hollenshead Speaks, gives a little more information on this game and its immediate precursor (Wolfenstein: Enemy Territory).

In the days of the telegraph, human operators sent telecommunications by hand using morse code. Each person had their own cadence and style of sending messages; the styles were so unique that another operator could unfailingly distinguish which person sent a message simply by hearing the message arrive. This style was called the operator’s fist.

A careful perusal of screenshots from different upcoming games shows similar “fists” emerging for the game engine driving them. Over the last several years, game engine licensing has increased in popularity as more and more is expected of a game and as the costs increase accordingly. Two of the major licensers are id Software, with their Doom Engine and Epic Software, with their Unreal Engine.

The Test

To even the playing field, all of the screenshots are from pseudo-futuristic nightmare worlds in which soldiers battle aliens. We can thank the inventiveness of the video game publishers for lending validity to this experiment. Each of the screenshots below are from one of the two engine mentioned above. Can you tell which?

(A)

(B)

(C)

(D)

(E)

The Answers

Before you read the answers, note how similar the game titles are: the first four shots have human soldiers fighting alien armies. The aliens look strikingly similar. Most of the weapons are ridiculously oversized. It’s only the style of the engine that lets us tell them apart.

(A) and (B) are from Enemy Territory − Quake Wars, which uses the Doom Engine. (C) and (D) are from Gears of War, using the Unreal Engine and (E) is from Prey, which uses the Doom Engine.

Why games using the Unreal Engine tend to have a more “cartoony” look—oversized weapons, child-sized heads atop enormous bodies, usually encased in full-body armor—while games using the Doom Engine tend to the more grotesque, is not clear. Even the Prey screenshot, a game with a completely different approach to gravity and gameplay, includes an alien that just screams Doom Engine.

The results don’t necessarily mean anything; it’s just interesting to see if you can tell which game engine drives a game just from a quick glance at a screenshot.

Neverwinter Nights 2 is a role-playing video game coming to the PC sometime in 2006. It hasn’t been that long since a “check out these awesome screenshots” article [1], but I like screenshots. The last batch [2] were of a hyper-realistic racing game set in real cities around the world—a very angular, metal and concrete world. Neverwinter Nights 2 takes place in a much more organic, open world with trees, grass and other bits of nature as the main backdrop to the adventure.

Using Those Pipelines

Cottage in the ForestBridge Through the Forest

As game developers expect the average graphics card to have more and more pipelines and be capable of applying more and more rendering passes per cycle, the quality—and realism—increase accordingly. Trees use many more polygons to approximate nature; fog and dynamic lighting effects are layered to “soften” the entire scene. Even the grass in the foreground gets enough treatment so you really have to look hard to see where it repeats. This all helps the immersion and keeps the gamer in the fantasy world.

Day/Night Cycle

Building at Sunset

Many games these days are far more sophisticated than their predecessors, incorporating a day/night cycle and weather effects to light a scene. Gone are the days of a single, global light source for base lighting. The shot above shows a castle on a cloudy day, with a late sun illuminating it. Note how the organic greenery blends with the building structure instead of being strictly demarcated. Even the cloudy sky is no longer recognizable as a “skybox”, if indeed it’s even rendered that way anymore.

Believable Structures

Harbor at SunriseRural Town

Even close up, towns maintain their smooth blending with natural elements and. Roofs are somewhat slumped without looking polygonal and the ground is uneven and worn. Wood looks like wood and the sun sparkles off of lighter surfaces realistically. Boats in the distance disappear convincingly into the haze and all components throw soft, convincing shadows. One shudders to think of the number of rendering passes required for this scene—multiplied 30 or 40 times per second.

Single Rendering Model

Warthogs in Wheat

The other screenshots are suspiciously empty of characters and creatures. In previous generations of game engine, the transition from character rendering to environment rendering is typically where immersion broke down. Characters floated and glided over rough surfaces, threw shadows through other objects, failed to cast shadows on other objects and were lit with completely different light sources than the envioronment. With the advent of the DOOM3 engine, that was all part of the past. More and more games have crossed over to rendering a scene as a single composite rather than as a statically lit environment with dynamically-lit models.

Note how the warthog characters above blend into the grass and how their shadows fall realistically on the uneven terrain. Even if the creature itself is too fantastical to be true, it’s so much easier to believe in when it interacts convincingly with the environment.

Role-playing isn’t really my thing, but I applaud this team’s efforts to bring natural environments to gaming—I don’t think I saw a single crate in the whole batch of screenshots. Bravo!

Project Gotham Racing 3 (Ars Technica) reviews the latest racing game from Bizarre Studios for the Xbox 360. The previous incarnations were known for their graphics. This one will be known for its photorealistic graphics. I’ve seen a couple of movies of the game in action and it is nothing short of breathtaking. Even when the action stops for a quarter of a beat as a Ferrari slews around in a 360º turn, the city backdrops are so perfect, you can’t tell you’re not watching a movie. The cars are perfectly rendered and the only giveaway is the complete lack of dirt and dust on them.

See? No more polygon look; no more clipping—just realistic environment mapping, specular highlighting and soft shadows. Even in action, like in the high-definition Project Gotham Racing 3 Movie (FileShack) or Project Gotham Racing 3 Summit 2005 Trailer movies, effects like motion blur and field of vision are used to create very cinematic gameplay.

The highly realistic environments are based, with an incredible attention to detail, on the cities of New York, London, Tokyo, San Francisco and Las Vegas (and possibly others—the home page for the game isn’t very forthcoming). The previous incarnation also had real-life cities (the same ones excluding Vegas), but screenshots from that game show good graphics, but the typically flat, overstretched, poorly stitched and blurred textures common in environments today. Good, but not photorealistic.

The realism in PGR3 is due to a painstaking level of detail both in the cars and environment—details that will never be noticed outside of a frozen screenshot, but that add that “little something” that make the graphics truly believable. In the head-on views of the cars above, you can see that the headlamp fixtures are all rendered in 3-D so that they look right from all the cool camera angles and don’t lift you out of the immersion. Similarly, the environments are wire-framed and built on architect’s tools, as seen in the map of San Francisco below (on the left). Textures, bump maps, and fine structure are added to produce screenshots like the facade in Chinatown (in the New York map) to the right, which take several seconds of study to distiguish from a photograph.

Now, realistic is one thing—that means it’s convincingly photorealistic. A fantastic accomplishment in and of itself. But are the environments actually real? That is, do the cities as rendered match up with reality? Awesome PGR3 vs Real Life Tokyo comparison pics (GAF) has a discussion of this topic and lots of screenshots with accompanying photographs of the same spot in the real city.

The three shots below have the photo and the screenshot juxtaposed on top of one another. They are all from Tokyo and it takes some examining to determine which one is the fake (hint … Japan has traffic :-). The results speak for themselves.

So it seems that one of the holy grails of video gaming has been achieved—photorealism in real-time is here today. Game companies are putting in the time and resources to make believable assets and real game worlds with accompanying physics. The X-Box 360, at least, is capable of presenting these graphics in all of their glory, especially when output as 1080i on a digital TV.

It’s enough to make you run right out and buy an X-Box 360, a digital TV and PGR3. Good thing for my wallet it’s not so easy to get an X-Box 360.

The third edition of Grand Theft Auto, San Andreas, is chock full of dozens of different activities. There are three large, main cities, a whole countryside with several smaller towns and a number of rivers and bays to jet around on. There are naturally the story missions, but also jobs to be had around the city as well as more collection missions than ever. The number of things to do seems to have tripled since Vice City. The game is far from linear, but neither is everything available all at once. Highly detailed online guides are available and there are lists of which tasks must be accomplished to get the vaunted 100% complete rating. The problem is that details as to what is available when are not easy to come by.

The following diagrams lay out all the 100% tasks in order of when they become available.

Los Santos

The game starts in Los Santos, with introductions to your gang, the game mechanics and how to live in the GTA world.

Los Santos Missions and Resources

Los Santos as PDF

Countryside

After a while, the action moves to the Countryside, showing off a side of the GTA world never before seen. A modern graphics card breathes amazing life into this natural world.

Countryside Missions and Resources

Countryside as PDF

San Fierro

After a few missions in the countryside, San Fierro becomes available.

San Fierro Missions and Resources

San Fierro as PDF

Las Venturas

There’s lots to do in San Fierro and after all the missions are complete, you’ll be able to move on to Las Venturas and the desert, which will include the dreaded flying missions.

Las Venturas Missions and Resources

Las Venturas as PDF

Finale

With Las Venturas and all of its missions complete, you return to Los Santos to reclaim the territory you lost since you left (were thrown out).

Los Santos Final Missions

Finale as PDF

Diagrams made with MindMapper 3.5.

The Next Unreal Tournament gives a preview of the engine and development process that will create UT2007, coming to a graphics workhorse near you in 2007. Epic is overhauling the gameplay in this version in order to address some of the issues affecting the previous two incarnations of UT. Namely, that they didn’t seem to be as much fun to play as Quake because of all the bouncing around, unbalanced weapons and medium- to long-range weapons play.

They’re taking a new approach, tweaking level design before finishing (or even starting on) textures and making it “pretty”.

“You can tell from our [demo] levels,“ Steve says, “we’re not trying to make a prettier UT2K4. We’re really trying to make a very new game.”

Detailed facial animationsPart of the reason for this approach is the level of detail afforded by the Unreal 3 engine; there just isn’t time to readjust textures and art if the gameplay of a level turns out bad. Now it’s a matter of economics in the age of multi-million-dollar games.

Bots

I mentioned Quake earlier for a reason. Reading the article, I get the impression that Epic is acknowledging several game issues that id Software had already ironed out with Quake III. For example, they note that “more people play offline than they do online”. That means that your bots can’t just be lame clones of one another and need their own personalities. The ones Quake III provided did this to the best of the technology’s ability, and the Unreal 2007 bots are finally bumping the level even farther.

The bots will speak (trash-talking and responding to orders) and also respond to spoken commands, relieving players from learning arcane “talk” commands that no one could ever really use. This will let a player control bots in the same way that players run teams at LAN parties (because they’re all in the same room and communicate effiently).

“For example, each level will have specific locations that you’ll be able to use voice command to tell the bots to act on, so you can say, ‘Go cover the cavern,’ and they’ll know what that means. They’ll also be a lot more responsive, so you’ll be able to ask them questions about their status: ‘Is there anybody down in the cavern?’ ‘No, cavern’s all clear.‘ Or you can throw out warnings like, ‘There’s a sniper on the tower. Go get the sniper.‘”

This is very cool, but more evolutionary than revolutionary, as UT is still going to be about a “tournament story [that] kind of ties together why you’re playing these different matches”. Sounds like the same complex back-story that Quake III had.

Player movement

UT is finally accepting that “ it’s not so cool when you can’t hit the people you’re playing against because they’re doing [so much jumping around]”. Wow, what a realization. Quake has long held the title for player movement, with Half-Life 2 adopting the same more staid approach to moving around a level in a way that makes the player feel a part of the world rather than a ping-pong ball.

Combined with the relatively overpowered weapons and lack of balance, “you miss the feeling you get from being in someone’s face and fighting with them” because players tend to run and hide in the immense levels and snipe from a distance or spam from medium distance.

The techonology is amazing and seems to be vying for the best engine-driven game crown with id — except that id stopped playing that game with Quake III and released a pretty good single player game in Doom III. I guess there has to be some game that can be used for one-on-one tournaments in the future; it might as well be one with the word “Tournament” in the title.

Half Life 2 is a breakthrough game for one main reason: its physics engine. Don’t get me wrong, the Source engine looks nice, but its graphics don’t impart the same atmosphere as Doom 3, which has much more detail and immersiveness. The graphics and sounds are good, but not revolutionary. There is a lot of attention to real-world detail and architecture, which pays off; the screen shot to the left is one of the best available, but isn’t representative of average in-game graphics.

The physics engine, on the other hand, is an order of magnitude better than anything I’ve seen in any other game. The levels and entities in them are incredibly well built to show it off. Barrels, propane tanks, circular saw blades, boxes, paint cans and other debris litter the levels. You can slice enemies in half, blow them up, set them on fire and knock them around using only the “gravity gun”; this is the kind of fun no other game has managed to deliver.

The Ravenholm level included with the demo is long and intricate, giving a potential buyer more-than-adequate play time. It’s built like a Rube Goldberg theme park, with traps and ample things to throw at hapless enemies below. Throw crates against walls until they break to get supplies hidden inside (though the box is actually marked with its contents now, so you’re not forced to break everything). Throw a cardboard box against the wall and watch it’s sides break open and flap around convincingly. Roll a barrel off of a roof at an angle and watch it tip exactly when it should, then plummet to the ground.

So, that’s the good news. Once you get into the game, it’s really cool and offers a new kind of fun for first person shooters. Note that I said once you get into the game. The in-game experience is great, the out-of-the-box experience not so much. That’s all of the time you spend from the moment you decide to play Half Life 2 until you’ve got your trusty gravity gun in your hands.

That part is far from amazing.

Getting to the game

I’m going to focus on this part because there are enough fawning reviews of Half-Life 2 that ignore the shift in content delivery. It deserves attention because this type of access to a game is heralded as the future. It’s the way content creators are going to treat their customers until we rise up and revolt (or stop buying their stuff, which is an even less likely scenario). The content delivery system is called Steam and controls which Valve content and games you have access to. It is also their multi-player platform and controls your whole Valve/online gaming experience. It’s basically a good idea, keeping gamers up-to-date and keeping out cheaters and pirates.

You download the demo through Steam (if you already have it installed) or as a massive 750MB setup. Yeah, the demo setup is almost 1GB. It gets better — the installed demo is 3.3GB. I’m honestly not sure what they were thinking or what their target market is here. The only logical reason I can think of is that they were just too damned lazy to repack their content from the Half-Life source files and just gave you the whole damned game content. I ended up installing the bloody thing twice because it took up too much space on my system drive.

So, you’re finished installing your demo and start the game. Here, you’ll have to create a Steam account (if you don’t have one already) and log in. It’s nice that the account doesn’t require much private data to be entered at all. Now comes the next problem.

Level loading speed

It’s slow.

The game itself plays really well on all types of machines (I even saw it looking and playing decently on a machine that was well below the minimum requirements). That’s not the problem. It’s the startup and level loading times that are atrociously long. I’ve got a relatively new laptop that’s far above the recommended system specs for Half-Life 2. It takes 30 seconds of thrashing on the hard drive before the screen goes black and another 40 seconds until I can select a saved game.

Loading a saved game for the first time takes anywhere from 25 to 35 seconds and reloading a quick save point takes from about 7 to 12 seconds. I know these games have a lot more content than games of old, but that’s just slow. There are also load points in the middle of levels, which some reviewers have designated as barely noticeable. They must be suffering from some sort of syndrome as they lasted anywhere from 10 seconds to about a minute for me. Then, if you encounter enemies in the newly loaded part of the map and backpedal into the save point, you’ll get to load the old part of the level and the new part of the level again, swearing the entire time (that is, if you noticed that it happened).

Steam gotchas

Actually getting into the game also isn’t always a given; Steam plays tricks sometimes.

You’re in the game and looking around and it’s awesome — you want to show a friend. So you quit out of the game, pack up the laptop and go to their house to show him (or her) the new demo. No luck. You’re not online anymore, so Steam can’t check your account, so you can’t play. You can’t play single player either; you’re just locked out of your game entirely*. I only had the demo, so I could hardly feel cheated (though I managed to somehow), but, had I bought the full game, I’d be a little pissed that I couldn’t play something I’d purchased. I’d actually be a lot pissed, but being pissed comes kind of easily for me, as you may have noticed. At any rate, get used to the screen shown on the right.

*There is, apparently, an offline toggle somewhere in the settings, bu I haven’t found it yet. Perhaps demo users aren’t allowed to play in offline mode.

So you have to be online to play, even if you’re only playing single player. That brings me to the next gotcha — Steam wants to keep all copies of the game in sync, so that when you play multi-player, everyone has the same version. Every time you start the game, it checks whether you have the latest version and, if you don’t, it starts downloading it. Without asking! I actually started this article because I had to wait while Steam dowloaded who-knows-how-much content to my hard drive (it’s not like it shows how much it’s downloading or how fast … that would be too much information). It was downloading for about 11 minutes, so, since I have broadband, it must have been quite a lot of content it needed to update.

So, here’s the point: I had to go somewhere in about 15 minutes and I figured why not kill the time by playing a game. Double-click the Half Life 2 icon and it starts updating itself — even though I only wanted to play single player and the copy I had was running just fine. It would be far better to ask if it should update and tell me that multiplayer is unavailable if I decline. Again, I’m honestly not sure what Valve is thinking here; they are obviously not afraid of alienating their customers with decisions that are only convenient for Valve.

My laptop was naturally still downloading as I left; maybe I’ll get to play a game tomorrow.
 

VulgarThat’s the name of the new expansion pack being created by Nerve Software using the Doom III engine and properties. DOOM 3: Resurrection of Evil (PC) (GameSpy) includes lot of details about it.

On piece of good news is that the Soul Cube makes a comeback. This is a device that you obtain in Hell in the first game that destroys one enemy instantly when used. In doing so, it sucks out its life force and passes it to you. Since you have to kill five other enemies to charge it, it’s a very well-balanced weapon and a lot of fun to use strategically (it also happens to kill the closest creature, so it’s also all about aim to avoid wasting it).

Half-Life fans will raise a tumult, but the “Grabber” will be making an appearance in the expansion, which is pretty much the same as the “gravity gun” from Half-Life 2. Doom also uses a variation of the physics engine that Half-Life has (Havoc), but didn’t enable it in as much of the world as Half-Life 2 (most likely to control the flow of the script more). The expansion pack will offer the Grabber as a weapon, enabling Half-Life 2-style rampages of just throwing stuff at enemies. It will also work on non-metallics:

“…you can grab projectiles from enemies and fling them back, or even pick up small enemies and toss them at each other.”

The story line is also kind of interesting, involving a hunt for three “hunters” sent out by Dr. Betruger: “Each of the Hunters has its own special ability, and the artifact is able to “steal” these abilities once each Hunter is defeated.” The expansion continues to incorporate (euphemistic, I know) features from other popular games, like “an ability called “Hell Time,” where everything in the world but you slows down.” This is analogous to ““bullet-time” from the Matrix movies and the Max Payne games”, but “it’s not quite the same, since your character still retains his normal movement speed.” So there.

Doom III was a work of art, simply a visual masterpiece from beginning to end. Sampling the best gameplay features from two other extraordinary titles (Half-Life 2 and Max Payne 2 if you weren’t paying attention) will only make this expansion all the more interesting.

So every year, EA Sports comes out with another soccer game for the PC and every year they make a new download just in time for the Christmas season. These downloads are always a catastrophe, in that you spend more time looking at ads for the game than actually playing it before it quits back to the desktop on you.

I stopped downloading these demos a couple of years ago, but couldn’t resist when a pretty nice-looking demo from CodeMasters showed up, touting club football with Real Madrid. Here’s what you get for your 530MB download:

• It starts off with a bombastic set of “made by” and “produced by” screens that are accompanied by the loudest possible volume your PC can attain. The normal volume you were using for other stuff is now way too loud.
• Alt+Tab to no avail, use keyboard shortcuts to decrease volume. Nothing. Hammer the “Esc” and “Enter” keys to get to some sort of menu from which to quit. Game proceeds at own pace, despite all.
• Finally get to a menu, click to quit, and get two screens of marketing for a game that has only tried to deafen me so far. Each screen stays up for ten seconds; keys do nothing.
• Sound adjusted, jump back in.
• Start up a demo game and the typical TV camera zooms in to show the field and the players and everybody’s stretching and damned if it doesn’t look exactly like EA did last year (and probably this year).
• Esc and Enter have varying effects, each dismissing different screens as useless rosters and menus flash by in my desperate search for a game to play.
• Finally get to the field and it plays and handles pretty much like I remember, except, despite using all of the same keys as EA, they managed to switch the meaning of a few, so you end up doing some pretty comical things.

  1. Attempting to strike the ball anywhere in my own penalty box or trying to clear always resulted in a fabulous rendered own goal
  2. Defending was almost impossible, as the run button is no longer “W”, but Shift (which triggers Sticky Keys in windows and kicks you out of the game, crashing it. Shut Sticky Keys off.)
  3. Shooting was interesting as every single shot I took left the stadium and went directly into orbit. Cannot figure out why the default is set to make shooting “impossible”.
• Once the computer has scored on you (or you have scored on yourself), you’re treated to an almost endless litany of replays, which are only somewhat interruptible with assiduous enough smashing of the keyboard and swearing.
• The game, as with EA, is also over in 2 minutes, but instead of quitting to the desktop, you get the aforementioned 20 seconds of advertisting, then get to start the demo over.

Believe it or not, I played a couple of times, but it’s hard to get a feel for the game with such a short demo. Typically, there are pretty scant instructions for how to shoot, etc. Presumably all becomes clear and fun once you plunk down your $50 for a copy of the game.

I think I’ll just wait for next year’s version.

Benchmarks

The official benchmarks have been released by id. id Software’s Official DOOM3 Benchmarks (HardOCP) covers it in detail, showing you that if your card was purchased in the last year, you can probably play the game at high quality. Let’s get the recommendation out of the way:

“There is no way for a $500 [ATI] X800XT-PE to compete with a $400 [NVidia] 6800GT when the GT is simply going to outperform the more expensive card by a good margin. … for those of you that are in the high end video card market, the GeForce 6800GT looks to very much be the sweet spot when it comes to playing DOOM 3 with all the eye candy turned on at high resolutions.”

That’s right, the NVidia part is noticeably better at playing DOOM 3. This is probably due, in part, to “the Nvidia drivers hav[ing] been tuned for Doom’s primary light/surface interaction fragment program”, but also due to 4 more pixel pipelines than ATI’s part (16 as opposed to 12). Take a look at the benchmark and ATI is pretty much 25% behind on all of them. DOOM 3 is purely bandwidth/pipeline limited in the graphic card.

If you’ve got the latest and greatest card, the 6800 Ultra, you can expect almost 70FPS at 1600x1200 in High-quality mode with 8x Anisotropic filtering (makes textures blend more smoothly and look better). With 4x Anti-Aliasing (no more jaggies and more photorealistic look), you can play at the same rate at 1024x768. The lower-priced GT part gets about the same here.

If you’ve got the year-old NVIDIA 5950 Ultra or the ATI 9800XT, you can play at Medium Quality at 1024x768 at 50FPS. You’ll get about 45FPS at High Quality with 8X Anisotropic filtering. That’s actually quite good performance if you’ve already got one of these cards.

Quality modes

Robert Duffy, of id Software, has a .plan update (Shack News) explaining some other stuff about DOOM’s graphics usage and the Ultra Quality level. Loading all of the graphics in uncompressed format (texture, diffuse map, specular map, normal map) will, “[i]n a typical DOOM 3 level, … hover around a whopping 500MB of texture data”. Since pushing this kind of data around can easily lead to “50+ MB … of texture data referenced in a give scene per frame ( 60 times a second )”, it gets choppy even on high-end systems, so DOOM 3 “require[s] a 512MB Video card before setting [Ultra-Quality mode] automatically.” I don’t even know where you’d find such a card today.

The high quality, medium quality and low quality modes basically “uses compression for specular, diffuse, and normal maps” High quality mode does not use compression for the normal maps, so bump-mapping still yields graphics that are “very very close to Ultra quality but the compression does cause some loss”. Low quality additionally “downsizes textures over 512x512”, allowing 64MB cards to play the game comfortably. Basically the 4 modes are directly equal to the amount of memory you have on your card:

• Ultra = 512MB
• High = 256MB
• Medium = 128MB
• Low = 64MB

The features supported by your card also play a role in how everything looks, but, basically DOOM’s quality is determined by memory, rather than numerous tweakable engine settings.

OS X and Linux

On DOOM 3 Demo, Linux & Mac Versions (Shack News) sheds some light on id’s cross platform approach. The game is going to ship for the PC, is already slated for the X-Box (and probably done too since the second-to-last demo movie issued was from the X-Box) and “Linux binaries will be available very soon after the PC game hits store shelves.”

OS X games will have to wait a bit, but “it’s definitely coming”. It’s just that “[m]ore remains to be done for the OSX version of DOOM 3 and that will take some time … we won’t release the OSX version until it’s just as polished as the PC version.”

Sensible, but probably disappointing to those that remember the first Quake 3 Test appeared only for the Macintosh because it was such a stable hardward platform. Guess they got all the kinks worked out in X-Box development.

Demo

The same article also mentions that, while a demo is definitely in the works, there is no fixed date for it. “We will release the demo as soon as it’s done, but this probably won’t happen until after the game has arrived on U.S. store shelves.” If you want to read anything into it, you can assume they’re already a good part done with the demo. If you’re a realist, you can assume that everyone at id is on vacation right now, after the gold release.

If, like me, your video card goes into a cold sweat just thinking about playing the DOOM 3 slide show for you, you’ll probably wait for the demo and see if it even has a prayer of running. Then you’ll go to a buddy’s house, see it running at High Quality mode and run out and kill your credit card anyway.

Todd Hollenshead’s .plan (Shack News) has the full announcement. It shows up in the US anywhere between the 3rd and 5th of August, Europe on Friday, August the 13th and … everywhere in Russia and Asia probably around August 1st or so.

PC Gamer has a review of one of the last few release candidates of Doom 3, and there are illegal copies of the text floating around, like this one: PCGamer Review − September 2004 Issue − 94% (NV News). id has stuck to their promise to make a better gaming experience and “those expecting a “classic” run and gun Doom gameplay, the biggest surprise may be just how substanial this game is.”

“If you try to blaze through any of these 28 missions you WILL be humiliated. Instead the only route to access is a slow and steady one, sticking to shadows, searching every nook and crany for health, ammo, and access keys, and generally advancing as methodical[l]y as you can. … But theres no need to worry that Doom 3 is as slow as splin[t]er cell − hardly a minute goes by without a furious exchange of hostiles with some manner demonic beastie”

So that’s the gameplay; the demos always hinted at a game that would deliver pants-shitting terror, and id apparently delivers a game in which you can “say goodbye to sanity for the next 20-odd hours” while you’re playing it. The single player mode sounds great and there is also multiplayer, including “[a]ll varieties of Deathmatch − 1v1, DM, Team DM, and Last Man Standing. There are 5 DM maps shipping with the game.”

The graphics are, as expected, just as good. Apparently, it’s nailed the photo-realistic look games have been reaching toward for years, where, in one scene, “[d]ust blew around the martian surface and the dull brown/red hue of the sand and the twisted metal of shredded structures all semed so perfectly plausible.” As you swing your standard flashlight around the empty base, “light floods through a room, swinging back and forth, shadows are cast perfectly; dust particles gently drift into the cone of the flash light, eerily visible”.

Carmack and Co got ahead of even the latest, richest gamer, by building an engine that has a “A higher level of quality and resolution is available, but the PC to run it well isnt”. Excellent. The sound engine and sound design are so good and so frightening that “[y]ou [are] basically subjecting yourself to a 20 hour car[d]iac episode. At times, death brought sweet, momentery respite from the fear drenched mayhem.” And once you get to Hell? The review describes it as a “balls shriveling nightmare netherworld”.

It is definitely time to upgrade.

But to what?

Let’s see … surround sound speakers seem to be a must. The sound engine is CPU-dependent, so no need to get a fancy card there. CPU should be at least 2GHz to play without hiccups, the higher the better. 512MB of RAM is perfect; 1GB is slightly better, but not really noticeable.

There are four graphics render paths within the DOOM3 engine itself. “ARB2 path is the “full package,” and is used for the R300+ and Geforce FX+ cards”. 128MB of VRAM should be good (textures are compressed in that case. 256MB doesn’t compress textures, but that shouldn’t make a noticeable difference. “a 500mb card is needed to run the game in Ultra Quality mode.” That’s the mode that no PC can actually run well yet.

There are two most amazing things about the Half-Life 2 case modification, Blackmesa HL² by piloux, documented here.

1. This is the first modification that I would consider a work of art. The attention to detail is really amazing. The person who made it must have really had a blast making it. They’re also bound to get a lot of attention at LAN parties.
2. This news was posted on Slashdot this morning. The page, including images, weighs in at 1.5MB. The server it’s on never faltered. Go to it now and it loads like lightning — it’s got more bandwidth than CNN. Kudos to Whabbit.

There are about 70 images of the case throughout the building process. The keyboard at the end indicates a somewhat abnormal attention span — but even the mouse is all rusty and crusty.

Version 3 of the Unreal Engine is in development. Tim Sweeney on UE3 (Beyond Unreal) has the latest info on it, direct from the lead developer/architect. He’s been involved since the first version which ran Unreal; the second version ran Unreal Tournament and UT2003. Since this engine is aimed at the kind of machine that will be “mainstream in 2006”, some of the numbers being tossed around are pretty formidable.

Sweeney offered no answer when asked how much space a game that uses “2k by 2k” everywhere might take up. “UT2004 required 5.5 gigs of hard drive”, so I guess the sky’s the limit. If you look at the rendered character to the left (in engine, as mentioned on the screenshot itself), you see the incredible detail afforded by the high-res textures. As far as the video card goes, the requirements are:

“DirectX 9 cards will be minimum spec, so any DirectX 9 shipping today will be capable of running our game, but probably at reduced detail. If you only have a 256 meg video card you will be running the game one step down, whereas if you have a video card with a gig of memory then you’ll be able to see the game at full detail.”

Only a 256 meg video card??!!?? Anyone else feeling old?

Unreal Engine 3 describes Epic’s new engine in detail, covering the advanced graphics, physics, animation and content creation technologies. The scene to the right had “[o]ver 100 million triangles of source content contribute to the normal maps which light this outdoor scene.” That’s in the editing tool — before generating the compressed game level output where “[w]ireframe reveals memory-efficient content comprising under 500,000 triangles.” A mere 500,000 polygons!

The engine’s completely written in C++, is Unicode-friendly, compiles for multiple platforms (consoles, etc.), has its own flexible storage format and includes script/level debuggers and much more. Check the licensing page and you’ll see it can be yours for $750,000 for a single-platform royalty-free license. Each additional platform is $100,000. But if you read through the engine specs again, you’ll find it’s probably worth it for larger game projects.

This engine is probably only available for development right now and may still be in development itself; as Sweeney said, they seem to be aiming for the 2006 market with it. Still, it really looks like Unreal already has their act together today.

If you dare step away from the video card world for more than a few months, you find, on returning, whole new vistas of acronyms, bandwidths, core-speeds, registers, pipelines and “technologies” waiting to eat you alive. The latest cards from the two top vendors, ATI and NVidia, are lining up to play tomorrow’s games at 1600x1200 with 8x Anti-aliasing and all effects on high. Quake 3 doesn’t even make a dent — assume it runs at 300FPS with everything on.

Radeon X800 XT

ATI’s latest card, which, though announced after NVidia’s should actually hit stores first, pushes “8.3G Pixels/sec”. Flagship Radeon X800 XT debuts (Ars Technica) gives more details:

“$499 price tag … 16 pipelines, a 520MHz core, and 256MB of DDR3 memory clocked at an effective 1.12GHz (550MHz GDDR-3)”

The ATI Radeon X800 XT Platinum Edition / PRO Review (Beyond3D) addresses my primary concerns with their “Cooling and Power” section. You’ll see why this is important when we take a look at Nvidia’s behemoth in the next section.

“The fan utilised proved itself to be a fairly silent solution on 9800 XT, as is the case here − however as part of the on-chip mobile technology utilisation, R420 features an on-die thermal probe which constantly monitors the temperature and different fan speed steppings are set in the BIOS to different temperature ranges, hence the fan will run at lower speeds when the chip is running cool and higher speeds when its hotter.”

The high-end version uses about “76W to 65W” on average, whereas the lower-end version (which is still ungodly powerful compared to most of our cards) uses about “58W and 49W” under the same conditions. The review itself contains an unbelievable amount of detail relating to how the card works with all of the aforementioned acronyms and buzzwords. It’s fast. It runs UT2004 at 1600x1200 with 4xAA (Anti-aliasing) at about 50 FPS.

NVidia GeForce 6800Ultra

NVidia has a new card too and it’s a monster. GeForce 6800Ultra Preview ([H]ard|OCP). As you can see on the left, the installed part is so huge, it completely overlaps one whole PCI slot, making it unusable. In fact, the article notes that “[i]t is our suggestion that you give any video card as much room to breathe as possible.”, meaning you should probably leave one more open, meaning you’ve got a powerhouse card that takes up 3 slots.

The last card they came out with was quickly named the “GeForceFX 5800Ultra dustbuster” because of the unbelievable noise it made. This one is, apparently quieter:

“When you first start your computer before the driver loads the fan will run at top speed, and the noise from that is noticeable. It is loud, but not as loud as the 5800Ultra. However, once you install your driver and the card is running in Windows the fan runs at a lower RPM. … In normal operational mode the fan is no louder than any normal 80mm case fan. It is definitely not discernable with it inside your case with the case covers on.”

I’m thinking that my tolerance for noise might not be as high, but we may just have to take the reviewer’s word for it. “not discernable [sic]” is quite subjective, I think, especially when you’re wearing headphones.

As with the previous review (in which this card is actually used in comparisons), there are charts and benchmarks all showing you that this care is really, really fast. Again, there are image quality comparisons, this time to the 9800 Radeon series (the precursor to the X800 XT), in which the 6800Ultra comes up short. NVidia seems to have the same image quality problem that Voodoo cards used to — before they went out of business.

Shader Models and shamming

In noted in another article, Shader Model 3.0 ([H]arc|OCP), NVidia has released screenshots in which they claim to “show the advantages of using Shader Model 3.0 in a game”. The screenshots are very convincing, showing the image quality that their card is capable of. However, they compare SM1.1 to SM3.0. ATI’s offerings use SM2.0, but do not currently support SM3.0. Hence, NVidia’s screenshots.

Unfortunately for NVidia, while SM3.0 does offer advantages over SM2.0, there are no games on the horizon (within the next year) that will take advantage of it. As the article says: “It is our opinion that SM2.0 is technology we are likely to see have the greatest quality impact on our gaming experiences through this next year”. Even worse, though, are the statements made by the lead developer of Far Cry, the game that NVidia used to make their screenshots. It seems that, while Far Cry takes big advantage of SM2.0 features to look as good as it does,

“In current engine there are no visible difference between PS2.0 and PS3.0. … In current generation engine quality of PS3.0 is almost the same as PS2.0. PS3.0 is used for performance optimization purposes.”

Looks like NVidia is so concerned about image quality, that they’re already starting with their fake benchmarks and false press releases again. At least some things never change.

Which one?

Radeon X800XT-PE and X800Pro Review ([H]ard|OCP) offers benchmarks which also compare to the latest from NVidia, including image quality comparisons. To the right, you see a striking example of the shortcuts Nvidia has always taken in order to get better performance. In this case, ATI’s card offers both superior quality and performance.

The [H]ard|OCP review’s emphasis on quality evidenced by the chart they include, showing the highest “IQ [Image Quality] settings that we found playable on our test system.”

So it looks like all of the toughest, most photorealistic, graphics-intensive games available today are playable with super-high anti-aliasing enabled and all visual options maxed out. The top ATI part manages all this at 1600x1200 for all but one of these games. The economy ATI card and the NVidia card do it all at 1280x1024. Unbelievable. Methinks John Carmack has, once again, predicted and timed the graphics card market just right.

Here’s looking forward to Half-Life 2 and Doom III.

Finally, here’s a roundup of the info I’m keeping in mind for my next card:

• NVidia’s card uses at best 2 slots (1 AGP/1 PCI), ATI’s fits in like a normal card.
• NVidia has a bad reputation for noise and power consumption.
• ATI makes prettier graphics
• ATI is faster
• ATI’s Anti-aliasing is better and faster
• Both sets of drivers are probably quite stable
• Both cards cost $499
• Both have a cheaper variant that is also ridiculously fast
• Both will probably be cheaper in 6 months

First Person Shooter − Under 100KBs of Code (SlashDot) delivers exactly that — including all textures and sounds. The screenshot below shows you that it’s a relatively normal looking 3d shooter, with a bit of a Quake2 feel to it, but with “Doom3-style graphics, that means full Phong lighting model with various light sources and normal mapping everywhere, and of course stencil based shadows”.

.kkrieger Screenshot

How do they do it? There’s a comment Explanations! by one of the developers, where he explains.

“All content, that is: the textures, the models, the map and part of the animations is generated procedurally. The basic concept is a modular graphics synthesizer which only stores the steps needed to opbain a certain image or mesh with their parameters. [sic − all of them]”

There’s a bunch more information at The Product, including downloads for the tools, so you can build your own 100KB game/rendered music video/whatever. You can even download other tiny products. Download in a heartbeat, double-click the EXE in the zip and you’re greeted by their stark white progress bar on a black background. Depending on your machine, it will take a little while to generate all the textures, models, sounds and animations needed to render the movie or game.

If techno/house makes your scream, avoid the downloads or turn the sound off. Hey, it’s 65k and it’s German/Austrian — what kind of music did you expect?

Is this seriously a screenshot?

Wow. Coming in April. Time to upgrade.

John Carmack, of id Software, has updated his .plan file (cached copy at earthli.com) with his impressions of the NV30 from NVidia. He compares it to the current king of the video card market, the ATI R300 part and describes how they handle his DOOM3 engine.

“At the moment, the NV30 is slightly faster on most scenes in Doom than the R300, but I can still find some scenes where the R300 pulls a little bit ahead.”

He mentions that there are several code-paths or renderers available for the DOOM engine, two of which (ARB and ARB2) are codes without card-specific functions. The ATI part will run the high-quality non-specific version almost as fast as it’s native implementation, but “[t]he NV30 runs the ARB2 path MUCH slower than the NV30 path. Half the speed at the moment”, so he can’t really do what he calls an “apples-to-apples comparison”.

The new part from NVidia is somewhat physically intrusive as well, as he mentions that “[t]hey take up two slots, and when the cooling fan fires up they are VERY LOUD.”

This ARB support is a new standard API for loading and executing programs on video cards with hardware pixel shader/T&L support. Carmack is making use of only standardized functions in as many pipelines as possible to avoid API fragmentation:

“Doom has dropped support for vendor-specific vertex programs (NV_vertex_program and EXT_vertex_shader), in favor of using ARB_vertex_program for all rendering paths.”

Worry not, Carmack is already looking to the future, beyond the DOOM engine (he’s been working on it for 2 years now) when he says: “[i]t is going to require fairly deep, non-backwards-compatible modifications to an engine to take real advantage of the new features…”. I take this to mean that the current DOOM engine required feature set is properly frozen, but tweaking to accomodate new hardware (and improve performance) is continuing. Completely new fuctionality exposed by the new cards — more programmability through the new APIs or larger program sizes (graphics card programs) — will only be explored in the next generation of the engine.

However, he has managed to work in some improvements over the existing renderer for those that purchase the horsepower: “[p]er-pixel environment mapping, rather than per-vertex. This fixes a pet-peeve of mine, which is large panes of environment mapped glass that aren’t tessellated enough, giving that awful warping-around-the-triangulation effect as you move past them” and “[l]ight and view vectors normalized with math, rather than a cube map … [which] give[s] you … a perfectly smooth specular highlight, instead of the pixelish blob that we get on older generations of cards.”

Both somewhat minor wins for a faster-paced game, but in a slower, environment-based game, any visual improvement is a good one if it enhances immersion. The next generation of cards will only improve internal data precision, with “[f]loating point framebuffers and complex fragment shaders” becoming very important, allowing “much better volumetric effects, like volumetric illumination of fogged areas with shadows” — again, something that will increase the movie-like feel to a scene as interaction of light, air and shadow improves.

John Carmack gave a long presentation (almost 3 hours) at Quakecon 2002 which covered the current Doom technology and the future of gaming as he sees it. Gamespy has the complete coverage in QuakeCon 2002 − John Carmack Speaks. The future of graphics technology involves rendering effects like lens flares and specular highlighting in a more realistic manner; not that they look more realistic necessarily (though they will), but moving more rendering into the ‘standard’ pipeline instead of handling so many effects as “special-case scenario[s]”.

“He believes that the most significant benefit to this type of advancement is that certain effects will become less of a special-case scenario. When rendering things “the right way,” one does not need to worry about covering up for an elaborate fake. The effect actually exists, and can be used in the world as-is, without needing to design a level or area around that effect. ”

This includes a “jump to 64-bit color” as well. The 64-bit color is for internal precision when rendering, so calculations are done correctly, then scaled to the output device (along dimensional or color axis) only at the very end. “[t]here’s still a lot of benefit to gain by doing all of the intermediate calculations the way you really should do it.”

Once again, we see that John Carmack is truly an engineer who sees that quality is a long-term goal directly affected by short-term decisions. He sees that game engines should be able to graphically render effects as they physically occur in order to provide the best possible model. If this is too slow, those become the ‘special case’ effects only at the end, in order to ensure proper performance on current hardware. Concessions to performance should be made only after correctness can be ensured.

One gets this feeling even more when he discusses features that he’d like to put in, like a much better ambient lighting system. “the DOOM 3 engine does contain some basic ambient light sources that will affect all of the bump-maps on the various models without giving a highly directed look to them, but what he feels would be optimal is to affect each bump on a per-pixel level, by having “each bump look up in a cube map.”” Unfortunately, this type of change, while basically done from his technology’s end, would require far too much reworking of existing artwork and content. It will have to wait, but you can tell it bothers him to have to ‘settle’ for a ‘hacked’ ambient lighting model, especially since any card after the GeForce-1 would be able to render ambient lighting correctly with no difficulties. In this case, he (and we) will just have to wait, as he has also said, “for the most part my work on Doom, the significant contribution to it, is done. Set in stone. The renderer decisions are all made.”

This desire for correctness is a recurrent theme in Carmack’s work and his designs. His approach to rendering and graphics is extremely abstracted and takes its goals directly from the real world. He sets his standards incredibly high.

“The way you should be calculating all graphics… the way it ought to be done is: you’re basically counting out photons that are, you know, imprinted on a surface. Lights spray out a whole lot of photons, that are collected on surfaces. … So what we want to do is do all of this calculation the right way”

The Doom III technology is actually based on “GeForce1-level hardware” as the baseline feature set, whereas “id will be basing their next technology on the features just now becoming available with ATI’s Radeon 9700 card, and NVidia’s upcoming NV30 chipset.” Though Carmack has repeatedly pushed hardware to its limits, he still falls a little short of predicting exactly how far and fast it will grow. These days, different cards support different features, so “[t]his has led him to keeping track of roughly half a dozen separate backends, each optimized for a particular card.”

This is actually a very sound design decision, which allows the rest of the engine the flexibility of interfacing with a standard API, but still allows the most performance to be squeezed from an individual card. “Carmack has chosen to have all of the basic rendering and most special effects follow one path, with lighting effects being the primary area of divergence.” This means that any card will be supported, but that popular ones with special features will have their own renderer.

“This means that Doom3 does not take complete advantage of every single feature found on every single card on the market, but that it will look very close to the same on a wide variety of cards. Carmack stressed that he prefers to have a title which appears the same on a variety of systems, and requires lower resolutions for increased framerates, than a title which turns off effects to achieve higher framerates.”

He’s already written an OpenGL 2.0 compliant renderer and continues to stress that developers should move to a higher-level shading language rather than specifying effects in card-specific APIs. It’s like the “transition from assembly language to C and other languages”; Sometimes a human can hand-code better than the compiler can create assembler, but not many humans know how, and the process is incredibly hard to debug and is very error prone. More often than not, developers won’t implement an effect because it’s too hard. Developers should be able to “specify what we want to have done, and as it needs to, [the card] cut[s] it down into multiple passes for different cases”. That gives the developer freedom to develop and design effects and the individual card vendors implement those directives as quickly and well as their cards can do it. It’s a very natural progression that’s happened many times before in programming. The transition to object-oriented programming was (and still is) similarly accompanied by complaints of performance. Carmack’s a visionary and doesn’t have such hangups; he realizes what will create the faster, most maintainable, extendable code overall:

“I am on the record as saying that the next game engine that I work on, after Doom here, is going to be written in a high level shading language. There’s just no doubt about it. There’s not going to be custom, per-card assembly specifications going on there. It’s going to be in a high-level language, and the drivers are just going to have to deal with it there.”

The future of video cards will only get more advanced (“Carmack points out that the ATI Radeon 9700 is over a hundred times faster and more powerful than ATI’s initial Radeon offering”) and developers should be happy not to have to figure out for themselves how to get the best performance out of it. Developers should be able to focus on doing the math for their environments to feel realistic, then specifying the commands needed to render it. The card can translate this into a scene on the display device. “Put quite simply, the 50-100 passes that Doom3 uses to render a scene will, within a short period of time, seem minimal in comparison to what graphics cards will be capable of.” Considering the videos from Doom, this is fantastic news for us.

As for Carmack? He’s getting ready to move back into research (lucky dog). He says, “I’m getting a little tempted now to start peeling off and working on some next-generation technology generations to research some of the things.”

Gamespy has another article called DOOM 3: Behind the Horror, covering Doom III’s new features. Doom III promises to loosen up first-person gaming from its often amusement-park ride feel. By that I mean that sometimes you feel like you’re in a cart on a track, where you can’t investigate this door because it takes you off the plot, or that light can’t be shot because you need to see the level and the shadows cast from it are pre-rendered. Doom III promises to remove most of those rocket-resistant light bulbs.

“Every light is dynamic and can be turned on and turned off,“ [Tim Willits] explained. After shooting out a few of the lights, noticeably darkening the area, Willits explained one way the new tech will influence gameplay. “The player could blast out too many lights and will have a hard time seeing that monster that’s coming sneaking up behind them.”

Many good design decisions have gone into the game, many taken because it makes the development of their own games easier, but also because id licenses their technology to so many companies. An id engine-based game generally has a mod/map-making community as well. “[W]e decided we wanted to make things simple on ourselves and also simpler for the people out there who play our games and also modify them once they’re released.”

To this end, even the HUD graphics overlays that form the user’s controls and interaction with a world have been generalized. “[T]he new GUI system — it’s an HTML-like system that will both drive features in-game, as well as handle the game’s menus” is very flexible and even allows “picture-in- picture effect[s]”, pulling rendered views from the game engine. Presumably the in-game, perspective-rendered videos that Graeme Devine added to Team Arena will also be available in the Doom III engine.

With the new fantastic, dynamic-lighting system in the engine, id can finally step up the interactivity of the rest of the physics engine, since movable objects now won’t cost any more to render than static ones (outside of calculating trajectories).

“First, Tim showed the physics system by shooting some boxes off a shelf — the boxes would react differently depending on where they were shot. Next, Tim shot the side of a lighting fixture, causing it to swing back and forth, and subsequently cast moving shadows on the zombie below.”

For those actually creating maps and mods, id has a welcome change that has been mentioned here before: no compile times and an in-game editor right out of the box: “you can just run the game, bring the console down, type ‘editor’, and up pops the editor we use to build all the worlds.” Any user can modify the levels, watching the changes happen in real-time. Robert Duffy demonstrated this: “lights were placed and moved around the level, with the appropriate reactions in the environment”. To be fair, Doom isn’t the first game to do this. The Serious Engine by Croteam, which looks great in its own right, had real-time 3-D level editing first, at least packaged as a single tool. Of course, they don’t support dynamic lighting everywhere, but the editor worked very smoothly in exactly this fashion.

Unlike most other id games, the Doom III engine contains a powerful, flexible scripting system, which should quiet many mod and map making critics who complain that they usually only put in a couple of event triggers that can’t even be combined (like putting a pendulum on a rotating disk, for example). Models are as much a part of the scene as any environment, character or world object, as they’re all treated the same now. This homogeneity in the “real-time lighting simplifies things with the scripting”. There’s no special cases for models anymore, so, “[i]nstead of having to define lighting for scripted events, the shadows cast off the moving parts are all created naturally in real-time”. Similarly, the object’s collision interacts autmotically with other game objects without any work from the designer. In a demo “involving a large mechanical loading arm”, scripter Matt Hooper:

“was able to shoot through the holes of the arm, and by zooming through, we could see the bullet holes way off behind the arm. So even though the arm was part of a scripted event, and the spin could be adjusted so the holes could have ended up in any position, the engine still treats it just the way you’d expect in the real world — if there’s a hole, you can shoot through it.”

With the Doom Engine, there’s not nearly as many tricks needed by level designers any more. It does what the real world does; it just works. The sound system is also very interactive, emphasizing that they want to “make it easy for people with studio background to add sound to the world”. The sounds interact with moving objects as well, dimming with distance and muffling from interference:

“[S]peakers could also be attached to scripted events, such as the spinning id logo in the beginning of the DOOM 3 demo. By attaching a virtual speaker to one side of the logo, you’d be able to “hear” the logo spinning as well as seeing it.”

Coming to a PC near you in Spring 2003.

ShackNews is reporting that John Carmack has updated his .plan file recently in Carmack On New Cards, Rendering − the actual .plan file is here − Carmack’s 2002/06/28 .plan.

If the other .plan link is broken, then you can get an archived copy here − Carmack’s 2002/06/28 .plan

His latest two updates concern the Matrox Parhelia. The first update pretty much trashes the card, calling it “really disappointing for the first 256 bit DDR card” and that the “[a]nti aliasing features are nice, but it isn’t all that fast in minimum feature mode”, which means no one is going to use it because it gives too much of a performance hit.

His second update reneges much of what he said in the first. In fact, he praised Matrox’s rollout of the new card for providing such good testing hardware and solid drivers right out of the gate:

“I was duly impressed when the P10 [Parhelia] just popped right up with full functional support for both the fallback ARB_ extension path (without specular highlights), and the NV10 NVidia register combiners path. … this is the best showing from a new board from any company other than Nvidia.”

Given that, he decided to “go ahead and write a new back end that would let the card do the entire Doom interaction rendering in a single pass”. That must have been about a good two or three hours of Mr. Carmack’s time, I bet. In addition, rather than using the NVidia-compatible extensions that the Parhelia supports and making use of the existing NVidia rendering pipelines, “[Carmack] decided to try using the prototype OpenGL 2.0 extensions they provide.”

This brought on a discussion of OpenGL 2.0, in which he mentions that “[he is] now committed to supporting an OpenGL 2.0 renderer for Doom…”, even though “[a] GL2 driver won’t give any theoretical advantage over the current back ends optimized for cards with 7+ texture capability”. He sees it more as a move away from “lower level coding practices” and toward “C-like graphics languages”. In addition, he “strongly urges [vendors] to implement GL2 instead of proprietary extensions”.

Thus, the decision to write a rendering pipeline for OpenGL 2.0. Any card that supports this non-proprietary specification should be able to run Doom III in all of its glory with no extra work. As mentioned above, since the other optimized pipelines for existing cards are already written, it won’t make any difference for Doom III whether he writes the GL2 pipeline or not, but from an open-standards and engineering perspective, it makes a lot of sense.

He sees the trend as moving toward using these C-like languages for all graphics-chip programming work instead of “current interfaces we are using”. For most compatibility, he suggests rallying around OpenGL 2.0 as the language of choice for this, as opposed to the CG language recently proposed by NVidia as an alternative.

“It won’t be too long before all real work is done in one of these, and developers that stick with the lower level interfaces will be regarded like people that write all-assembly PC applications today.”

AnandTech has a whole article reviewing the Matrox Parhelia, Matrox’s Parhelia − A Performance Paradox. Once again, as with the Radeon 8500, there is a card that, on paper, should beat the top-of-the-line NVidia cards. And, once again, the actual hardware fails to reach expectations.

As with ATI, Matrox focuses on visual quality above speed. “There’s no doubt that the analog output quality of Parhelia is excellent, it’s definitely the best we’ve seen to date.” However, the sacrifice for that visual quality is still too high, since a sufficient frame-rate can’t be sustained with full high-quality rendering enabled. The best feature of the card is Fragment Anti-Aliasing, which is Matrox’s new algorithm. With anti-aliasing enabled, the Matrox card is slower than the comparable GeForce4 Ti 4600, “[b]ut rest assured that Matrox’s Fragment Anti-Aliasing looked nothing short of amazing.” At 1024 × 768, it was a little faster than the GeForce4, but both were in the mid-40s for frame rate.

Beyond3D has some extremely detailed questions with John Carmack in John Carmack on DOOM3 rendering. Of interest is the mention that “[t]he game characters are between 2000 and 6000 polygons”, which makes sense, given the massive number of rendering passes in Doom 3. Since all shading and lighting is done in real-time, the number of polygons becomes a lot more important.

Other engines, like the Unreal engine, boast much higher character-polygon counts, but they use only limited dynamic lighting. Of course, they generate nice shadows for the models, but the shadows won’t interact nearly as well with the rest of the environment and the other models as the Doom engine promises to do.

With the heavy use of bump-mapping to generate lifelike, photorealistic detail with extra polygons, Doom’s use of dynamic lighting will still produce much more lifelike scenes, despite only a small increase in number of polygons per model. With the new, fully real-time rendering system, more video horsepower will always help. He mentions that “[t]otal number of render passes is (greatly) influenced by the number of per pixel lights used.” and, since the engine’s render path is far more straightforward that it used to be, the expected speed of a scene is somewhat easily predictable. The speed will be proportional to the number of lights used, excepting some unusual cases, like “making very jagged models with lots of little polygonal points”, which would slow things down inordinately as the engine rendered a lot of little shadows on a lot of little surfaces.

This guy is from another place altogether…

John Carmack Interview on Voodoo Extreme

The big news from E3 (Electronic Entertainment Expo) is the first official showing of Doom III. The video they are showing there is available for download includes in-game/engine footage and interviews with members of id, along with a history of Doom. The in-game footage is amazing and pretty photo-realistic. The models breathe, roar and move in unbelievable detail.

The first scene is a monster pawing around a dead zombie in a bathroom. It drools, the zombie is batted around realistically, a lamp on the ceiling is swinging, so the light source and shadows are all moving…you can even see it pull the zombie apart. The amount of detail that bump-mapping brings to the skins is amazing. Veins and muscles all appear to stand out.

The video is available in QuickTime format, which runs 11 minutes, has a lower framerate and is 97MB. That one is available from ShackNews from the article: Doom III Stuff. The other video is a ‘High Quality’ version and weighs in at 90MB in WMV format (Windows Media Player movie). The framerate is much higher and if you have Windows Media Player, you should download this one. You can download that video from ShackNews as well, from the article: DOOM3 Legacy Vid High-Res.

For the faint of heart (90MB videos are big), 79 screenshots captured from the videos are available at Custom Q3. The shots show off nice stills from the movie and give you a pretty good impression of the enormous change that the Doom engine represents. It’s going to be like playing a movie.

Doom III will depart from standard id games in another important way. There are plans for a plot. The game engine’s amazing sound and video capabilities (see Doom Technology) allows the artists and designers to create a terrifying atmosphere and focus on the environment’s role in the game, rather than using many monsters (Doom) or other players (Quake III) to create fear. GameSpy has a full article, DOOM III: The Very First Look, with more details.

Watching the video (see Doom III E3 videos), id looks like they’re returning to monsters in the style of the fiend (see right) from Quake I, with “… many seemingly non-human creatures present — many of whom attacked with quick, lunging attacks, and were all modeled and animated exquisitely.”

Further good news is that the “The game has already been storyboarded from start to finish” and “id has enlisted the services of science fiction writer Matthew Costello to pen the game story and dialogue. Costello is no rookie to this sort of thing — you may recognize him as the writer of the 7th Guest and The 11th Hour games.” Graeme Devine, now a programmer and project manager at id (he wrote the new sound engine) was a founder of Trilobyte, who made the 7th Guest and 11th Hour. Furthermore:

“You can expect the plot to be furthered along through a combination of in-game scripted scenes and in-engine cinematics, but nothing rendered — everything will happen inside the new engine.”

The dynamic lighting enhances the storytelling significantly, helping set the mood:

“… there’s a scene where a scientist sits in front of a computer, which casts a glow upon his face in addition to the other light already present in the room. As the computer display changes, so does the overall lighting, similar to how a television will light up a dark room. Shadows also play a big role — at a number of points in the demo, you’d see a shadow approach before the actual enemy, giving you a chance to react before being attacked.”

The new animation system is skeletal with blended animations, with “lots of animations for eyes and mouths, as well as syncing, and they’ve also developed an entirely new scripting system as well as a “fairly robust” camera system.”

The new physics engine is also extremely robust. In the demo, “one scene shows a character falling down a flight of steps, reacting to each step along the way”, so maybe the days of watching bodies float off the edge of a set of stairs are gone. “another [scene] shows boxes getting knocked off a shelf and bouncing to the floor in an entirely realistic manner.”

The interactivity with the environment doesn’t just stop at larger objects, like stairs, boxes, tables, etc.:

“One of the more interesting technical innovations we saw is a new GUI system that allows for Flash-like animation on any surface, and can be completely interactive. Unlike most current games, where a keypad or switch might be portrayed with a few simple textures, these items can now be presented with much more detail.”

Doom III is going to be ground-breaking in many different ways. First of all, it doesn’t seem that it will attempt to replicate the super-fast gameplay style of the original Doom. In There’s no ignoring Doom III on MSNBC, John Carmack says “If you … have everybody running around at 100 miles an hour, you would lose the immersion.” It will be “about the fear, the scary, extraordinary, unknown environment and not knowing what can happen.”

There are many departures for id from their standard ‘formula’ of game construction. Attention to content is paramount in this game, with quite high hardware requirements in order to assure that artists don’t have to dumb down their models or textures. From a post on Slashdot, High end hardware reasoning, again by John Carmack:

“The comfortable minimum performance level on this class of hardware is determined by what the artists and level designers produce. It would be possible to carefully craft a DOOM engine game that ran at good speed on an original SDR GF1, but it would cramp the artistic freedom of the designers a lot as they worried more about performance than aesthetics and gameplay.”

The renderer for Doom does all light-source and shadow rendering in real-time, with no precompiled lighting information. In addition, environments will be much more interactive than they have been in id games in the past. “With the new Doom, there is no separation [between static and dynamic elements]”, so gone are those unshootable lights and unmovable tables. The lack of static lighting information allows the engine to use movable objects throughout, including moving or rotating lights, breakable walls, etc. In a demo, “… a buffalo-sized hound from Hell nearly broke through a solid wall trying to get at the player.”

As Tim Willits says, “You can move around in it and knock boxes over and break stuff. It’s ‘in your face’ immersive.””. Carmack chimes in that “You’re not going to know where there is a safe piece of a wall.” That is, any portion of the environment could move, break or give way, with completely believable physics (look at the railing being squashed by the hellhound near the end of the video, or in the screenshot to the right).

Carmack describes the new engine in Master of Doom: Carmack Speaks on Gamers.com. The renderer also uses “discrete, projected lights, full-time bump mapping, and global unification of light-surface interaction” to make a more believable world. This means that creatures have the appearance of much more detail (bump-mapping) and all polygons in the scene are treated the same, with respect to occlusion, shadowing, transparency and lighting (global unification or light-surface interaction). He goes on “Once you experience the consistency of the Doom world, other games start to feel more like puppet shows.” With Doom, all surfaces are dynamically lit and there is no distinction made between static (world or environment) and dynamic (model or player/creature/entity) surfaces so that “… lots of effects with light and shadow that people have always wanted to see in games … just work naturally now, with no special hacks.”

The sound engine is “Dolby digital, six-channel” (“full dynamic 5.1 channel sound mixing, and multichannel playback of studio sounds”) and the physics engine has been completely rewritten. In fact, most of the engine has been rewritten. An article at GameSpy, Q & A with John Carmack quotes Carmack as saying, “after our move to C++, there is very little code remaining from the Q3 codebase at this point.”

And, he goes on to explain his approach to programming (which is why you just can’t help admiring the guy):

“There are some tough judgement calls to be made during development about whether something is an elegant extension of our chosen technical paradigm, or if it is unjustified work. Having the inclination and authority to just say “no” to feature requests has been an important aspect to being able to write quality code. Too many programmers agree to random feature requests without thoroughly considering the impacts. I try to err on the side of elegance in implementation, rather than feature coverage.”

For implementers of mods and levels, the Doom engine promises big changes as well. Whereas the current game engines require a long precompiling stage before viewing a level, Doom’s “…tools are now built in to the game … The game also uses the same map file that the editor uses, so the original source data can be opened up with any copy of the game.” So, while the amazing renderer means artists will spend more time making models and level textures that take advantage of the engine, level builders will have a much faster turnaround for testing.

“What this means is that you’ll be able to walk around in a level, press a key, bring up the editor, place some lights, and go back to the level.”

Don’t think the choice of an NVidia card is cut-and-dried yet. Sure, the latest Radeon offerings from ATI are slightly slower than the Nvidia cards, but what about the visual quality? Speaking of visual quality, an old hand at making slow, nice-looking graphics has a new technology, called Parhelia. The Matrox site has pages of information, with screenshots and vidoes galore. They hope to lose the “slow” reputation with this one.

Ars Technica has an article, Aiming High at Matrox with some details about the new card:

1. 80 million transistor (0.15 micron process) 512-bit GPU
2. 256-bit DDR memory interface with up to 256MB DDR unified frame buffer
3. AGP 8x with Fast Writes
4. OpenGL 1.3 and DX 8.1 compliant
5. Fourth Generation DualHead Technology (400MHz 10-bit RAMDACs up to 2048x1536 @32bpp)
6. Support for 3rd RGB output (3 display desktop up to 3840x1024 @32bpp)
7. Quad Vertex Shader Array, Quad texturing per pixel
8. 36-Stage Shader Array
9. 64 Super Sample Texture Filtering
10. Hardware Displacement Mapping
11. 16x Fragment Antialiasing
12. 10-bit DVD playback, filtering, scaling and output

Hardware Zone has even more in a very thorough run-down in Matrox Parhelia-512 : The Technology. The article mentions that they have demonstration versions, but availability is still a ways off, with prices expected to be in the sub $400 range for a 128MB DDR card.

That’s a lot of stuff for one card, with 3-monitor support at insane resolutions, DVD hardware playback, 4 vertex shaders per pixel, each with 36 stages, and 16x Antialiasing. Whew! If Matrox can deliver, then this could give NVidia a run for their money, since NVidia always wins the speed crown, but can be a bit slower in making an all-in-one card that is capable of rendering as nicely as ATI or Matrox.

When you’re playing the latest and greatest game at 1280 × 1024 at 75FPS, who cares if the NVidia card will do 90FPS? Still, the jury is out until Mr. Carmack weighs in with his opinion of the new technology. Here’s hoping they’ve made their shader language compatible with OpenGL and/or DirectX.

The latest in a long line of great Q3-engine games arrived in stores almost 2 months ago. If you’re looking for a good game to get you ready for Attack of the Clones, which debuts on May 16th, this is it. Reviews have been very favorable, with only some of the more bizarre and pointless puzzles drawing criticism. The GameSpot U.K.‘s review:

“However, Jedi Outcast soon transforms from a typical first-person shooter to an exceptional Star Wars action game that contains some of the best combat sequences since Half-Life, the most distinctive control mechanics since Max Payne, and the most involving plot in a Star Wars game since Jedi Knight. … There’s nothing quite like walking into a room of storm troopers and sending them all crashing to the floor with a force-filled flick of your wrist, or fending off a cantina full of pistol-wielding drunks with nothing other than your lightsabre.”

The Adrenaline Vault also has an excellent review, which is very well-written and nicely laid out. Considering how much time it spends discussing the lightsaber, it seems that you can pretty much play the whole game with just your lightsaber and force powers, which is very cool.

If you do get the game, make sure to patch it to version 1.03 to because it has “four new Duel levels, fixes several multiplayer and singleplayer issues, and adds EAX 3.0 and OpenAL support” (ShackNews | LucasArts).

If you’d like to try before you buy, ShackNews reports that Jedi Knight 2 Demo [Coming] Soon:

“LucasArts sends word that they’ll be releasing a demo of Jedi Knight 2: Jedi Outcast this Friday. It will contain a level not found in the retail version of the game, though this level was shown at E3 2001. Five weapons will be available in the map, as will various Force powers.”

Since the game is based on the Quake III Team Arena Engine and was developed by Raven, a veteran of the id engines, the graphics are amazing, the multiplayer is top-notch and you can create your own maps and levels. JediKnightII.Net has instructions for creating levels in QERadiant, the official editor for the Quake III Team Arena Engine.

id Software Partners With Activision on DOOM III™ at Yahoo Finance News officially announces DOOM III, the latest and greatest game from id Software. The game will be shown at E3 (Electronics Entertainment Expo), but probably to a limited audience. Still, we can hope that a video emerges soon after.

“DOOM III will change what people expect to see and experience in a PC game,“ said Todd Hollenshead, CEO, id Software. “We couldn’t be more psyched about DOOM III, and are thrilled to debut the title at E3 with Activision. Get ready to be terrified.”

The game is based on the latest game engine, appropriately named the Doom Engine, built by John Carmack (Graphics/Networking), Jim Dose (Scripting/Animation), Graeme Devine (Sound/MacOS) and Robert Duffy (Tools/Editing). The game will support Windows 98/2000/XP, MacOS X and Linux. It will be optimized for video cards with onboard Pixel Shaders/Vertex Shaders, but will still run on GeForce2/Radeon cards, though not nearly as realistically. Apparently, all portions of the Quake III Arena engine have been rewritten for this new engine. The sound was redesigned/rewritten by Graeme Devine; the scripting/animation/object structure by Jim Dose; and, of course, a brand-new renderer from John Carmack. Even the content creation is new, with the tedious vis/lighting stages removed and maps loadable directly from text files because the lighting model is completely dynamic now. The turnaround time for editing changes promises to be orders of magnitude faster.

For more information, go to the id software home page, which has, believe it or not, actually been redesigned for the first time since Quake II.