Scop’s Blog

I was standing next to a Gears of War poster the other day (one of many), and it had Emergence Day written on it:  11.12.2006.  I found it hard to believe that the game was nearly three years old already.  It has even been eight months since Gears 2 came out, now.

Oddly for me, it got me thinking about how graphics have evolved in the past few years, or, rather, how they have not.

Consider this.  2006 saw the release of games like Oblivion, Tomb Raider:  Legend, Dead Rising, and Splinter Cell:  Double Agent.  In 2008, we had games like Fallout 3, GTA IV, Ninja Gaiden II, Metal Gear Solid IV, and Soulcalibur IV (yes, Firefox, one word).

While some of you might argue that the jump from Tomb Raider:  Legend to Underworld is significant, consider that the span between 1996 and 1998 was enough to get us from Quake to Half-Life.  For you visual learners:

(note:  None of these pictures are mine!  All hosting sites are shown in the tool tips!  Formatting in this post is going to be a nightmare!  I can’t stop exclaiming!)

From this...

...to this.

Compare this to the four year (and significant hardware upgrade) between Ninja Gaiden I and II:

Xbox, 2004

Xbox 360, 2008

Yes, I am fully aware that the second image is sharper, and the lighting is more dynamic, but for Pac-man’s sake (hmm, no?) it is a difference of four years and a console generation!  Those were my senior years in high school and college.  Chew on that.

I am not saying that game artists have gotten lazy.  Were that the case, I would be scared to go into work tomorrow (not that any of them read this site).  No, I am saying gaming has reached a point where even sandbox games can render damn good character models, so maybe it is time to start looking in different directions to squeeze more out of our rigs.

I would be foolish to say that we have reached the pinnacle of computer generated imagery, even in the case of on-the-fly rendering (as opposed to pre-rendered movies).  No, if you hear me out, you will find that I am suggesting something of a paradigm shift that would change every aspect of future games.  Not least the graphical side.

Physics.

Again, I would be foolish to say that no work has been done to make our games mimic the world around us.  I remember the first time I saw an enemy ragdoll in Max Payne II.  He was perched atop a painter’s platform, and needless to say, it was a long way down.  Havok changed the way we game, and PhysX cards (dedicated physics processors) made a splash back in 2007 or so.

Yet if engine developers started tackling physics with the same vigor that put Valve’s Source Engine on the tip of everyone’s tongue back in 2004  I do not think it would be long before we saw a resurgence of that hunger.  The need for a playbox.

I remember watching a Half-Life 2 video and practically drooling.  It was just a tech demo.  It contained things like dynamic in-game cameras (Hello, G-man), objects that floated only if they were buoyant, which sank under the weight of non-buoyant objects, and a giant pachinko board.  I loved it.

Hell, I remember being in a game shop back in ‘96, even, and listening to a couple of guys talk about Duke Nukem 3d and Quake.  Even then, the reason why they chose the latter over the former was how the player would fly across the room when he got hit by a rocket.  That little extra bit of realism sold them over all of the cheesy one-liners and boobs.  (‘Cause it sure wasn’t the polygons)

And where should they take it?  How about starting with the environment?  Consider this fan-made video from Crysis.  It is not just that the game can calculate collisions for 3,000 individual objects, but look at how he arranges them into simple shapes half way through, like the house and the chair.  Imagine if every object in a game world was composed with such granularity, with that much potential for interaction.

Instead of objects that follow the same fracture paths, like in the Source engine (see this), imagine a surface that can warp when a bullet hits it.  Bones that snap exactly where the force is applied.  Fabric that shears when cut and stretches when pulled.

There’s a scene in the pilot of the USA show, Burn Notice that I absolutely love.  The protagonist, Michael, brings a hostile drug dealer to the door, then shoots his leg through the wall, between previously located studs, using a homemade silencer.  Then, as the dealer is bleeding in the foyer, holding a gun and waiting for Michael to come in the front door, Michael goes around back to a weak exterior wall, breaks in through it, and sneaks up on him.  This scene is not just an example of the MacGyver-like ingenuity in the show, it is something I want to freaking do in a video game!

Imagine a fully-realized virtual world with this kind of physics system.  Imagine the tactical choices that would crop up in, say, a first-person game.  Do I take the time to run around the wall?  Do I use my last brick of C4 to blow through it?  Do I move that table over and hope it will support my weight, so I can climb over the wall?

As you can imagine, a greater level of environmental realism would require a greater level of environmental interaction, which in-turn means greater attention to the physics of characters in the world.  The Euphoria engine (SW: The Force Unleashed, GTAIV) is a great start in this respect.  I really want to see it used more, especially in tandem with the level of environmental interaction I have in mind.

Let us dream up a little situation to use as an example:  one character tackling another next to a wall.  First, imagine that the tackling character is smaller, and moving fairly slowly.  He hits his target square, the tackled character tips off balance, and reaches toward the wall for balance.  Only the wall is made of paper, and gives under his weight.  Both characters topple through the wall.

Second, the tackler is larger, moving fast, and hits low.  His shoulder hits his target’s knees hard enough to break one, and his upward trajectory knocks the other off his feet, and over the tackler.  Unfortunately, this leaves the tackler with most of his momentum intact.  He slams into the wall, this time made of drywall, with underlying wooden studs and steam pipes running throughout.  His head breaks through the drywall with minimal damage, but his collarbone hits a stud with enough force to crack the bone.  As he reels back in pain, the tackler reaches out and grabs a pipe.  As his weight shifts back away from the wall, he bends the pipe out of its fitting, and steam begins shooting up into the air.

Third, the tackler is smaller again, but moving much faster.  He hits the other character high, transferring most of his weight at the shoulders.  Because the other character is so much heavier, he absorbs more of the tackler’s momentum, and the two separate this time after the impact.  Good thing too, as this time, the wall is made of glass.  His arm shoots out for balance again, but this time punches right through the glass.  The top half of the pane breaks outward, but the bottom half remains in a jagged, deadly edge.  The tackled character comes down almost on top of it, and the glass cuts at his fatigues.  They hold at first, but as he twists and slides along the edge, the fabric catches and the glass cuts through it, and deep into his armpit and chest.  As more of his weight is transferred to the glass, it too breaks, leaving a number of shards in the character, and he topples through the pane.  Blood is dripping and torn pieces of his uniform may be seen hanging from what remains of the window.

Damnit, is anyone else excited about this stuff yet?  I know I am not the only one who would lose his shit if I saw this in a tech demo.  I just want to know it is in the works.

(Last minute addition)

From what I hear, Red Faction: Guerilla is a step in the right direction.  I played a demo and was not particularly impressed, but I’ll give it points for enthusiasm.