Pixel Counterrr

The history of simulated space, as I wrote in my thesis, basically consists of going through several stages of development, most of which relate to the level of spatial latitude. My observation was that the freedom of movement the player, or the “camera”, has within a game has increased one dimensional/rotational axis at a time over the years, starting from a fixed game space with no spatial freedom at all (think of, say, PacMan), to the current state of six degrees of freedom (think of any modern game).

So games like Wolfenstein 3D (1992) and Doom (1993) were positioned somewhere between the two extremes, in that they lacked, thanks to the ray-casting technique, the ability to tilt the view up or down. The vertical lines remained parallel however player positioned the view within the game, there was no escaping it.  Interestingly this restriction liberated developers worrying about something that came with the first-person view, which is as follows.

As Quake (1996) was de facto the first FPS to introduce the full six degrees of freedom which allowed the player to genuinely look up and down, it also brought forward one of the biggest issues with the first-person view by and large. It’s as simple as when looking down in a FPS, one should, according to all logic, see the rest of the corpus, or at least the legs, especially since there are usually other body parts visible, i.e. hands holding the weapon or what have you.

The issue here is that in most cases the rest of the body simply isn’t there for people to behold. Even such prominent series like Half-life and Call of Duty employ a first-person view that bears little of the feel of controlling a physical corpus, but a floating, massless camera with arms and a gun attached to it instead. Yes, Call of Duty games do show the body of the character aside from the hands here and there during some of the cut-scenes, but, to my mind, that just doesn’t cut it.

The thing is, the first-person view should be all about the illusion of being in someone else’s shoes, quite literally, so it surprises me to some extent that so many games of that genre ignore the rest of the body altogether. I know it’s mainly a technical thing and something which many people presumably don’t even care about so much. Nevertheless, I would argue that if FPSs want to move forward as a genre, the corporeal nature of the action has to be integrated more and more not only into the visuals, but core gameplay mechanics as well. I do acknowledge that such an endeavor probably isn’t as straightforward as when dealing with the third-person view, but pushing the envelope never is.

Much hyped Jurassic Park: Trespasser (1998) was an early, highly ambitious attempt to put some meat and bones into the FPS experience, failing miserably at almost every level. F.E.A.R. (2005) is worth mentioning as an endeavor to incorporate martial arts into the genre using a visible lower body, not to mention Dark Messiah of Might and Magic (2006).

To me, it was Crysis (2007) that showcased the impact a decent first-person body can have on the visuals overall. Indeed, to see one’s own shadow hovering over the sandy beaches while shooting at enemies was something of a breakthrough in my eyes, which would’ve not been feasible without the body casting the shadow. It still continues to amaze me whenever I pay attention to it, and boggles my mind why, for instance, the Call of Duty series still ignores completely this aspect.

Furthermore, Crysis earns extra kudos for the way weaponry and other movables are picked up using the character’s hands, and not just teleporting them into the inventory. Or, levitating objects at the center of the screen in the manner of Half-life 2 and derivatives. In Crysis, there really is a strong illusion of a physical connection to the game space through the character, and the brilliance of the system can’t be emphasized enough.

Then there are instances where the first-person view wouldn’t make much sense without rendering the rest of the body. Mirror’s Edge (2008), for one, is a great example of a first-person game which makes great use of the body when parkouring over the rooftops, or delivering kicks and punches. In addition, the corporeality manifests also through the gameplay itself in terms of inertia, mass, and momentum that each have to take account when jumping and landing with the character.

That said, I still wouldn’t go so far as Shift 2: Unleashed (2011) which featured so-called Helmet Cam that put the view inside, you guessed, a helmet. Sure, that sounds like a logical extension to what I’ve written about here that everything related to the game character should be modeled in the first-person view. However, for some reason, such a solution feels gimmicky and, moreover, wrong.

So what my position is regarding the first-person view is that when presenting a game from it, everything below the neck should be there visible to the player, not merely arms and hands like most games seem to operate. I guess I have to make a follow up post to open that “neck-line” dogma further, but it most probably has got to do with the fashion the player’s real head and body relate to the screen.

If there’s one concept regarding video games to which everything written here at my blog comes ultimately down, it have got to be the concept of simulation. What’s so intriguing and unique, then, about simulation in relation to other forms of representation is that simulation doesn’t replicate only how a certain entity look, but how it functions beneath the appearance. Simulation is always a dynamic system of input and output, which is the beauty of it.

As I noted in my thesis, simulation as a notion has basically two sides: An educational/scientific, i.e. more “serious” side and a playful side, of which the latter video games are more closely representative. The main difference between the two is that simulations in the video game space are mostly aesthetic-driven and less accurate regarding the physical reality than the educational/scientific simulations. So simulations for merely entertainment purposes need merely to look good and be truthful enough to provide a credible impression, which makes such simulations that more interesting to a visually oriented person, like myself.

Bearing that in mind, video game simulations, such as of lighting or physics, are more than anything about compromises and trade-offs, and thus generally far from perfection, but decreasingly so as the technology moves forward. Consequently, it’s pretty given that practically every simulation technology that comes along is mere temporary, waiting to be replaced with more advanced one. Of current technologies, for instance Ambient Occlusion in all its forms is such a lighting solution of which obsolescence is only a matter of time once genuine indirect illumination schemes become more feasible.

What, then, comes to simulation of space, there was this endeavor in the 90s to use voxel ray casting technique to depict 3D landscapes in a relatively detailed, but highly constrained manner. The first commercial video game using such a method was NovaLogic’s Comanche: Maximun Overkill released in 1992, which blew visually the competition out of the water. There wasn’t anything near as impressive as Comanche on the market at the time, and everyone with a decent PC must have thought that the future belonged to some kind of a voxel system, not filled polygons.

We all know what happened to the mainstream use of voxels later on, but there was that brief moment in the early 90s when voxels did arguably better job in conveying the appearance of a bumpy 3D landscape. Of course, the visual glory of voxels came with the typical limitations associated with ray casting. Still, polygons in 1992 were too few and far between to be able to depict even remotely the complexity provided by voxels, and as such, far behind purely aesthetically speaking. Eventually games like Flight Unlimited released in 1995 started to establish the idea that texture mapped polygons was really the way to go even for detailed scenery in which voxels excelled so well, and today we really have no credible alternative paradigm what comes to geometry.

To yet further illustrate my case, Comanche 4 released in 2001 was the first one in the series to abandon voxels for hardware-accelerated polygons, the technology that was the final nail in the coffin of software rendered voxels.

The demise of voxels proves if anything that in the realm of simulation technologies, some solutions are hot at one point, and obsolete in the next. And who knows if the contemporary polygon-paradigm which we so certainly believe in, will be somewhere down the line substituted with some uncanny technology that our current minds can’t even grasp.

After all, technology by nature is not built to last, but to be replaced.

In 1946, an essayist and poet Jorge Luis Borges wrote a frequently referred to fable about ambitious cartographers who drew up a map so detailed that it ended up covering the territory of an Empire exactly. This, of course, defeated the very purpose of the map, so it eventually decayed to shreds under the feet of successive generations.

In the digital realm, however, we don’t luckily have above kinds of spatial issues, which enable us to create as detailed representations of reality as we are willing and able to without worrying how much virtual space it may occupy. The world is ours to model.

And that is exactly what must have been on the minds of the people at C3 Technologies when they decided to develop their cutting edge aerial 3D imagining technology. I saw first demo videos of it few years back and couldn’t believe the level of definition and overall quality of the imagery: It looked too good to be true.

But now that the C3’s mapping technology has put in more mainstream use by Nokia, the significance of it becomes clearer and clearer: This changes the geological cartography as we know it. And the true beauty of the tech is the presumably minimal amount of manual labor that goes into the creation process, so the cost and time per square-kilometer should remain reasonable, which is crucial for the future of the technology.

Admittedly Google Earth does include a similar project to 3D model at least the major cities, but the fact is, it doesn’t play even in the same ballpark what comes to the level of photorealism. Currently, C3’s offering simply kills the competition as far as I know, and it’s hard to see anyone else to come up with a better (or even equal) solution in any foreseeable future.

Of course, C3’s maps are far from perfect 3D representations of their real-life counterparts since the maps basically simulate only space [1] and not light, which becomes quite apparent and distracting when looking at metropolitan areas with highly reflective skyscrapers. Plus, the shadows are completely depended on the weather conditions of the day of shooting, so the overall look may vary drastically within a large region, but these are known issues of photo-based modeling in general.

When I was younger, I was very much into scale models, like dioramas and such. They fascinated me beyond comprehension and my theory is that real-time imagery, like these 3D maps, intrigues me for similar reasons. I’m yet to pinpoint exactly what those reasons may be, but my gut feeling says it has got to do with this god-like, omnipotent perspective on the reality. That one can seemingly twist and turn a piece of reality as one pleases, which, of course, isn’t the case with the “real reality”.

[1] see my thesis, Chapter V: Simulation of Visual

The original Test Drive developed by Distinctive Software was my first introduction to something I call first-person driving. Of course, the game is rudimentary at best judged by today’s standards, but it popularized a number of game play elements that we now take for given, such as a roster of differently handling civil cars, police chases, manual shifting, and so on.

TD was not indeed only an inevitable reaction to the so-called yuppie culture that dominated the early 80s, but also an ideological ground zero for, for instance, Need for Speed –series that is easily the most important intellectual property for simulated driving at the moment. TD proved that racing doesn’t have to take place in a dedicated track and that it can be actually more fun and intense outside one. Just like in real life.

It comes as no surprise that in TD, considering it was released in 1987 for home computers, the simulation of depth was carried out with jumpy sprites. The biggest problem when using sprites to depict 3D space is of course their complete lack of depth, so cars, traffic signs, and so on are nothing but cardboard cutouts of their real life counterparts.

To be fair, sprites in TD worked generally fine in the distance, but the “closer” they got, the flatter they obviously appeared. And when overtaking another vehicle, the fragile illusion of three-dimensional objects broke down completely.

TD’s sequel, The Duel: Test Drive II released two years later, kept the graphics paradigm basically intact, but in 1990 Test Drive III: the Passion finally entered the realm of (jumpy) polygon-based imagery.

Sure, the graphics engine of TD III was a bit of a mess and as a game almost unplayable, but one thing that struck me heavily back then was the pure idea that now you could actually see the side of a car when driving past it without the illusion disintegrating. And it didn’t bother me one bit that the polygon cars were in some ways much less detailed and more abstract than the sprite ones. No, that was beside the point. The point was real, genuine depth of objects no matter the visual cost, which was, of course, substantial back then.

I have always found it extremely interesting whenever a certain visual concept is translated into a different graphics principle. The thing is, when immigrating from sprite-based imagery into polygon based such (or the other way around) more often than not the viewing paradigm shifts drastically in the process. Think about for instance Super Mario World and Super Mario 64, how they both depict the make-believe reality in two completely different ways.

That said, what ultimately made the jump from TD (or in fact TD II) to TD III fascinating was the fact that the fundamental viewing paradigm (first-person) was the same in both instalments, even though the underlying technology wasn’t. It was like the developer was forced to admit that the original sprite-based solution was indeed all smoke and mirrors, and now with TD III they were actually serious about it, first-person driving that is.

The reality is, to come across aesthetics generally associated with polygon-based imagery outside the digital domain can be considered a somewhat rare occurrence, most likely due to the distinctive fundamentals and principles the medium is based on.

But it does happen.

One of my favorite cases of such are sculptures made of archival card and foamcore by Susy Oliveira, in which an aesthetic principle basically jumps off from the computer screen into the tangible reality, in a quite elegant manner.

So, as we noticed in the previous discussion, there are indeed times when reality uncannily imitates art and not the other way around. And to be more precise, computer-generated such. Furthermore, in addition to the case of NASCAR headlights echoing old-school texture-mapping, I made a brief remark about a fighter plane known as F-117A Nighthawk, implying how it resembled to an extent of something made out of polygons, and relatively a few of them.

I do acknowledge that a visual appearance such as of a stealth fighter is designed for various, non-aesthetical reasons, like aerodynamics and more importantly, to deflect radar signals. Even so, I find it extremely fascinating when aesthetics found in the physical reality collides with of the virtual world, like it does in this mysterious, pointy plane.

Better yet, as said in the NASCAR post, where things get really interesting is when this kind of “mimetic art” (if you can say so about a military vehicle) is in a way brought back to its original, aesthetic environment. As one can notice from the above Nighthawk model extracted from Tom Clancy’s H.A.W.X., the polygon-mesh isn’t just an estimation of the surface of the plane, it is the surface.

Obviously, there’s a myriad of real world shapes with similarly minimal geometrical properties, and thus equally analog to low-poly imagery. Examples that come to my mind are certain buildings, designer furniture, home electronics, and so on. But what then made Nighthawk so special in my mind must have been the context, as fighter planes at large tend to be more complex in terms of overall geometry.

Indeed, when comparing the model of Nighthawk to other select aircrafts found in TC’s H.A.W.X., the Nighthawk’s ▸polycount was noticeably lower than the average, and thus substantially more “polygoney”. Sure, F-22 Raptor came close too, as it bears similar stealth design, but in my mind, Nighthawk is still a premium example of what I’m trying to convey here. Additionally, the difference would probably have been even more dramatic if I had compared only the fuselages of the planes, since a chunk of the polygons go to the landing gears and other minor detail.

All this actually makes me think of the famous Death From Above –scene in Call of Duty 4: Modern Warfare, which could be seen as a video game mimicking a weapon mimicking a video game, and I believe the whole point of that scene was exactly to emphasize that absurdity.

I have always had warm feelings towards polygons as pure forms, even (or especially) back when they were such a scarce resource that one could literarily count the number of onscreen polygons – a task which often low frame rate made quite feasible to undertake. There is some elegant, cold beauty to be found in old-school flat shaded polygons, and a game like Virtua Racing still looks in a way as absorbing as it did back in the day.

Obviously things have changed since, and today polycount isn’t considered as a major issue any more within the real-time medium. And the use of polygons aren’t limited only in flight simulators and such either, like in the early days, but are been employed across the genres. In fact, polygon-based imagery has become so mainstream and mundane that polygons go generally unnoticed in these days: they are simply the stuff what everything is made of.

That said, an interesting pheonemon happened when I fired up Reckless Racing on my iPod Touch one day. The geometry in that game really jumped out of the scenery, even though the game wasn’t supposed to be all that three-deey in the first place, but rather resembling an old-school top-down racer. For instance, the extra-tall ▸churchtower appeared really impressive on the screen when driving past it, which got me thinking what actually made that seemingly generic geometry so special?

Let me tell you.

I believe the reason lies in the 2D gameplay itself, meaning the game could have been carried out in theory without using a single polygon, like they did in the ol’ days. The thing is, any game using, for instance, a first-person-view needs some sort of a 3D engine to work out even in theory, since the gameplay operates usually in all three axes (the z-axis is actually the key issue[1]). But games with strict x and y –axes (2D) gameplay, such as RR, don’t require a 3D engine to be fully functional, thus in a way polygons are there solely as luxury items without any genuine need or purpose in terms of the gameplay. And the church tower in question was obviously the pinnacle of that excess-ness and unnecessary-ness that for some reason fascinates me. Or what fascinates in luxury by and large, I might add.

Additionally, part of the aesthetic charm must stem from the disorientation the visual muscle-memory of hundreds of sprite-based 2D games with a similar perspective causes to one’s brain. Perhaps the mind projects now-obsolete limitations on top of the 3D imagery and acts surprised when the fantasy limits are suddenly “broken” by polygons sticking out of the ground.

Of course, there are number of similar cases which fascinate equally, like Street Fighter IV and Bionic Commando Rearmed to name but a few. The latter even includes completely unnecessary, but beautiful, ragdoll physics as a bonus.

[1] For instance, Wolfenstein 3D operates only in two axis: x and z, but needs a 3D engine because of the z -axis.