Right now I’m reading a book on Byzantine religious art called ‘The Sensual Icon’. It’s enlightening in some deeply strange ways. It seems to me that a lot of art history gets things wrong, at least at the survey level. Ancient and medieval art is tacitly presented as if it were modern, but older. Hang it on the wall in an evenly lit room with a name, date, and likely attribution. It was never meant to be seen like this, whether it was a plastered skull from Catal Huyuk, or a gold bas relief of an archangel from the Hagia Sofia. The angels look walleyed and stupid in electric light, but in something harsher than the flames of candles the central effect is lost. It’s like the gilt and silvered glass of Versailles: Shadow is needed to sculpt the light. In daylight these things are too shiny, too soft in profile, and too delicate in color. It’s like Liberace had a yard sale. Placed in the deep night however, and the warm glow of small open flames brings life to dead metal. The mirrored corridors of the palace become starscapes dotted with the magnificence of kings, and the round and fishlike face of St. Michael becomes something else again in the light of the polykandelia through a haze of smoke. The eyes don’t follow you, to duck the cliche. Rather, as the light source is moved, the eyes look upwards, then across the spreading scene, then down into mystery and the dark. What seem like anatomical errors and crudity of technique turn instead into a mastery of light. The face is not perfectly rounded, but rather faceted. The eyes are the same. What might be a cheap effect becomes far more lifelike because the faceting breaks it into discrete expressions with soft but distinct transitions.
It might all be called theater, and I think this is true on some level. This degree of dismissiveness does little to credit either the scene or the viewer, though. Myself, I see it through hungry and lecherous eye of the apprentice. While the goals of the concept artist and level designer are admittedly more modest, a boatbuilder might look upon the work of Harland and Wolff with some degree of an insider’s appreciation. This last month, I celebrated, in private, two extraordinary (and coinciding) ten year anniversaries. I’ve been making art for games full time for a decade, and also mark a like amount for an enduring friendship.
Ten years. I’ve noticed a few things, but feel cautious as to how to communicate them. I wouldn’t have you think that I should be regarded (at best) as something other than a first among equals. I do think that concealing what might be helpful to others would be an error as well. This is a fabric that we’re all weaving together, though the task is both larger than us and something that will likely outlive us. Anyways, here goes. Feel free to disagree or mock my pretensions:
Keep in mind where your camera is if you’re designing things for video games. An FPS and an RTS need very different silhouettes for vehicles and characters. In an RTS the camera is going to be up real high and will be real far away. You’re also going to usually have a lot more units on the screen. Depending on your lighting model you could lose the silhouette of your unit real easily. Things to look for here are ways to differentiate units and teams from each other with massing and color, respectively. With the larger texture sizes available to us on current hardware a somewhat more sophisticated approach might be taken to team colors. That is, colors might be done as actual textures, rather than palette swaps or vertex color. If you can zoom in a lot and you have texture space for it, you can add detail that can be appreciated up close. The trick is to keep it relatively low contrast, so that when you zoom out to low mipmap levels it just goes away, rather that turn into something that sparkles in a displeasing way. Likewise, try and keep the small details relatively sedate in your normal maps. Things like gun barrels and antennas will tend to get shorter and fatter. Partially this is for style reasons, though there’s something of a complex dance going on there which could be an essay all its own. Much of this is again to play to the camera and to make sure that you call easily tell one type of tank from another when the camera is zoomed all the way out. The more information that you can share coherently with the art, the less has to be done with expensive article effects and cumbersome UI overlays. RTS unit design occupies one end of a broad spectrum, and something like an FPS lies at the other end.
A first-person shooter lies at the other end of this spectrum. While characters, vehicles, and props can be of cinematic quality, the concerns and priorities are quite different. In a movie, the lighting and camera position can be controlled more or less precisely. In an FPS, the camera represents but a single view, which moves with the player. You (aside from cutscenes) don’t have a stationary camera to draw attention to particular actions in the same way. Likewise, lighting isn’t as sophisticated as a movie. Game engine lighting has improved a great deal over the past decade, but if you’re shooting at a bum rush of slime people you can’t really count on the lighting to distinguish one opponent from another. Silhouette is still king here, though with the lower camera angle you can’t just use an RTS model and have it look right. The information on something like a vehicle need to be laid out differently. In an RTS it sticks out to the sides, while in an FPS it sticks *up*. It’s one of the core, but seldom articulated, skills of a concept designer to communicate gameplay information while making it attractive and not breaking immersion. There’s been a increasing tendency to use the visible weapon elements to tell you where you’re aiming, how much ammo you have left and whatnot. There’s a lot of potential there, but you also have a set of challenges related to making the weapon you’re using appear interesting (and believable) from the back and the oblique side.
The over-the-shoulder camera you see in MMOs and squad shooters is a third case, which while it’s tempting to think of it as something in between is more of a distinct problem. One of the biggest peculiarities from the art production standpoint is that you’re staring at the back of your character most of the time. That’s a lot of overhead for something that doesn’t add to gameplay. The way Dead Space used your suit as a health indicator was an interesting solution to the problem, and let the designers ditch the UI in the bargain. One of the problems here is making buildings look right with the gigantic doorways you need to get the camera inside.
I bring all these examples up no such much to recommend specific solutions as to suggest things to be aware of when you’re designing units, props, environments, or whatever. Without getting too deeply into it, there’s a lot of competition for a relatively small number of game jobs. You may *just* want to model, concept, or animate, but if you understand and have at least some ability with other parts of the pipeline it gives you an edge, and potentially greater opportunities.
Going forward, I’d like to share some of the things I’ve learned, modest though they may be. It strikes me that’s there’s a problem with how information on game art is shared, and I’d rather not become part of the problem. You see, most of what a game professional does can’t legally be shared until well after the fact, if ever. Even if your NDA expires, a lot of it can be quite gossipy and impolite, even career suicide, to talk about openly. Hence, the people that know (in a practical sense) can’t say, and those conversations that do happen in public are often dominated by journalists seeking readers, ideologues with no real industry knowledge, amateurs filled with the certainty of ignorance, and a variety of has-beens and hangers-on looking to set themselves up as experts (for a modest fee). This entire shadow play is fueled by the immense hunger of the next generation of professionals for knowledge and a chance for a leg up in a highly competitive field.
I’m coming at this from a slightly different perspective, though I recommend you be as suspicious of me as of anybody. The internet is full of crazy people. For my own part, I’m trying to share some of what I know out of a respect for the work. In the end, we’re just making video games I know, but it’s also what we’ve chosen to do with our time, and as such is worthy of doing well. I won’t tell you what you have to do, because I don’t know. What I will try and do is make observations and suggestions to the best of my abilities.
To that end, I’ll go back to about where I began: it may seem strange, or even blasphemous, to speak of Byzantine sacred art in the context of concept design, but the thing here is to realize that what the modeler builds from your concept is going to be greatly modulated by the lighting, mood, and camera location in the level. The more you understand the contributing factors the better handle you’ll have on designing something right the first time and saving everybody else time during development. The concept artist sits near the beginning of the production pipeline, and the decisions made by them and the art director determine much about how the entire team does their jobs. Understanding both the technical and psychological aspects of the implementation of your designs makes you more effective as a concept artist. I like art history for this reason, among others. You may find than an adaptable solution for your current problem was worked out a thousand years ago, which gives you more time to play video games or re-read Lackadaisy.