Inside the striking art and design of Hawken

Hawken greeted the game-playing public with a video -- a simple gameplay trailer that quickly gathered almost 2 million views, and the interest of dozens of publishers. But the team at Adhesive games wanted to remain independent, so instead a publisher (Meteor Games) was formed around them with a $10 million investment from Benchmark Capital.

Perhaps the most striking aspect about this game is the art. It seemed "next-gen" before we even knew what "next-gen" would look like. This is due in part to the pedigree of the team, which was five people at the time of that trailer. Almost all of them had worked on the similarly-striking Project Offset, a fantasy game project that was acquired by Intel, but was never released.

Hawken leader and Adhesive game's CEO Khang Le has been doing visual development for movies and games for years -- and he took a unique approach to the game's art and design. Rather than aiming for the moon, he and his team tried to create a game that played to the strengths of its members.

Let's talk about art methodology -- you seem to be creating a high level of detail with a pretty small team. What is your production pipeline?

Khang Le: Originally we only started with five people. So, at the time we actually didn't have any idea of what to do, yet. One of the things I know from working on many previous art productions is that sci-fi is a lot faster to do than fantasy, because you can repeat assets. So, let's say you have a broken column in a fantasy game. You can't really repeat it that many times because, you know, it's very obviously being repetitive, that same broken shape. But for a broken sci-fi column, like a very tech-y looking column, you can repeat it as many times as you want to and the audience just accepts it.

That's one of the reasons why we went with sci-fi. We only have one animator, so, I didn't want to do anything that had a lot of complex humanoid animation involved. So, robots were the logical step. And I'm always very much been a big fan of designing robots. I love to draw and paint them... so, all those other things just came together.

That's where Hawken came out from. It sort of came out from working around the limitation of the team, instead of just blue sky-ing. It was less like, "What do we really want to do?" It's more about, "What do we have? What's possible? What can we deal with in such a short time that still feels impressive?" That was the decision we made.

Basically we focused more visually on the overall picture of the game. The specific assets, like the little props that a lot of bigger companies have time to put in... They put a lot of attention into little details, where we just sort of made the overall picture look good. If you actually walk up close, it's there, it's good, but it's not super-polished.

But since we do that, our assets are much lighter. They don’t have their own normal map or diffuse textures or speculars, so we're able to have a lot more objects than other companies can put into their game engine, because each asset is pretty light. Our scenes have more objects than most games, but each object itself is less detailed and intense on the machine. 

It seems to have made an impression, because the game's early buzz was very much because of the art. Like you say, if you look at a screenshot or a video of it, it looks extremely impressive, but you tend to be zoomed out. I guess it's built more for scale and speed than for staring an asset real close.

KL: Yeah, especially, our game is not a single-player game where you just hang out and talk to NPCs. You're going to be focusing on just attacking the enemies on the screen and the other player. You're mostly going to be viewing the overall picture anyway. So we don't really put our emphasis on, you know, a light fixture to look good up close, we just make sure the picture looks good and impressive.

The camera frames everything you look at in the game -- and the cockpit camera and general motion seems very believable, even though you're piloting this thing that doesn't exist. How did you arrive at what you have there, and the tradeoff of visibility of the environment versus the aesthetic of shaking around and being in a mech.

KL: So, the cockpit -- I mean, you're right, especially, if you think about it, it's just going to distract you from looking at the world, right? It'll cover up, maybe 30 percent of the screen. But as far as a mech game goes, I think that's very crucial. Being inside a HUD is what I loved back in the day with the old MechWarrior and all the Steel Battalion; the feeling of being inside a machine. That's why we chose to go first person and not third person.

By creating these elements that block off 30 percent of the screen, we also try to make it make sure it's also functional somehow. So a lot of the UI elements are right on the HUD -- right on the cockpit. In order to make it feel like you're in a machine, I did a lot of research and watched tons of videos... I guess a lot of people, they record themselves when they're working those industrial machines, like bulldozers and things like that. And just really studying all the motion that's going on, the way how your head turns, your vehicle will get delayed behind. Things rattle. Every sort of physical detail like that, we tried to put in as much as possible. And there's a fine line between making it very realistic and giving someone headache. So, we're doing our best to find that middle ground, making sure it's fun, but at the same time very immersive.

How do you build in that sense of weight that a mech has, where every footfall has a lot of power to it and also when you're rocketing up into the air, you're falling perhaps even faster because of how heavy the thing is.

KL: For comparison, the mechs aren't that big in MechWarrior. They're actually about 18 feet, 20 feet tall. So they can move a build faster, but our turn speed is currently slower than a typical FPS now would turn, just because we don't want a mech head to whip around like a human's. It would look very cartoony like that.

Another thing is we delay how the cockpit follows the mouse slightly, and then we're also doing a lot of things like when a mech walks past you or another mech flies past you, it'll rattle the cockpit you're in. So, just the little details like that. Anything related to the physical mass and the sense of weight to the mech, we like to add in. At the same time, playing the game is all about... we're a much faster, twitch kinda gameplay. It's a fine line to walk in between them.

Right, that's what's made mech games an acquired taste for a while, because for some people -- it either feels too slow when they're walking or when they're boosting they're kind of on rails almost.

KL: Yeah, yeah, they have that sort of Armored Core feeling, like you're like on an airplane. The closest game that we might compare to is Heavy Gear, you know? Back in the day? It has a sense of weight but at the same it has a sense of speed. So we're pretty close to that. We're somewhere between Halo and, let's say, MechWarrior. Somewhere right in the middle. We're not as fast as Halo. But we're not slow. 

How do you go about designing mechs? Do you try to build them for actual plausibility, so they could actually be made or do you go more mostly for aesthetics?

KL: As far as mech aesthetics go, there's a lot of different variety that people are used to, and they like a certain type. There's Transformers, Gundam, Evangelion, the humanoid-looking mechs that has a head and fingers. There's the American-style mechs with the lumbering gait, feels like a machine -- the mechs in MechWarrior or the robot in Robocop (ED-209). There's also very sleek-looking... like Bubblegum Crisis, that type of mech.

The one I really like, my personal favorite, is just an old, basically 1980s kit bash style, it's from this Japanese designer named Kow Yokoyama. He does a line of robots called Maschinen Krieger, so they kinda look like World War I, if World War I had tech robots and mechs. That's what they look like. Still very industrial machines, but they also had a bubbly-looking element to them. Almost insect-like. But at the same time we're trying to please the typical mech crowd. So we also have mechs that feel very much like tanks. Very square-looking. Basically an Abrams tank with legs. We have that style, also. So we got, basically, two styles to cover for now. And eventually as we keep expanding we'll go for more variety to make everyone have something that they enjoy looking at.

There seemed to be a bit of a Masamune Shirow-type chunky mechs in there, too.

KL: Yeah, definitely. Very much like the, slightly insect-like, but at the same time very mechanical.

But as far as the actual process of designing these mechs, do you where do you start and how do you build them? Do you just draw them whole cloth or do you start with the cockpit? Do you start with the weapons? What do you do there?

KL: We have three different styles of mech: light, medium, and heavy. And they have their own animation effects that we had designed beforehand, so all the mech designs, we build them off those skeletons. What we do is just start painting. I would design the whole mech all at once, and hired another really amazing concept artist named John Park, who basically took all that from me.

But, yeah, we just sorta just design them, we just draw them, and another way we actually do it is here is we just kit bash, like the old Star Wars ethic -- we just have a big library of tank parts and helicopter parts in 3D and we cut them up and assemble them like little toy LEGOs and see what we can come up with. It's very fun. A very organic process.

Sometimes you'll just build them straight -- the models straight from scratch?

KL: Yeah. In Sketchup or Maya or something. We have a big library of parts now that we can slap together. The thing about the old World War I tanks is they were pretty much built the same way, and they were built from cars. Ford actually had cars and they just slapped on armors and stuff like that to get them ready for war, so, we're trying to get the process the same in digital and 3D. 

And do you plan the sound and animation for these things simultaneously? The sound design is very integral to the mech experience.

KL: Yeah, we're really lucky. Our original sound designer, back in the day, was Sean Neri. He did the initial sound for the first trailer, but since then it got super popular and he got his own job somewhere. We have another sound designer here now, so all the latest trailers you've been watching now is from another guy named Shadi Muklashy, and he's awesome. He's our sound guy, all the music you hear, and all the current UI interface guy, he does it all.

For example, we love Battlefield 3, the sound quality they have in there, which kinda teaches us that everything has a sense of weight and variety. Like, the machine-gun sound is very different in the distance compared to up close. So it's not just about muffling or fading out the volume. The sound actually changes in the distance. It actually gives off a different sound in the distance than when it makes a sound up close.

So for us, when you go into a tunnel, they become very echo-y, and even the footsteps of the mechs, we want to make sure they feel very heavy and machine-like, but at the same time it doesn't give you a headache. Because if you ever think about it, if you were to actually drive in a mech, you might get that groaning metal sound consistently pounding through your ears, and your ears will get very tired of it. So we have to make it sound heavy and very metal-like at the same time, so it's musical and pleasing to the ear, so it doesn't get tiring for the player.

When there's changes in distance between, I mean, with machine-gun fire or moving in and out of echoy spaces, do you use blends between them or do you just switch between the two?

KL: For echo and, you know, for normal stuff, there's different ways -- like the volume, and in the Unreal engine you can change it to make it kind of more echo-y. But for, let's say, a distant machine-gun sound compared to a close-up one, we actually create two sounds and we just change it by distance.

Let's go back to the launch of that first trailer, and where you are now.

KL: When we first launched that video, we just sort of released it on YouTube and we emailed a couple of small indie sites to let them know we have a video up. We didn't expect much of anything. We woke up the next day and it was, like, 200,000 views. And I think in a week or two it became 600,000. That was our first video.

And within three days we had 80 percent of the large publishers out there contacting us: EA, Activision, Sony, Square Enix. Basically every company came out wanting to publish the game, and eventually we found that it's awesome to get that sort of support, but we still want the freedom of being an independent team, and the creative control, and the fewest middlemen possible between us and the consumer. We just want to go directly to them without going through, you know, Best Buy. Without going through a publisher.

The best business model for that -- at the same time, like, bringing new content, updating all the time for the consumer... since we're such a small team we're unable to bring out a finished product immediately. With free-to-play we can actually bring to people in small chunks at a time and grow our team slowly and that was kind of confirmed by someone at benchmark capital, who was heading our funding at Riot. At the early stage, he approached us and he pitched us the idea of being free-to-play. We always had plans to make a free-to-play game, but our plan was more like Valve, with Team Fortress. You know, going digital download first and then free-to-play later. But he recommended just going straight to free-to-play and making it our focus so we don't have to keep changing our product. And, yeah, it worked out.