King of 2D: Vanillaware's George Kamitani

[Japanese independent developer Vanillaware has created modern 2D game classics such as Odin Sphere and the upcoming Muramasa: The Demon Blade, and founder George Kamitami sits down with Gamasutra to discuss his company's roots, passions, and plans.]

There's a small but passionate group of people who still care about 2D gaming; the majority of the industry, and gamers, have moved on to 3D -- years ago, in fact, at this point. 2D is only routinely used on portable platforms.

However, somehow, Japanese developer Vanillaware has carved out a successful and critically-acclaimed niche creating 2D games. Its Odin Sphere was released for the PlayStation 2 in 2007 -- a gorgeous dark fantasy that, despite its apparent limited appeal, sold in the hundreds of thousands globally, surprising many.

The company's 2009 release is Muramasa: The Demon Blade for Nintendo's Wii. Released in Japan in April, the localized version of the game will reach North America in September via publisher Ignition, and Europe in November.

It's generating a lot of buzz -- fans who played the import version speak of a step up in gameplay quality from Odin Sphere; the game won several best of E3 awards from enthusiast publications.

Here, Vanillaware's founder, George Kamitani, talks about his ambitions for the company -- sticking with 2D, going to HD resolutions, and maybe even making an online game.

Can you talk a little bit about Vanillaware's origins?

George Kamitani: Well, it's not a simple tale, definitely. (laughs) Vanillaware was founded around the time I went to Tokyo. At the time, I was working with Square Enix directing the development of Fantasy Earth. It started as a small project, and I was just contributing to it on a personal basis, but it ballooned in size to the point where an entire team was established to complete it. In the beginning, the company was called Puraguru.

Before that you were at Atlus, correct?

GK: Well, directly before that, I was at a company called Racjin, a game developer. They worked on Mario and Sonic at the Olympic Games most recently, I think. Largely I participated in their projects on an outsourced basis.

Princess Crown

How many people that are at Vanillaware now actually worked on [Vanillaware's 1997 debut game] Princess Crown?

GK: Within Vanillaware right now, there are three people who worked on Princess Crown, myself included.

And how many people are now in Vanillaware?

GK: Currenly, 21 people.

And what percentage of that is artists?

GK: Basically, 100 percent. (laughs)

That makes sense. It seemed like Vanillaware suddenly exploded in 2007 with a bunch of titles. How did that happen?

GK: Well, a lot of announcements did come out all at once, certainly. It wasn't our aim, really, because when it comes to consoles, we only have one development line going at any one point.

In the case of Odin Sphere, Atlus instructed us to have the game done within 2006. We completed the game fully within 2006, but sales on Persona [3] were going so well for Atlus around that time period that the publisher pushed the release date back a few months to keep from cannibalizing its own market.

We began serious development on GrimGrimoire after Odin Sphere was completed, but that wound up coming out in Japan about a month ahead of Odin Sphere.

Do you find that it's difficult to find hi-res 2D artists in Japan these days, because not many people are doing that?

GK: In terms of pure art, there are a lot of people out there with the talent. However, most artists these days are simply unfamiliar with the older styles of 2D animation, so our only option is to train them in that field.

Some time ago, I spoke to someone from SNK about King of Fighters XII. He said it was quite difficult to find people who already knew how to do 2D graphics at that level.

GK: Definitely so.

What do you think are the main challenges to getting to 720p-level of hi-res? Right now, Muramasa is 480 -- how big a leap is it from 480 to 720 -- what KOF accomplishes?

GK: Well, the original art we draw is all done in double-size -- in the case of Muramasa, the animation frames then get compressed down to the Wii's native resolution. As a result, producing a fully HD title would not be a great deal of extra work for us; it would just mean our original art is displayed in higher resolution. It wouldn't be a simple insert -- we retouch the compressed graphics here and there to make sure they look as good as possible -- but it's not restarting from scratch, either.

Can you explain your art production process from the beginning? It can be detailed if possible, because I think people would be very interested to know. For instance, concept art exists -- is it then done in Photoshop, or do you have your own type of programs you use? Or, for animation, do you use a 3D model that you map 2D sprites onto?

GK: The very basic process begins with me messing around with whatever I like, then coming up with a screen visual and building the game around that.

How does the process of creating one full animated sprite work? What is the progression?

GK: There is a proprietary character editor that we use in the studio, a package that took about a year to develop into its current state. The editor's largely based off Flash and other well-known packages, so it doesn't have a major learning curve for most people.

With KOF, the artists make a 3D model for the animation and draw art on top of it. Do you do things like that, or do you just draw it dot-by-dot from a storyboard or something like that?

GK: I take the more traditional approach like that, yes, starting with a basic rough character sheet and working from there. It's a process of gradual refinement. I create a set of basic poses for the character, standing poses and so forth, and I hand that off to the animator/designers and tell them to make it look as "cool" as possible.

Muramasa: The Demon Blade

The animation style on the larger enemies is very consistent across Vanillaware games. It does feel a bit Flash-like at times. I'm wondering if that's a conscious style you've chosen, or if it's just because of the tools you use.

GK: Certainly, part of it is because the toolset that produces the graphics is heavily inspired by Flash. Some of the enemies we create look 3D, but are actually all handmade. We call it tebineri, or hand-shaping.

Why does Vanillawave have such a commitment to stick to 2D?

GK: Well, because we like it.

It's nice that you can actually make that happen. 3D keeps advancing and getting more beautiful, but 3D still can't match high-quality 2D in terms of detail. But people are still afraid to do it, so it's nice that you can.

GK: Thank you very much. However, when you're talking about realism, that's one area that 2D couldn't hope to match 3D in.

What I've been waiting for is for someone to push 2D even further. Even now, Muramasa and KOF -- they're on the top-end of 2D games, but it seems like it's possible to get even more layers of detail and get true hi-res 2D at this point. I was wondering if you think that's possible.

GK: I would like to try and make that happen, certainly.

Do you have any current plans to move into hi-res as a company?

GK: We do. We're in the experimental stage on that right now.

What, for you, are the major difficult points of these experiments?

GK: The most difficult issue to deal with is the fact that current platforms aren't developed with 2D image generation in mind. They're geared toward automatic 3D generation, so we're coming up with ideas to figure out how to facilitate that process. They may not come to fruition immediately, though.

When you're creating hi-res 2D, it seems very high-risk, because if you decide you need to throw something away, then you've lost many months of work. If something doesn't work, you've lost a lot of time, while with 3D you can probably reuse it somewhere.

GK: Certainly. It depends on your development path, too, of course. We're definitely aware of those and other risks involved with the process, and we're always thinking about how to streamline our development to be as efficient as possible.

It seems like you have to be very clear about what you're going to do; you have to determine a plan and stick with it. It doesn't give you as much opportunity to experiment with different ideas.

GK: Indeed, you're always going to go through something of a trial-and-error process whenever you're trying to create something new. That's unavoidable. You have to balance this trial-and-error stage with actual development progress, or else you're simply throwing money down the drain. Of course, if we didn't experiment at all and just went with what we knew, our fanbase would get bored pretty quickly. We wouldn't evolve at all.

It may be a difficult question, but I heard that Vanillaware had to finish Muramasa before you felt that it was ready to go. Is there any possibility of upsizing the graphics, as you said, and releasing Muramasa 1.5 on XBLA or something like that?

GK: Well, you could say that we're officially finished with Muramasa at this point. We did have to cut out four bosses and a fair bit of the story, but if you asked me whether I want to make Muramasa all over again with that stuff included or not, I'd say that if I had that sort of time to work with, I'd prefer to devote it to some new idea instead. There are so many things I want to make, still.

What do you think about the downloadable platforms, like XBLA or PSN or DSiWare?

GK: I'd love to tackle them. I'd love to, but finances wind up being an issue still, at this point. We're a small company, and from an income standpoint, we can't really devote our full staff to a downloadable project yet. We'd run through our money in a flash.

Well, that's why something like Muramasa 1.5 would be a good idea, no?

GK: Ahhh, now I see what you mean! (laughs) Still, I don't think I can see myself doing that. If Muramasa sells a million copies and I get all the royalties from it, then maybe I can think about it. (laughs)

It's kind of funny that you'd say that about downloadable, because most smaller teams think that it'd be easier to release something on a download platform and get a quick return on it. Why do you feel it's not as possible to make money there?

GK: Well, ahead of any of that, the only real dream that my staff and I have is to keep making games for as long as we can. Me, personally, as long as I can keep myself fed, I don't mind devoting the rest of my time to game development. If we have enough capital to explore it, I would absolutely like to try the downloadable marketplace sometime in the future. Even if I made a ton of money off games, I'd just put it back into development in the end, after all.

I was just responding to the idea that you mentioned, unless I misunderstood, that it's not as easy to make money for a small team on the download platform -- because many small teams I've talked to actually think it's much easier to get a return on investment. I was wondering you may feel that way. Not that I'm suggesting you're greedy or anything!

GK: Well, we'd certainly like to become a big-name publisher sometime, but when it comes to game development, no matter what platform or service you're aiming for, you need some kind of monetary base for it. In our case, that comes from our clients, the publishers -- and if you want to catch the attention of the publishers and put them on your side, then you need to think big. That's largely it.

We funded the development of Muramasa completely off the royalties we received for Odin Sphere, which means that we technically earned zero profit off Odin Sphere. (laughs)

As you say, you want to eventually publish your own games, but that's what downloadable platforms offer to you. For example, [with your DS game] Kumatanchi, I imagine that wasn't all 21 Vanillaware staff making that game. If you could find some people to work on downloadable stuff, you could become your own publisher, you could get a more consistent revenue stream going. Just a suggestion. (laughs)

GK: Well, It's not that I'm against downloadable, keep in mind. In the ideal case, I'd like to grow Vanillaware up to the point where I was not the main guy responsible for every game we were working on; once that happens, I could devote myself fully to the ideas I find the most personally interesting, and that's where the download market might make more sense.

It does seem quite possible, especially with 2D. Have you played Castle Crashers?

GK: I have, yes.

It broke the million mark [on leaderboards]. It's possible, anyway.

GK: Yeah. Oh, it's very impressive, definitely.

Will Vanillaware be doing any more DS projects?

GK: Not for the time being, no.

Can you say anything about what you're working on next? Will it be hi-res?

GK: Well, we're experimenting with hi-res. I can't say that the next project is going to be hi-res, but we are experimenting. We don't have a lot of money, so we don't have a lot of leeway with our company's time.

Well, I hope that someone will give you enough money to do that sort of thing.

GK: Oh, yes!

We've been talking for about half an hour -- how much more time do you have?

GK: Could I--could I talk a little about Vanillaware's dream?

Yes, please!

GK: Well, I worked on Fantasy Earth for Square Enix a few years ago, and a lot of us on the team want to try our hand at another online MMO project -- 2D, 3D, whatever. We've got Fantasy Earth II running through our minds, all of us.

Before that, though, we'd like to make a few 2D online games first. To make that happen, we need money, but money has a habit of getting spent pretty quickly around the company. (laughs)

Well, maybe you can propose to Square Enix, because they are continuing to do more online stuff.

GK: Ahhhhh, I dunno. (laughs) We didn't exactly part on amicable terms. (in English) No, Square Enix good company!

Mmm, that's too bad.

GK: It was really a matter of them taking it [Fantasy Earth] from us. Moving on, though! (laughs)

Have you ever played Dungeon Fighter Online?

GK: I haven't heard of that one. Is it recent?

Two or three years old. It's a Korean MMO similar to Capcom's Advanced Dungeons & Dragons from the arcade. It's very popular in Korea and it's getting launched in America now. It's a side-scrolling 2D MMO.

GK: Oh! You mean Arad Senki [the title in Japan]! Oh, yeah, I've played that. It's pretty safe to say that if we did a net game, it might look something like that.

It seems very natural for Vanillaware.

GK: Definitely. If I had a chance, I'd love to make something that surpassed that game.

What other MMOs are an inspiration for you?

GK: I don't know if I'd call it inspiration, but I love playing Warcraft and StarCraft online. A lot of the core Odin Sphere staff were really into StarCraft, sticking around the office to play it after work and everything. We wanted to put some RTS components into Fantasy Earth, even, but it didn't work out in the end with that project. That's a field I'd definitely like to revisit at some point.

Did StarCraft play any role in your creating [PS2 RTS title] GrimGrimoire?

GK: Well, GrimGrimoire got its start when we all said "Hey, let's make StarCraft" to each other. (laughs)

You should try to release it in Korea, then.

GK: We did, actually. Sony published it. I think it sold like 500 copies. (laughs) Odin Sphere did about the same. The ad copy went on about how popular it was, though!

I bet many times that number of people played it over there, though.

GK: Yeah. It's hard to say.

I can tell you that I've seen it available to buy on just a disc in Yongsan, the electronics district of Seoul, so -- Sorry! (laughs)

GK: Aw, you're making me cry here!

If you do some kind of net game, do you feel it has to be an MMO with community, or more like a StarCraft type of game, where it's one-on-one, or a party against another party?

GK: I've been involved in projects that took both approaches, so it's hard to say, but it's the RTS genre I'm interested in, so it'd likely be a competitive sort of online game. MMO makes for a better business plan, though. I have an interest in both styles, really.

I always have a lot of ideas buzzing around in my mind, but publishing another online game's definitely one of my goals -- and from a business perspective right now, it's one of the more feasible goals to reach.

You know, if you do that, then you have to grow a lot as a company. You need customer support, the servers, and constant community management. You may have to grow, like, five or ten times as large.

GK: Yeah. Dreams are faraway things, after all.

You can bring it closer!

GK: (laughs)

Personally, just to tell you what I'd like to see from Vanillaware in the future -- I'd like to see a pure action game of some sort.

GK: That's one of my favorite genres, too. I've been involved almost entirely with RPGs or action RPGs up to this point, but action is where my real wealth of knowledge lies. You can definitely expect an action game from me.

The action in Muramasa is quite good...

GK: Thank you very much.

Tell me more about those dreams of yours.

GK: Do you mean Vanillaware's dream? My own dream -- well, I wouldn't call it a dream, really, but all I want is to keep on making games until the day I die. I kind of wonder sometimes how many of my ideas I'll be able to put out.

What, then, is the ultimate game that you personally want to make?

GK: The ultimate game, huh? Hmm. I don't know if I have an ultimate game in mind. I have a bunch of ideas buzzing around in my head, and I know I'm never going to have the time to make all of them a reality. I'd like to get to as many as I can, though, constantly upping the quality as we did from Odin to Muramasa. Vanillaware may be a pretty poor company, but I'm happy with it because it gives me a chance to create what I want to create. The staff complains at me sometimes, though. (laughs)

Treasure has kind of a similar idea going. "We don't make much money, but at least we get to make what we want to make," sort of philosophy.

What is the feeling that you really want to get across with your games? Because they seem to be building in a certain direction. Do you have some kind of personal goal that you want to reach, aside from continuing to make games?

GK: Well, my definition of the ultimate goal would be a game that implements everything I like about video games -- a little RTS, a little action, a decent story. If I can find the perfect mix of every aspect I like and be happy with it all, that would be my ultimate title.

It could be quite difficult to integrate all of that without confusing the player, so good luck.

GK: Oh, definitely.