My alt.GDC series came to a close on Friday, but I still have one story I’ve been wanting to tell. In my day job, I work as a game designer for a company called Toys for Bob which recently released the wildly popular game called Skylanders: Spyro’s Adventure. The game comes with an external device (called the Portal of Power) that communicates with your console and action figures that can be identified by the portal and brought to life in the game. You swap out characters as you play by changing the character on the portal. Players can also expand their selection of playable characters by collecting more toys.
Before I went into GDC on Friday, I swung by the Toys for Bob offices in Novato, CA and sat down with Robert Leyland, the technical engineer behind the Skylanders hardware, and I-Wei Huang, the character and toy director (also known to MAKE readers as magazine and site-contributor CrabFu), to discuss the process of developing this innovative title.
MAKE: What can you tell us about the early development you did on the Skylanders concept; where did you start? Robert: We started with a bunch of prototype ideas… When the Wii came out, it had these new controllers and people started thinking: We could make more devices, more accessories, and you saw this with things like Guitar Hero, Wii Fit, the Tony Hawk skateboard. So we wanted something that would connect to the Wii to create a physical device.
There was a period there where we said “But how are we going to connect it? Will we need to mount a Wii-mote in it?” And right around that time, on MAKE or Instructables, I can’t remember where – we saw it on Gizmodo, somebody published a hack about how to use a Wii Nunchuck to control an R/C helicopter and we thought: Whoa, that’s really pretty cool. And we looked and sure enough they weren’t just wiring to the potentiometers in the controller, they were using the output signals and there was a bunch of software published that showed that you could get the signal data from the Nunchuck controller. And I thought that was neat that they’d done that, but I also thought, if I can pretend to be a Wii Nunchuck controller I can talk to the Wii and we can go the other way.
So a few weeks later, we have an Arduino and a circuit board and some wires, and on the Arduino I’m pushing a button and its going through the controller and being registered in our script as a button press from the Wii. So if we’re controller 2, we can talk to the Wii while the player’s controller 1. We did a few tests and proved that it all worked. But then we talked to Nintendo and found out that there’s only about 15mA of power available to the controller which would barely run an LED, let alone a CPU or anything else, and there were way too many restrictions so we thought we’ve got to find some other way. Eventually, we ended up going through the USB port. And then somewhere along the line, somebody came up with the idea of “making toys come to life” and we thought: That’s good, that’s fantastic, in hindsight the best idea we’ve ever had. And we looked at different ways of connecting the toys. We looked at little contacts on the toys, different RFID techniques, and we finally settled on one of those, partly because it was an ISO standard, and partly because there were open source libraries published to access the RFID controller.
So, we got an RFID controller from Turkey, the Arduino from Italy, and the USB connection which was wired here in California with a software library from some guys in Germany – you can just pull all of these pieces together and gosh, you can quickly prototype something which is really neat.
The first connection we had was only 5 bits; we were pretending to be a keyboard and whenever you got a keystroke it said “Oh, that’s key X, so that’s Spyro” or “Key 7, that’s Bash.” So with those individual keystrokes you could essentially see a character, but we knew we needed more IDs than that, so we eventually wrote a protocol that would send longer data sequences. And it evolved from there.
MAKE: What kind of hardware did you use to build the original portal prototype? I-Wei: We spent a lot of time going through hardware stores, office supply, and craft stores looking for things we could use. It’s fun, it’s really similar to what I do with my steam-powered robot stuff (CrabFu Steamworks), you go into a store with the mind of “I need something that has this shape and size.” It’s a different way of shopping, you’re pattern matching. “What could I find that would fit that look, if I painted this thing, or if I cut this part off, I don’t need this section, but if I take this other part and put it here…” Robert: And it can be any part of the store; it could be the toy section, it could be the kitchen section, it could be the plumbing supply section… In the plumbing section we found a pipe that was the right diameter and flared at the bottom. It was for a toilet! But it enabled us to put a light in the bottom, a sheet of translucent plastic on top and a ring of LEDs around it to create the effect we wanted. I-Wei: And on the bottom where the main LEDs are we used a kitchen sink drain because it had a nice reflective surface, it was already a bowl shape, so the light on the bottom would be like a flashlight, shining all the light upwards.
MAKE: And you were using Arduino for the electronics? Robert: Yea, I think the early prototypes used a Boarduino from Adafruit. Actually even earlier than that I was using two 555 CMOS chips timed together to control the LEDs and they were wired to a “magic moment” button, so in our demos someone would put the toy on the portal, but I was off to the side pushing the button going “pay no attention to the man behind the curtain.”
MAKE: At what point in prototyping the portal technology did you show it off to Activision? I-Wei: Pretty early on. This was the kind of thing that we had to see the idea internally to prove to ourselves that we weren’t crazy, but also if you just imagine Skylanders as a pitch on paper, I don’t think it would have been very successful. But as soon as we made the first hand-made sculpey figure and you put it in your hand, you start to get it, and then having a portal that actually worked, where as soon as you put a character on it you see a character on screen running around a simple world, that’s all you need to sell it and we had that really early on.
MAKE: I-Wei, what can you tell us about the making of these incredibly popular toys? I-Wei: The quality of the toys that are on the shelves today is the result of our ignorance in the toy industry. We weren’t toy makers, so we spent a lot of time going through every little detail with these characters because we wanted a cool toy, not just a good-enough toy for our age group. If you look at other toys in our price range, the amount of paint, the amount of detail isn’t there. You’d say “Oh, that’s a cool toy for a kid,” but to an adult it wouldn’t be as interesting. We just didn’t have that filter of “This is good enough to sell.” We wanted you to be able to look at the toy by itself, apart from the game, and say “Wow, that’s a cool toy!”
At the time we had no experience making toys and the kind of mold-making you need for mass-producing toys so we relied on an external company to make them. We would give them sketches of characters and in-game models, animations, screenshots, and everything, and they would produce a toy that we would then tweak and tune to get the likeness and personality to show through a little bit better and that process could take many many months. We did many iterations of the character and paint and pose and everything.
MAKE: How has this process been improved since the completion of the first game? I-Wei: Now we have a 3D printer in our studio, so we can do everything internally. We take our in-game model and we res it up in Z-Brush so it looks great and has all kinds of detail in it. Then as soon as we have the high-res version we put it in different poses because it’s also an in-game model which already has a rig. We usually try about 3 poses and then we just export it, hit a button and the next day we have a 3D print (It’s actually a bit longer process to prep a 3D print in color). So this time around we have much more control over how the toy’s look. The 3D printer has sped up a process that could take, from final concept to a toy that looks great, around four months. Now we can have a model roughly done in about a week. We’ve greatly streamlined this process.
MAKE: In closing, do you have any advice for getting publishers or investors to subscribe to a crazy idea like this? Robert: When you pitch your prototypes, you want to make sure that they are good quality, they’re neat and tidy. Take the time to put it in a box and package it up so it’s not just written on the back of a fish and chips wrapper. I-Wei: Especially for me, on the visual side that’s vastly important because you’re really selling your idea and you want to make sure that the first impression is good. Make sure everything is as polished as you can make it. It’s harder to sell a tech demo. If it’s a beautiful looking thing and you can grab them by the emotional strings you don’t have to explain the tech. Robert: Another piece of advice when prototyping: Go with standards. Don’t try to own the entire space. The more you can do something unique, find a niche that works with other standards, the better off you are, because people will find ways to use your stuff that you didn’t envisage. It seems obvious, but sometimes it’s not.
More: See all of our alt.GDC coverage