top of page

Making a 3D Arcade eye-tracker game that doesn't suck: A camera problem.

For the last 3 mounths of my first year in game-school, I worked on an eye tracking game named Patin Grappin.

It's an ice skating game where you can throw a grapple at any object visible on screen, and just enjoy being some kind of spiderman-person in the air, or just enjoy ice skating on the ground.

The whole thing is playable with just one button, to trigger the grapple, the rest is played with head and eyes : head rotation to move the player, gaze to determine where you're gonna throw your grapple.

The eye tracking part works incredibly well. And when I say that I mean :

I thought it'd go very badly and it worked out okay and I think it's incredible.

From what i've seen, games that are specifically designed for eye tracking devices are usually games with a static camera, maybe 2D graphics.. things that don't move all the time. Appart from that, AAA games like Hitman uses eye tracking as an assistance for aim, for example.

We wanted to make an arcade 3D game with lots of speed, that would be designed to be played with the eyes. It was probably the worst idea but that's what we wanted to make.

We encountered a lot of problems due to the eye-tracker: Anything can break the technology: people with glasses who can sometime play the game, sometime not (as they need to move their head up and down, the device has trouble to deal with the glasses not being oriented always the same.), callibration problems, etc.. Even a simple light change got my day of work ruined: I was sure I was looking at the objects, but because of the new lighting, my gaze would focus around the objects but not on the objects themselves. Took me a while to figure that out..

...But the main problem that we faced really quickly was the camera : in a 3Dimenssional space, with huge speed, the camera had to move a lot. Which meant that, if the player was looking at an object on the screen, the object would move with the camera and it will be super hard for the player to focus his gaze on the object.

The ice-skater movement could have cause issues with the camera for example: at first, it was making the objects change radically on the screen when you moved the character aroun, as the camera was following the character.

When we found our head rotation mechanics, the problem kinda sloved itself : First, the player was not moving the character around, but changing its direction with small head movement. Secondly, when testing the mechanic, we found out that the players would constantly compensate by themselves the camera, by orientating their head just a tiny bit. Just like when riding a car, making constantly sure that it's not going off the road. That way, player have a precise and direct control on where the objects are on the screen. That would also mean that we should not make the player turn a lot, and if we do, that'll be thanks to the grapple.

The second camera issue we had was that, when the player was in the air, they would fall and that would, of course, make the camera move, and the player obviously can't control that. It made it impossible for players to grapple an object while being in the air, so could not have the spiderman-feeling that we wanted to give them.

As a designer, what I did to fix this quickly is that I took the lerp parameters from my camera and took their values down drastically so that the camera would move as slowly as it could.

And it worked !

The players could have their spiderman-like experience, grappling in the air and in the air again !

….But the character was not on screen anymore until the character hits the ground ! The camera, of course, couldn't catch back the character.

(Woops, where's my character?)

And it was not that much of an issue actually, it stayed like that for a whole mounth and from what we saw during playtest, the player didn't really noticed. But it made it odd to see someone else playing, with no character on screen.

We could have left it like that, but as we noticed that the camera would always be super close to the character, we choosed to make it move just enough to make the character be always visible.

At the end, what always made the eye tracker work in patin grappin was that, when we added a feature, even a simple one like movement or being able to move the camera, we tried to find the control that would help the eye-tracking part of the game. Help the player's gaze stay on the objects on screen. And then focus on making that control feel good as well.

We didn't want the eye tracker to be just a funny feature for the game, but to be a real part of it. We wanted to help our player with their gaze point, because it was already a hard task to ask.

Hopefully the headtracking and slow camera motion were good solution enough and for most players it worked out.


Playful stupidity.

bottom of page