CALL me a creature of habit, but I approach any new computer interface with a sense of apprehension. I鈥檓 downright inept when it comes to playing video games on the Nintendo Wii: the wand controller is just too foreign to my mouse and keyboard-entrained muscles. I feel that familiar sense of unease as I stand in a nondescript brick warehouse in downtown Los Angeles.
I am at the headquarters of , developers of the G-Speak gestural computing interface, and I鈥檓 about to trial its system for controlling computers through hand gestures.
I find myself surrounded by a cage of metal scaffolding, which houses the system鈥檚 16 near-infrared motion detectors, as John Underkoffler, Oblong鈥檚 chief scientist, boots up the system. I鈥檓 amidst three large screens, and above me three projectors beam images onto them. A fourth overhead projector, pointing onto a white table, serves as a fourth screen. Underkoffler insists that the G-Speak is targeting hardcore number-crunchers, not gamers, but the rig looks like it would be more at home in a rock club than an office.
Advertisement
Underkoffler hands me a pair of black gloves with tiny reflective balls attached to the back of every digit except the pinky. The gloves help a camera follow my hands but add to the feeling that I鈥檓 about to start a performance. Yet again, Underkoffler stresses this is serious computing, before adding that: 鈥淭he goal is to get rid of the gloves entirely, and we鈥檙e not far from that.鈥�
First up was a basic training program. A grid of white letters hovers against a blue background on the screen in front of me. It feels playful when I鈥檓 told to form a 鈥済un鈥� by extending my index finger and pointing my thumb upwards. But this hand position forms the basic shape for interacting with all on-screen objects, allowing me to move a star-shaped cursor around the display. I 鈥渟hoot鈥� the gun by depressing my extended thumb and in so doing grab a letter and move it with a gesture to the table in front of me.
So far, so Minority Report. Having mastered the basics of point and shoot, I move on to a 3D application. Here I go from shooting to flying. On the bank of screens, a universe of regularly spaced boxes extends in all directions. To navigate, I start with the gun hand and bend my middle finger so it鈥檚 at 90 degrees to my index finger. This action creates three axes on screen. Pushing up, in the direction of my thumb, moves my position up; moving my hand sideways, or forward and backward, has a similar effect. Twisting my hand rotates space.
After a bit of flailing, I gain control. I learn to target a particular box and head slowly towards it in a smooth spiral. Then suddenly the flight jerks to a halt, my gestures no longer in control. Underkoffler points out that I have let my extended fingers curve just a bit. The system tracks my fingers to within one-tenth of a millimetre in all directions 鈥� it feels surprisingly sensitive to my hand, which is used to the imprecision of a mouse. I straighten my fingers and I鈥檓 off again.
We move on through applications for video editing, photo analysis, even air-traffic control. By the end, I am really starting to appreciate that G-Speak is meant for real work: the interface allows me to sort vast amounts of on-screen information with far greater ease than I could with a mouse. Vast sweeps of my hands create rapid movements, but the system is sensitive enough to allow precise movements to be picked up too.
Will gestural computing become a mainstream technology? Perhaps, but this set-up is hardly going to fit into the standard office-worker鈥檚 cubicle. Oblong promises desktop versions will soon be ready to demonstrate, though. Mastering it would take practice, but probably no more than touch-typing, and it鈥檚 a lot more fun than that.
