The project was started in 1991 and was based on image processing. We wanted to show that VR is for everybody. The goal is to study the VR arcitecture, i.e., the closed man-machine loop that all VR systems have. This is the same architecture that was envisioned by Sutherland around 1954 (?) and it was first implemented on Whirlwind. The system was a 33M.Hz 386/33 with a Targa image processing board and it was later upgraded to a 486/50. But most of the fun we had was with the original, slower system. If you think 386 was too slow, think Whirlwind and thank God.
Today the computing community has extended the term VR to include photo-realistic rendering, and other aspects of computing.
A camera was installed on the computer which displays whatever it saw. (Usually the player) Then the machine introduced an artificial ball. Based on the movements of the player, the machine interprested the ball movements. To the player, he or she felt the effects of his/her actions towards the ball and reacted accoprdingly with their hands. (The objective is to keep the ball in the air.) The machine perceived the hands movements as input and adjusted the ball movements. Thus a partnership between the player and the machine was achieved. This was the original intent of Sutherland.
The interactions were based on simple law of physics and the players could anticipate the ball reactions from their own hand actions. Occasionally, surprises were introduced. While the ball hit the side wall, instead of bouncing back, it was re-introduced from the opposite wall. Most players adapted the situation quickly and had found it was hilarious.
Following is an .avi file that plays back what the computer saw, i.e., from the computer's perspective, and was recorded on a video tape. Virtual Juggler in action.
After we had gained experience with the V.Juggler, we asked a hard question to ourselves: What more could VR do for us? - which is still being asked even today - 40 years after Sutherland. Armed with a 486/50 now, we were more adventurous, i.e., we thought we coul do some 'real' meaningful work (However little did we knew could be a blessing.)
A V.mouse would follow the hands of the operator. The machine displayed a certain, simple window that had some toggle-switches, some spring loaded switches, and a lone slider. The objects all had a hot zone. If the V.mouse (operator's hands) enters the hot zone, the object was considered selected. If the V.mouse stayed in the hot zone for some time, then the object was considered activated. If the V.mouse left the hot zone then the object was deactivated. Slider slided, switches toggled.
It was during the taping of the final project that an errie accident happened. Taping can be done in two perspectives, from the point of the computer and from the point of the world we live in. The formal from the computer monitor tap-off and the latter just a home video recorder pointing at the operator, the V.mouse. We had found that by feeding the world tape to the computer through its video eyes, we could instruct the machine to perform some steps by taping the operator movements on the recorder. This could be the ultimate in remote control, no tethers attached, no radio signals to instruct, no seeing-eye feedback, etc. The computer would be instructed by an on board tape.
The tape of teh V.mouse is still under format reconstruction. When our students finished their work, we will put it inline here. Meanwhile, we lead you back to look at the Virtual Juggler in action.
Return to Soong's home page.