University of
	Bielefeld University of Bielefeld - Faculty of Technology - AI Group

VR Lab NG construction site

This page loosely documents some of the stages during our major lab reconstruction. If you are interested in some of the problems to consider when installing a large screen projection system, the following pictures will give an idea about that. The pictures roughly follow the time line of the reconstruction and start with a small blueprint and a couple of 180 degree shots that give an overview of the labs layout.
The blueprint of the lab. Spot 1 is the camera spot for the following 180deg pictures. What is found "left" in the blueprint will be found "right" in the pictures as will be clear for all of you people with a "CG & VR trained 3D imagination" ;).

Almost the beginning. The responsive workbench (formerly on the right) has moved to the other room (on the left). The old electromagentic tracking system is still in the place.

Cabling and painting (we have to reduce light reflections) has begun.

Its getting dark. Almost a cave without screens.

Now there is a floor and the framework for the screen and camera attachment.

The screens arrive. They are optimized for the use with circular polarisation filters and are made of laminated synthetics. As you can see, they are a little delicate to handle. Their size and weigth together with the "don't touch the surface" generates some problems while we try to move them in the lab (no elevators are large enough and the lab is on the 4th floor). The same is true for most of the doors. In this case, we just had to break the walls to the lab to allow the screens to reach theitrfinal destination.

Ups, almost forgot to set up the subsonic sound drivers on the floor. Thx goes out to the guys of the installation team to re-lift the quite heavy wood/aluminium construct once more. And then: Let it be light - well at least to test the floor projection quality for the first time. We prefer the atlantis demo to get into the right mood ;). Brightness and contrast of the floor front projection is quite good and the first test did satisfy our expectations.

From left to right you can see the back projected front screen from the back, the foldaway construction for the left screen to change the tri-space to an l-shaped projection system, the left screen from the back and finally all screens together from the front.

The mounts for the video projectors. Each screen needs two of them (left and right eye). Left and front screen are back projected. The left screen is directly driven (no mirrors), the front screen needs a mirror due to the limited distance between projector and screen in our setup. The last picture on the right shows the floor projection (with one projector missing during installation) wich uses a mirror under the roof.

Yeah, there we are. The first thing to do: test the quality! As you can see from our sceptical facial expressions we are searching for tweaks and small inconsistencies ... but we did not find too many. So we could finally enter into beta mode!

Here we see MAX, our embodied conversational agent, standing the first time in our virtual environment. Success! Of course, a project like this new lab will never be really finished totally. There are still hard- and software problems which we encounter and which must be solved. But the overall system in general has now moved into "production state" (as far as you have a production state in an ongoing research area). Most of our project demonstrations now take place in the new lab and utilze its facilities. Our webside will inform you about ongoing research work.


Marc Erich Latoschik (E-Mail:

Back to the homepage AI Group

Marc Erich Latoschik , 2003-01-17