Stereoscopy, 1st testnight

Post by: Joris Weijdom Add comments
Stereoscopie

IMG_7431_exposure.jpgOn 23rd of August we had our first testnight for the “ProjectieLAB” (Projection Labratory) at the office and workspace of Z25.org Foundation. We have tested a very simple setup with two beamers, polarization filters and glasses and a laptop with the NVDIA stereoscopic drivers. Although we didn’t manage to get two identical beamers and had only quite amateurish means to align them the result was very promising.

IMG_7423.jpgAs you can read in the last post in the category of Stereoscopy we had prepared a small silverscreen by painting a plate of MDF with aluminium paint. Also we constructed a frame for this plate so we could easily change different screen types and have a means to actually hang the screen in the air. Due to our impatience to actually try out some stereoscopic projection we just put the screen on top of the Football Table.

IMG_7425.jpg After extensive research, and based on the experience of Z25 we chose to work with polarization as the stereoscopic form. Mostly because the results are graphically the best in terms of color (as opposed to Anaglyphe), most relaxing for the eyes and reasonably cheap (as opposed to Shutter glasses). Normally you use two identical beamers for a typical setup, we couldn’t arrange this however so settled for two different types of Optoma beamers.

IMG_7426.jpg Calibrating and aligning the beamers in a way that both projections are exactly on top of each other was simply impossible because the lenses of the beamers where actually different. So we tried to align at least the center of the screen. We ended up having a difference in the vertical axis of about 1.5 centimeters both at the top and the bottom of the 1.2×0.9 meter screen. The difference in luminance we could mostly compensate with the menu options of the beamers.

IMG_7427.jpg To have two different images for both beamers and to be able to have realtime 3D stereoscopic images we used a laptop with a dualhead Quadro FX 2500M graphics card with Forceware 94.22 and Stereo drivers 91.31 from NVIDIA installed. It speaks for itself that you don’t need necessarily a laptop or Quadro card. It would be a lot cheaper to have a workstation with a GeForce installed. Only important thing is that the card is from NVIDIA so you can use their Stereo drivers AND that the card has two physical connections for two beamers.

Another reason to go for the workstation instead of a laptop is that the drivers for the Quadrocard in this laptop being a Dell precision M90 is actually released by Dell and not NVIDIA. Point is that you need a Forceware version that comes close to the Stereo driver version. NVIDIA updates it driver versions regularly so you have a lot of choice, Dell however is not famous for its driver updates…

IMG_7428.jpg To perceive the illusion of depth you need n different image for each eye. In polarization this is done by putting two different polarization filters in front of the two beamers. One filter must be placed in front of the beamer lens in such a way that its polarization angle is at a 45 degrees. The other filter is then placed in front of the lens of the second beamer at an -45 degree angle (so there is a 90 degree difference between the two). You can easily check this by putting on the polarization glasses so you can see which eye sees the image of the corresponding beamer with the filter in front. Filters can be bought online easily, you have cheap ones from plastic and expensive ones from glass. For our test we used plastic 3X3 Linear Polarizing Filters for Projectors form Berezin.com (yes, we could have used circular but for a test we decided to start with the cheaper linear option).

IMG_7429.jpg First we runned the embedded NVIDIA 3Dtest application from the advanced screen properties settings in Windows (make sure you clone your desktop and choose Dual VGA Output (Dep3D) from the Stereo properties within the NVIDIA graphics card settings tab). And to our surprise we saw 3D. The NVDIA text seemed to float above the football table in an instant! Even with the crappy alignment of the beamers, the luniminosity difference and cheap filters we managed to create a 3D image that was working without getting a headache.

IMG_7431_exposure.jpg Having tackled most of the first basic setup hurdles we started to play and experiment with a simple object in Virtools. As long as you run the OpenGL window on fullscreen the NVIDIA stereo drivers kick in and you’ll have your own custom realtime 3D stereoscopic images. Their is a lot we have found out in this short test night. For example the angle of the screen should be preferably 90 degrees to the viewer for the optimal light reflection. It was easier to see depth inside of the screen than in front. Make sure your object is not close or hitting the edge of your screen because your brain will immediately loseWeight Exercise the illusion of depth. Also using a black background in the 3D image made it easier to get the idea that the object was floating in front of the screen.

To conclude we have found it pretty easy to create this simple setup and where suprised by the quality of the stereoscopic 3D images. Even with the crappy circumstances that we mentioned. Our next test will be about interaction with the 3D objects.

Comments are closed.

WP Theme & Icons by N.Design Studio
Entries RSS Comments RSS Log in