For three years, researchers in the Camera Culture group at the MIT Media Lab have been developing a design for a multiperspective display for 3D video that does not require special glasses, which they hope will be a practical, if short-term, alternative to holographic video.
Using this technology, they designed a projector that they will present at the Siggraph conference later this year.
The MIT Media Lab 's projector can also improve the resolution and contrast of conventional video, which would make it an attractive transitional technology as content producers gradually learn to exploit the capabilities of multiperspective 3D.
The multiperspective 3D differs from stereoscopic 3D – which is now common in cinemas – in that the objects depicted reveal new visual angles as the viewer moves around them, exactly as would happen in the real world. This means that there could potentially be applications in fields beyond entertainment- such as architecture or medicine.
Gordon Wetsstein, Matthew Hirs and Ramesh Raskar built a prototype using simple materials that are commercially available. They also created a prototype of a new type of screen that widens the angle at which the projector images are seen. The screen combines two special lenses, similar to those used to create the impression of three‑dimensional effects in old children's books. For each video frame, a total of eight different viewing angles are produced.
From an algorithmic perspective, «key» in the system is a technique that calculates how much data can be retained between the different viewing angles and how much must change. Retaining as much data as possible allows the projector to create a brighter image, and the final set of angles and light intensities is encoded into the projected patterns, via a specially configured (based on the architecture of graphics processing units used in games) algorithm that runs almost in real time.
