Glasses-free 3D display for theaters
Three-dimensional films have become massively successful since 2000, and audiences are more satisfied with 3D movies, which provide more vivid and realistic scenes than 2D versions. It is not surprising that almost every movie is now produced in 3D. An additional incidental benefit is a reduction in illegal in-theater recording of movies. However, there are several reasons why some people are against 3D movies, in particular the discomfort of wearing special glasses. Yet, even though glasses-free 3D display technology has been commercialized for personal devices, such as the Nintendo 3DS, significant limitations must still be overcome for theater-scale projection.
For glasses-free display, special optical parts must be installed in front of the screen, which interrupts original images from a projector and is incompatible with conventional theaters. Until now, the only practical option for large-scale 3D display was using high-definition projectors and wearing polarizing glasses.1 In fact, many IMAX theaters have set up two projectors instead of one, and covered the set-up cost with additional ticket prices for 3D movies: see Figure 1(a). It is possible to realize a glasses-free 3D theater display by positioning the projector behind the movie screen: see Figure 1(b). However, such an approach is not feasible for existing movie theaters, because they have insufficient space behind the screen. To meet this demand, we have developed a glasses-free, front-projection 3D theater display system.2
Our method locates the projector behind the audience, as in existing theaters: see Figure 2. We employ a special optical film array of transparent films in front of the screen. It consists of a polarizer with slits (like venetian blinds but vertically oriented) and a quarter-wave retarding film (QWRF) to make the audience see different directional views. A QWRF is a polarizing film with a ‘fast’ and ‘slow’ axis at 90° to each other. When it transmits light, the waves polarized along the slow axis will emerge from the film a quarter-wavelength delayed relative to the light polarized along the fast axis. In this way light polarized at 45° to both the fast and slow axes can be transformed into circularly polarized light.
Projected light is polarized by a filter as it leaves the projector, for instance along the y-axis. It passes unchanged through the parallax barrier polarizer, and becomes left-circularly polarized as it passes through the QWRF (which has its fast axis at 45° to the x-axis). The light is then reflected by the movie theater screen, which changes its polarization state to right-circularly polarized. As it passes through the QWRF a second time (this time with the fast axis at 135° to the x-axis), the projected light becomes x-polarized. It is then blocked in some directions by the y-polarized parts (the ‘vertical venetian blind slats’) of the parallax barrier. In this way, the audience sees simultaneously separate left-eye and right-eye images: see Figure 2. In other words, after reflection by the screen, the polarization state of the image is orthogonally changed and so the slits in the optical film split the image into several signals for the different eyes of the audience.
The advantage of the general concept—directional control of the light signal via double transmission through a QWRF and polarization changes—is that we can set up the optical film array as either a parallax barrier (as I have described) or a pinhole array for integral imaging.3 The optical film array can also be replaced by a polarization-activated lenticular lens, increasing optical efficiency (however, a lenticular lens would not be appropriate for use in theaters, as the 3D effect requires the viewer to be in a particular location). Our method requires only optical film, and so can be implemented by compact, passive devices. It is thus both space saving and cost effective.
We have further extended our approach to improve the quality of the 3D image and express real scenes by capturing real 3D objects. The disadvantage of the method we have described is that it blocks some of the light to the audience's eyes; currently the reconstructed 3D image is fairly low resolution. Temporal multiplexing and using a high-definition projector can improve the resolution.4 Another approach will focus on a passive polarization-activated lenticular lens array that could also enhance optical efficiency for personal uses. We are also now working on a real-time multi-camera 3D video capturing method.5
Youngmin Kim received his PhD degree in 2011 from Seoul National University, Republic of Korea. His research focuses on interactive 3D displays, holographic displays, and visual fatigue associated with 3D displays.