I think you may be blending two senses of focusing here...
depth of field is caused by the lens in your eye or camera adjusting for one
distance, and objects nearer or farther away being blurry depending on
aperture (F-stop). So a conventional computer 3d display is like a pinhole
camera--effectively zero aperture size and infinite depth of field--all
objects are in perfect focus (although mipmapping and filtering can give
some depth-of-field like effects within the surface of a textured object)
The other effect of depth is the two images being out of place horizontally
(a la beer goggles) due to parallax between the two eyes (stereoscopic). 3D
glasses of various sorts can give the stereoscopic depth effect, but not the
focusing effect. The 3D glasses I've tried are problematic because the
binocular depth information from your eyes is in conflict with the focus
information. However, I think if the screen is "too far away" rather than
"too close", the effect is better, such as at an Omnimax with 3D glasses.
I believe there is a technique where you have a reflective "subwoofer" that
flexes at a pretty high rate, with an image projected onto it from a device
capable of insanely high refresh rates (such as vector plotters)...
basically the woofer acts like a variable convex mirror, so the image can be
layered in depth. This gets you both 3D effects with no glasses involved.
The problem is to render 60 FPS with 100 depth layers requires 6000 frames
per second! I think most demos of this have spinning wire-frame cubes and
such, but one could conceive of some *** new way of doing this with
full-motion 3D graphics, such as a graphics card that could render triangles
to multiple depth planes simulaneously, etc, etc. (Caveat: my knowledge of
this device is extremely sketchy)
However, I think games still have quite a ways to go to achieve even 2D
photorealism...
