You're confusing the mechanics of how the images are computed with the methods the finished images are seen by the viewer.
Think of 3D glasses, images hovering in air. That's what he means.
Although, what rules and code they use to artificially impart 3D depth to an already-flat image...? Hell if I know!