There's been an explosion of interest in 3D recently with 3D CGI hitting the big screen (e.g. Toy Story, Monsters Inc, Up, MegaMind etc), computer games, the use of stereo cameras for broadcasting sport and live events, stereo goggles in cinemas and the introduction of 3D TV sets. 3D stands for three-dimensional or the third dimension and in media terms its desireable because the real world is three dimensional and anything that approaches this level of realism can't be a bad thing because we get to use more of visual system which evolution has spent millions of years perfecting. This article will give you a basic understanding of how 3D vision works.

At this point we should disambiguate 3D content with 3D presentation. 3D content, as popularised by World of Warcraft and Call of Duty utilises scenes and characters that were created with 3D modelling software but this content is usually presented on a 2D screen. The graphics look better and more appealing because they can display realistic shading, textures, motion parallax and perspective effects. With 3D presentation this 3D content takes a massive leap forward in terms of fidelity and therefore believability because the depth dimension is presented with one of the most powerful visual abilities we have, stereo vision or stereopsis. This article will be concerned with how we see depth in 3D presentations.

What is Stereo Vision?

Stereo vision is the ability to use 2 different views of the world in order to extract depth information. We have 2 eyes which have around 65mm distance between them measured from pupil to pupil. When we look at an object in depth we fixate on it: that is, the two eyes converge and the object is brought to focus on the fovea (the area with highest sensitivity) at the back of each eye (a light sensitive surface called the retina). Objects in front of and behind the object of fixation will also be pictured on the retina BUT at different distances and directions from the fovea depending on how far in depth they are in front or behind the object of fixation. This difference in retinal distance is called binocular parallax and it gives the brain a direct measure of relative depth.

So how does a 3D presentation work? 3D presentations are created either by using two cameras suitably positioned left and right like the two eyes of a human, or they are shot digitally and software is used separate each frame into left and right stereo images. The left and right stereo pairs are then combined into a single image. How they are combined depends on how they will be viewed. If, for example, the presentation will use the old-fashioned red/green anaglyph glasses the left image might be filtered red and the right green.

When the combined frame is viewed you will see a jumble of coloured lines. However, wearing red/green glasses with a red filter on the left lens and a green on the right the left image will only be seen by the left eye (the right eye filter won't let the red content pass through) and silimarly for the right eye. As the two images are separately fed to each eye, the retinal disparities between the images in the two eyes will be interpreted as depth by the brain.

Other technologies also exist for stereo viewing based on stereo pairs as used in anaglyphs. For example Shutter Glasses open and close each eyes view around 60 times a second.

The motion picture success of avatar was supported by its use of 3D cameras and a viewing system based on polarising glasses from RealD. The principle is again the same as anaglyphs. The left eye filter allows the left camera view to be seen and not the right and the right eye filter ensures the right camera view is seen and not the left. The brain combines the two images and derives depth from the ensuing disparities on the retina