Virtual Reality Bowling

3D Stereo Movie

3D Stereo Movie

3D stereo immersive short film

In this project an immersive 3D stereo short film was created using both real live footage and CGI animation. The 3D stereo in the animation part of the film was created using a custom programed 3D stereo plugin for Maya allowing two cameras to be mounted on the same axsis at all times, turning in relation to each other to simulate a binocular spectrum. The real live footage requried a custom made rig, mounted on a tripod to achieve the stereo setup for the cameras.

The film was rendered for two different stereoscopic methods: anaglyph and LCD (liquid crystal display) shutter glasses.

The LCD method runs on active stereo infrared controlled LCD shutter glasses and the nVidia system to alternate left/right eye images on a CRT (Cathode Ray Tube) running at 120 Hz.

The anaglyph method uses colour filters to limit what each eye sees. The most common colour combinations are green/red or cyan/red.  The images for each eye are then combined into a final composite image; this stereoscopy technique uses multiplexing in space. 

Special interests:

3D modelling and animation, character representation & expression, immersive techniques and 3D stereo

Programs and tools used in development:

Alias Maya 7, Premiere, LCD shutter glasses, anagplyph glasses, 2 Sony HD DV cameras, custom made 3D stereo plugin for Maya

Immersive technologies

Immersive viewing technologies rely on maximising the field of view (FOV) of the audience.  The most common examples of this technology are panoramic screens, Head-Mounted Displays (HMD) and 5 and 6-sided caves.  Panoramic screens generally afford a FOV of up to 160 degrees.  A 6-sided cave allows a total sense of immersion in the projected environment.

What is 3D stereo?

The human eye and brain interact to perceive depth, based on different kinds of visual cues.  Depth cues can generally be divided into two main categories: Monocular cues and Stereoscopic cues. Monocular cues can be perceived by anyone with “normal” vision.  In this case, two eyes are not necessary as the monocular cue appears in the same way to either eye.  They provide the viewer with information concerning light, size, perspective and relative placement of objects, backgrounds and foregrounds.  This is achieved through the use of light and shadow, occlusion of distant objects, perspective angles, relative sizes of close and far objects and so on.  While monocular cues give the viewer an understanding of depth, they do not provide the true sensation of depth, which stereoscopic cues provide.  They do, however, provide essential visual information allowing an effective 3 dimensional representation of objects in 2 dimensional images and moving images.  In other words, they allow us to understand that certain flat images are representing objects which in reality have a third dimension of depth (3D).

In contrast to monocular cues, stereoscopic cues are based on the different information received by each of the eyes.  This method of perception is called binocular vision. Each eye perceives the world from a slightly different perspective viewpoint, which causes retinal disparity, a difference in the images that appear on the retina of each eye.  The distance between the average pair of human eyes is about 6.4mm, or 2.5 inches (Ibid p 8).  A 3D model is created in the viewer’s eye-brain based on the separate information received by each eye. 

In the case of normal stereo vision in the real world, the visual perception system uses a combination of two completely unrelated perceptual functions: accommodation (focus) and convergence.  These two functions use completely different sets of muscles. 


Accommodation, or focus, is achieved through the use of ocular sphincter muscles that change the shape of the lens.  By making the shape of the lens fatter or slimmer in the middle, the eye’s focal length is changed. Conversely, convergence changes the rotation of the eyes towards each other for close objects (in the case of becoming cross-eyed), or to a maximal convergence angle of parallel vision for distant objects.  Any rotation beyond parallel is called divergence and is usually unnatural and uncomfortable.

Stereoscopic cues for stereoscopic images are all based on what is known as parallax.  Parallax is a general scientific phenomenon, and in regards to stereopsis is fundamental in producing 3D Stereo images.  In this case, it is the distance between corresponding points of a stereoscopic image if they were to be overlaid.  Parallax results from capturing images from two different viewpoints, relative to a background or foreground.

When viewing a stereoscopic image, the degree of parallax (spatial separation) between objects represented in left and right eye views allows objects to have the appearance of being behind the viewport, in front of the viewport, or flat on the viewport (as normal 2D images are).  The names for these perceived object depths are:

Positive Parallax – objects behind the viewport, further away from the viewing
Negative Parallax – objects in front of the viewport, closer to the viewer
Zero Parallax System (ZPS) – objects flat with the viewport, at the viewport depth