...stereovisuals¹

By analogy to ``audiovisual''.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... comprehensible²

As is the case when being directly observed by a human.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... system³

This is not to say that they can be ignored completely, but that the system will ignore them for all intents and purposes. The user will be expected to handle them them properly to obtain satisfactory results.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... CCD⁴

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...fig:stereocam3).⁵

As shown in Figure 3 the cameras were strapped together with rubber bands and placed in a small box to stand them upright, however the base of the cameras has a threaded hole which would allow two cameras to essentially be bolted together for better stability. This could even be extended to have an adjustable bolt controlling the separation of the cameras.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... depth)⁶

For example, SGI's webcams can capture any of a ``full-frame'' (640x480), ``half-frame'' (320x240) or ``quarter-frame'' (160x120).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...sec:preliminary-method)⁷

The shutter speed of the cameras must also be set identically on both cameras, as this directly influences the amount of light entering the aperture of a camera and hence the brightness of the resultant image. This isn't a problem, as there is rarely any need to alter the shutter speed anyway, but still deserves mentioning.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... another⁸

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... seamlessly.⁹

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... appropriate.¹⁰

The compression of stereo images as a whole (that is, CompressedStereoImage) rather than the compression of the component left and right images (that is, two instances of CompressedImage or two instances of CompressedStereoImage, each with a single component image) may allow significant improvements in compression to be obtained. This results from the observation that the component images of a stereo image are, almost without exception, very similar. This could be exploited by some sophisticated compression techniques to increase the amount of compression of the images as a whole.

Unfortunately, most of the images transported across the network would tend to be individual images (see Figure 5) and not stereo images.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... exists)¹¹

Or the left component counterclockwise and the right component clockwise, as required and specified at the class's creation time.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... image¹²

The term NOP is derived from NO OPERATION, commonly used in assembly language and other areas of computing.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... these¹³

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.