The basic geometry of a parallel camera setup is shown above. The two cameras are arranged with parallel axes and the zero disparity plane (ZDP) is set to coincide with the region of the scene (the yellow square in this case) that we want to appear in the plane of the screen when the image is shown on a 3D display.
Notice that the camera view angles are chosen so that the perspective frustum have the same width in the zero disparity plane. In computer graphics systems this can often be arranged directly by defining asymmetric perspective frustum exactly as above. For other computer graphics systems and almost all physical cameras this is not possible and a symmetric frustum is used. In this case the left and right cameras take images wider than required and the images must be cropped using a software tool such as StereoPhoto Maker before presentation on a 3D display.
The zero disparity plane (ZDP) is also confusingly known as the zero parallax depth (ZPD) and the HIT point, it is also equivalent to the vergence point used in toed-in camera setups. The importance of this distance in all cases is that it identifies where in the scene will appear in the display plane in the final stereoscopic image and divides the scene into objects that will appear in audience space (in-front of the display surface) and in screen space (behind the display surface).
The anaglyph stereo image below illustrates the results of using parallel cameras, it is worth comparing this with the similar image on the toed-in cameras page. You should see noticeable differences, in particular there is no vertical disparity in the image and no curvature of perceived depth across the image.