Three-dimensional displays that do not require viewing aids such as spectacles, termed "autostereoscopic" displays, present two or more distinct perspective views to two or more distinct viewing locations such that the viewer's two eyes may receive a stereo pair of images and perhaps may also move from side to side to enjoy further depth cues such as motion parallax. The management of the image-bearing light, and in particular its focusing or convergence to distinct viewing zones, is typically effected with the help of optical components at or near the apparent location of the image in space. Those components may consist of an array of micro-optical components such as cylindrical lenslets (a lenticular sheet, for example), or a single macro-optical component acting as a field lens or mirror (perhaps with a directional diffuser). The differences between these two design approaches, micro- and macro-optical, present different choices in determining whether the resulting spatial image will be ortho- or pseudo-stereoscopic, and also present different limitations to image depth based on diffraction by the relevant apertures. Following Weiss, we refer to the second category of optical systems, the macro-optical types, as "specular" autostereoscopic displays, even if some of the components are replaced with their Fresnel-optical equivalents (1).