The source of the input signal for a raster display can be either a television camera or a computer image generator. The input to a raster display device can be modeled as a sampled version of a continuous function and the output as the result of a reconstruction process. Regardless of the source of the input signal for a raster display, the postsampling filter is usually, although not necessarily, limited to the spatiotemporal characteristics of the display device itself. A nondistorted low-pass version of the original image may provide a perceptually acceptable level of resolution. Information concerning such a retinal image is presumably important if the perceptual system incorporates precise information about the movement of the eyes. The diagonal sampling grid of an interlaced system is more efficient than a rectangular sampling grid if the pre- and postsampling filters pass comparable, diamond-shaped spectra.