A challenging problem associated with performing hardware- in-the-loop tests of imaging infrared seekers is projecting images that are spatially realistic. The problem is complicated by the fact that the targets may be small and unresolved at acquisition and grow to fill the field of view during the final guidance updates. Although characteristics of the projection system are usually thought of as determining the spatial realism, the imagery used to drive the projector is also important. For a pixelized projector, the driving imagery must be sampled at a rate determined by the sample spacing of the pixels in the projector. If the scenes contain important information that is small compared to the projector pixel spacing (that is, if they have important information at high spatial frequencies), then information may be lost in the sampling process if the images are not adequately bandlimited. This bandlimiting can be accomplished by prefiltering the scenes. At acquisition, targets are usually small; thus, prefiltering is necessary to preserve information about the target. Without such prefiltering, for example, infinitesimally small targets would never be seen unless they just happened to be at the exact location where the scene is sampled for a projector pixel. This paper reports the results of a study of various filters that might be used for prefiltering synthetic imagery generated to drive projectors in the KHILS facility. Projector and seeker characteristics typical of the KHILS facility were adopted for the study. Since the radiance produced by projectors is always positive, filters that can produce negative values were not considered. Figures of merit were defined based on the sensor-measured quantities such as radiant intensity, centroid, and spot size. The performance of prefilters of various shapes and sizes and for typical projector and seeker characteristics will be reported.