Existing models for the scintillation camera are limited to designs based on pulse-mode operation, although some cameras are now able to process two or more gamma-ray interactions at a time. Techniques for improving count-rate capability basically do so by reducing camera deadtime, and include division of the detector into multiple geographical zones, use of special algorithms to identify temporally-overlapping events that are spatially independent in the crystal, and use of pileup prevention circuitry (PPC) for recovering one or more events from multiple pileup interactions. Minimization of deadtime effects in other ways has been confined to limiting the camera's operating range to the 20% count-loss point, despite mathematical models showing that this can be extended to the peak response point by correction of count losses, with residual errors of 10% or less in the corrected rate. The present paper describes a comprehensive camera model for single-zone and multi-zone operation without and with PPC, including correction of deadtime losses. While multi-zone architecture and PPC both increase camera sensitivity, and hence improve contrast resolution in the image. Multi-zone operation, however, also reduces the generation of misplaced pileup events, which is one major cause of image blurring. Correction of count losses restores the detector's sensitivity and contrast-to-noise ratio to the values they would have if deadtime effects were not present. The operating range of a five-zone detector, with PPC of order 1 and with correction of count losses, for instance, enables the input-rate range to be extended to as much as 100 times that of the basic camera, with less than 1% residual error in the corrected rate. Because of greatly increased camera sensitivity obtainable in this way, it becomes possible to trade some of the detector's sensitivity for increased spatial resolution in the collimator and the detector crystal. Since nuclear cameras and PET scanners currently have substantially lower spatial resolutions than CT and MRI, and nuclear images are now increasingly being integrated with images from these imaging modalities in hybrid, or fusion, imaging, improvement of the spatial resolutions of nuclear cameras becomes even more desirable. A realistic model not only is a useful aid to a good understanding of camera operation and the interplay of its parameters in dictating performance, but also a logical starting point for the researcher seeking to develop improved new designs, or to optimize image-processing software for existing ones.