The effect of noise, including system noise, background noise, and cell-to-cell nonuniformity (spatial noise), is mathematically treated and experimentally verified for staring-mode infrared cameras. Spatial noise is shown to be dominant in high background environments (3 to 5 Am or 8 to 12 Am imagery at background temperatures greater than 0°C) even after compensation. Camera sensitivity is quantified by a contrast signal-to-noise ratio that includes the effects of system, background, and spatial noise. Past analysis of camera performance has assumed that the cell-to-cell nonuniformity can be completely removed by using nonuniformity correction techniques. We present data showing that neither variations in detector spectral response nor excess low frequency noise can be fully corrected using existing nonuniformity correction techniques. Furthermore, we show that even the small amounts of nonuniformity that persist after the application of correction algorithms will significantly degrade camera performance.