Cosine-to-the-fourth effect

Discussion The projected solid angle for pixels in an array will depend on their position in the array, and, as a result, the output and responsivity of pixels in the array will be nonuniform. This nonuniformity is not a flaw in the array, and we may be allowed to correct the outputs to show what the uniformity would be in the absence of this geometrical effect. We refer to this as the cosine-to-the-fourth effect - but the cos" ( ) relationship is an approximation, and more rigorous formulas are available - see Sections 2.3.2. Because of reflections within the cold shield, the actual effect of the cold shield will probably be stronger than predicted by the geometry. In at least one instance, a correction of cos - ( ) worked better than any theoretically predicted function. 

The NUC process not only corrects much of any intrinsic array nonuniformity but also automatically corrects for the cold-shield ( cosine-to-the-fourth ) effect if we do a NUC, we do not need to do any other correction for cold-shield effects. (However, the NUC process should be redone if we change cold shields.)... 

---