By numerical analysis of the general formulas for standard ellipsometiric measurement, positions of the compensator, polarizer, and analyzer can be found such that the errors of Δ and ψ are minimized. Such optimal conditions can be found for any value of Δ and ψ from plots that are given. The improvement of precision over that of usual methods can be one order of magnitude or better in some Δ, ψ regions.
E. SCHMIDT, "Precision of Ellipsometric Measurement," J. Opt. Soc. Am. 60, 490-494 (1970)