Abstract

The influence of noise in the measured intensities on the calculated phase from a phase-shifting interferometer is analyzed. A dependence on the particular phase-shifting algorithm used is investigated for the cases of independent intensity noise and intensity noise due to quantization. The method used for the analysis is a straightforward Taylor series expansion of the phase in terms of the multiple intensities necessary for phase-shifting algorithms. When statistically independent intensity noise is present, the variance of the calculated phase improves for algorithms using a larger number of intensity frames: the variance is shown to be independent of the algorithm for the case of error in measured intensities due to quantization. Good agreement is obtained between this method of analysis and Monte Carlo simulations.

© 1988 Optical Society of America