For the best flexibility in ground and on-orbit calibration modeling of the AXAF telescope over its entire field of view, including off-axis calibration evaluations, AXAF synchrotron reflectance calibrations require that the measured reflectance data be reduced to optical parameters analogous to n and k. We have developed a method for AXAF witness mirror analysis which is a modification of the NKFIT optical constants algorithm published by D.L. Windt. The algorithm assumes uniform layer thicknesses using a recursive, exact formation of Fresnel's equations, with a modified Debye-Waller roughness correction factor. The recursion formula has been modified to include an explicit double-precision formulation. The results of most of the fits of AXAF calibration measurements yield residuals less than 1 percent of the reflectance value levels down to R approximately .03. The precision of the measurements is smaller still, which compromises the (chi) (superscript 2) fitting algorithm; however, the results will most likely prove adequate for AXAF witness mirrors calibrated in the 5-12 keV range. Coating density determined from the refractive index n is approximately 98.5 percent of the bulk for iridium. Derived coating thicknesses are extremely consistent with the photon energy, giving still more significant calibration information to the program.