Abstract
In an earlier paper, an aberration-balancing merit function, ΦD, was described in which the modulation transfer function squared, with suitable weightings, served directly as the image criterion. Further details concerning the optimization procedure are now presented: the coefficients of the normal equations are computed from the intermediate data generated in the numerical evaluation of the optical transfer function of the system. These coefficients are effectively the weighted values of the intermediate data in question, the weightings being design-parameter dependent.