Introduction

For a particular wavelength, given two FOC scenes of the same extended astronomical object taken within a few hours of one another and an observed FOC point spread function taken at nearly the same epoch, deconvolution of the extended images using the observed PSF should lead to similar small-scale structure. We have found that this is not the case. The probable reason for such differences is that the FOC system is time variant, coupled with the complication of noise on both the images and the observed PSF. It is well-known that observed PSFs taken in a sequence seem to ``breathe'' as if the focus of the camera system is changing with time. Parametric Zernike polynomial modeling of the point spread function core and wings, coupled with phase retrieval techniques constrained by the observed PSF, can produce a calculated noise-free (analytic) PSF. It may be that this analytic PSF can be used in the deconvolution of both observed extended scenes to obtain similar small scale structure by empirically changing only a few Zernike coefficients (e.g., the focus term).