One of the questions surrounding the use of videokeratoscopic height data is the ability to extract useful parameters in the presence of noise. Parameters such as spherical equivalent power, radius of curvature and conic constant need to be accurately calculated from noisy height data. There have been suggestions that submicron accuracy of the corneal height is required to faithfully determine quantities such as axial and instantaneous power.1 This accuracy requirement, however, appears stringent and further examination of height precision is necessary.