| The precision analysis of optical surface, which is to determine whether it will satisfy user's demands, is the important aspect of optical surface testing. It needs to select appropriate specifications to assess different optical elements for their different application. It is generally to use PV, RMS and Strehl ratio to assess surface errors, the inherent relationships among these conventional optics specifications are discussed in this paper. The proportionality relating PV and RMS is studied theoretically and statistically, and the result is that PV will generally show a factor of 3 and 5 greater than RMS if they are from the wavefront error fitted by Zenike polynomials while will typically show a factor of 7 to 10 if they are directly from actual measurements. Ultimate curve of Strehl ratio and RMS is obtained by computer optimization, furthermore, features of surfaces between two extreme situations are gained by computer simulation, then a conclusion is drawn that Strehl ratio of surfaces with a certain RMS value is related to the distribution of spatial frequencies, it will be low if the medium-high-frequency component of surface error is high, which is the reason why surfaces with the same RMS value have different Strehl ratio. The use of power spectral density(PSD) to measure mid-spatial-frequency errors in the wavefront is discussed. The calculations of 1D and 2D PSD are given and a method is proposed to calculate 1D PSD from 2D PSD for a given direction across the circular aperture by rotating 2D PSD or wavefront data. |