Re: ATM P-V vs. RMS

[Anthony Stillman <atmer@flash.net>]

A few remarks.

The Rayleigh criteria was originally proposed as a means for comparing the
resolution of spectroscopes.  Subsequently it was adopted to relate
aperture to the resolution of two adjacent point sources. In the presence
of aberration the Rayleigh criteria is invalid.

The Rayleigh quarter wave limit, as Jim Burrows noted, originally
associated one quarter wave of primary spherical aberration in the
wavefront with a 20 percent reduction in the intensity of the center of its
circular diffraction pattern at the Gaussian focus.  This amount of
redistribution of energy was considered tolerable.  Later, it was found
that this limit when applied to other low order aberrations produced
similarly tolerable image degradation.

Systems which meet the Rayleigh quarter wave limit are often described as
diffraction-limited.  A more appropriate description is tolerably-degraded.

The quarter wave wavefront aberration limit is at best a crude requirement
for tolerable optical performance.  And this, only with the caveat that the
wavefront is smooth.  P-V values can be meaningful, when combined with
additional information.  At the very least a qualitative description is
needed.  In contrast a reasonably good Strehl ratio or RMS wavefront error
value can alone aptly predict system behavior.

Of course, all of this is moot if the values for these metrics are derived
from bad and/or inadequate data.

Anthony