I've been using the SubframeSelector and FWHMEccentricity scripts to sort my images and I'm wondering what other users are seeing when they run these scripts. It's all very well to select frames based upon the capabilities of your own equipment and technique, but I'm curious about what values other users are getting. To make any comparison at all we'll need to know what your capture scale is (arc seconds per pixel), whether you are using a monochrome or full-color sensor (assumed Bayer pattern), and the aperture, field coverage (focal length and sensor size), and type of your telescope. Reports should also include the stage in image development where the images were measured. Typically, that might be immediately after calibration and debayering (if required), but before stretching or any other digital development.
My greatest concern (or interest) is in eccentricity values as I very seldom see results that are better than the 0.42 limit that is supposed to indicate round, undistorted stars. My FWHM values seem reasonable, but my typical median eccentricity values are always above 0.45 (values in the 0.5 to 0.6 range being quite common). When viewed visually at 1:1 scale, the center-field star shapes run from slightly elliptical to what I'd call very close to perfectly round. The odd thing is that even on very short exposures (say, one second) my center-field eccentricity values still don't reach that 0.42 limit. I recognize that some of the poorer median values may be caused by field curvature and edge aberrations, but I struggle to get values below 0.5 even in the center of my field (on multiple telescopes, not just from one sample).
As an example, here is the summary from my Astro-Tech AT72ED (a non-flat-field, 72mm aperture, f/6 doublet ED telescope mated to an APS-C Sony NEX-5N camera). The capture scale is 2.3 arc seconds per pixel with a 4.8 micron pixel size and the diagonal field of coverage was 3.7 degrees. This data shows a lot of field curvature (as would be expected), but even in the center of the field the eccentricity is between 0.6 and 0.55. Over this same center the FWHM ranges from 4 to 6 arc seconds. Here is the data from the full frame (includes the "bad" edges):
FWHM 4.76 arc seconds
Eccentricity 0.725
StarSupport 4168
FWHMMeanDev 1.54
EccentricityMeanDev 0.121
The sample was taken from a stack of ten images that were exposed for 10 seconds each, there was no image calibration, the ten images were debayered (VNG), registered, and integrated with no normalization or pixel rejection. The measurements were then taken on the linear image using the SubframeSelector script (with only one image, the stacked result as input).
Is anyone willing to put forth there own measurements or add some comments on their own use of these scripts.
