Which FWHM is SubframeSelector Computing??

[Astro61]

The documentation says it's using the DynamicPSF process to fit a star model. When I confine the area to one star and then use the DynamicPSF on the same star, I get two different results:

SubframeSelector: FWHM = 1.725 a-s,  Eccentricity = 0.6775

DynamicPSF:         FWHMx = 2.86 a-s,  FWHMy = 2.05 a-s,  Eccentricity (computed from r: e = sqrt(1-r^2)) = 0.6981

I tried all of the star fit models in DynamicPSF and none of them produce 1.725 a-s (it's set to use the Moffat4 model).

Thanks

Jay

[mschuster]

I am guessing a bit here, here is a possibility and a suggestion:

If you set "Subframe region" too small, the script will ignore it and give a typical FWHM for all detected stars in the entire frame. Check the process log for the warning message.

To run a test on a single star, do this: Create a preview around the star (say 10x10 px or so), drag it to the workspace to make a new image. Run Crop to expand the image to 1000x1000 px or so, where the star is basically centered. Run PixelMath "iif($target == 0, median($target), $target)" to replace the zeros with the median. (The median($target) will give a clipped, nonzero median of the original 10x10). Save this image and run the script with the point spread function you want to use. Run DPSF with the same function. The results should basically match.

In DPSF set scale mode to pixels, and similarly in the script set scale unit to pixels.

Thanks,
Mike

[Astro61]

Thanks Mike,

I gave it a go. I created a file from a small (320 px by 320 px) preview of a luminescence image containing one star and verified that SubframeSelector was calculating the parameters for that star via the output map. The results match...sort of. SS seems to output a quasi-average of the FWHMx and FWHMy estimates that DPSF comes up with. Just trying to nail this down as I want an approval criterion based on the maximum FWHM value. Thanks

Jay

[mschuster]

OK,

FWHM = sqrt(FWHMx * FWHMy)

Eccentricity = sqrt(1 - (FWHMy / FWHMx)^2) assuming FWHMy <= FWHMx.

You can try this estimate of max(FWHMx, FWHMy):

FWHM / (1 - Eccentricity^2)^0.25

which should work as an approval term.

Thanks,
Mike

[Astro61]

Yep. Thanks for the help. I should have realized that FWHM reported by SubframeSelector is the diameter of a circle with the same area as the ellipse with major and minor axes of FWHMx and FWHMy from DynamicPSF.

Jay