Quote:
Originally Posted by Placidus
Very sharp image Colin.
Trish and I seem to be in the minority in much preferring the HSO palette, but this is a fine rendition.
I think you're saying you remove the stars, deconvolve the starless image, and then put the stars back. Two questions:
(1) Why not deconvolve with the stars in? Is it to avoid panda-eyes around the stars? We certainly do wavelet sharpening on the starless image, for that very reason, but with deconvolution we use an anti-panda algorithm.
(2) What do you use for a point spread function if you are deconvolving the starless image? Do you load the PSF from the with-stars version, or do you use a synthetic PSF?
Best,
Mike
|
I have in the past done a SHO of this region and I do love the contrast that narrowband provides but I do also like the star clusters that only show nicely in RGB.
1) It's definitely to avoid panda eyes
I've found that what goes into protecting panda eyes is what causes the deconvolution worms so if you don't do ANY protection you don't get worms. I routinely use 50 passes of deconvolution and don't ever have to worry about worms because I'm not protecting any part of the image.
2) I model the PSF before I do star removal. Modelling the PSF is usually one of the first processes I do anyway so it's a non issue