Sorry, just saw this post...been busy with work.
Before I digress, we should all acknowledge that there is no magical smoke and mirrors algorithm to make shotty stars look perfect. We are talking about pin point sources of light which are best controlled through good optics and guiding amongst other things. Manipulating stars in post processing should ideally be an act of enhancement not recovery. Try not to get into the situation of the latter.
Quote:
Originally Posted by Alchemy
not sure , but jase says .... or my interpretation of what he said was..... that digital development gives that effect
might have to do a search to confirm that though.
|
Close... MaximDL has a Kernel Filter which can be invoked by deconvolution or Digital Development function. You can configure the "User Defined" Kernel filter to sharpen stars in a non uniform manner. This can be used to alter edge to centroid mapping. A word of warning, from experience this method can be a little hit and miss. There is an assumption that all stars across the field have the same problem - common with guiding errors - difficult for optical errors such as off axis coma. There is no "selectiveness" in the routine - all stars are given the same treatment. However, you can work around this by creating different Kernel filters and integrating them into a single image using masks. You'll pull your hair out in the process - not for the faint hearted. More information is available here;
http://www.princeton.edu/~rvdb/images/deconv/deconv_MaximDL.html