Quote:
Originally Posted by Placidus
.........
A philosophical thought: if there was a very bright star, burning out to a radius large compared with the seeing, then detail out to that radius would be forever lost, because the information was not in the original photo. To be able to reconstruct a starless image even in principle, we require a very sharp original image, such as you have taken, and stars that are not burned out to any appreciable distance relative to the seeing.
|
I think there is some logic to that. I tried various colour images and was confronted with artifact central. Tight h-alpha data worked best for me. That said in the original 4k file, there was clear artifacting from the star removal routine. With a little Gaussian blur and just under 50% web resolution it cleaned up quite nicely.
Quote:
Originally Posted by gregbradley
Remarkable detail. I like it.
Greg.
|
Thanks Greg. The process has been around for a time
I recall reading about a photoshop method over a decade ago, which was picked up by Fred Vanderhaven who did well at the Malin awards with a "starless" image.
Starnet++ automates the process and certainly works well...but I think you need to pick the right data set for it to enhance the image rather than become a fashion statement.