I have a fast newton (270@ f/3.6) with a coma corrector, I found that the focus of the stars at the edge of the field is very critical. Extremely small defocus can affect drastically the star roundness.
In SGP I can crop the edge to perform auto focus, but having the possibility to crop the centre may be sometime useful. If you perform a good focusing on the edge of the field, it will be a good chance that the focus will be good also in the centre, of course the correction is not too bad on the entire field.
May be I am totally wrong.
Assuming an even star field (bad assumption, I know!) I would think the HFR of stars further from the center would dominate the average because there is more surface area. But, I was thinking a crop center would be useful, too … I had a situation a couple of months ago where I couldn’t get the nebulosity rejection to work effectively, and was getting a lot of false stars in the object I was imaging – they were very, very low HFR contained to the object in the center 20% of the image. This was at a long focal length so there were relatively few stars to begin with. I don’t even remember what object it was, but I do remember thinking, “dang, if I could just crop the center!”
2.4 is supposed to have an improved focus routine, but maybe the guys could implement a plate solve focus routine to guarantee nonstellar rejection.
A center crop already exists:
It sounds like Pascal is looking for an inverse center crop? IE only focusing on the edge stars and not the central stars?
The changes in 2.4 only effect how the data is gathered and allow it to be more deterministic about what happens when you are a ways out of focus. It does not change the underlying focus mechanism…more how data is gathered based on trends.
Probably a more generic solution is to add the ability to define a ROI for focusing, you could reuse the ROI selected in the frame and focus panel.
may be the tittle of this topic is not clear, yes I would like to have the ability to do the autofocus only on the edge stars.
Having the ability to define an ROI or ‘just’ an excluded central area could be a good thing.