Phantom "Star pairs" ... bizzaire !

andrea tasselli:
They are hot pixels. Try a more selective rejection algorithm and run CosmeticCorrection, if using PI.

Hi Andrea - thanks for the info; I figured that it was that but stuck me odd as to the placement and predictability but now, upon closer examination of the RGB masters I do see it in the R and B masters and the subject framing was different by exactly that much explaining the spacing issue but what I don't get is why I haven't seen this up until now? especially in so many places - the pixel correction was working and suddenly now isn't with no changes made. 

Maybe  just a fluke and it goes away on my next run? 

Pete
PS - by the way, locking the mirror (manually for now until I install the secondary focuser)  has definitely greatly reduced those earlier aberrations that I was getting - your analysis of that was much appreciated!

How are you dithering?  If not with a random dither, you can start to see patterns like this. Or if you are not dithering enough.

Pete,

Dithering is essential for rejection algorithms to perform properly.  It's best to dither every frame if you can, although sparse dithering every 2 or 3 frames is ok as long as you collect a lot of subs. 

It sounds like your starting to capture a frame after dithering before your scope/mount has had enough time to settle.  This can be addressed by increasing the settle time, or addressing something physically with your setup that might be unstable or prone to shaking after moving.  The easiest is to just increase the settle time.  Depending on your mount and other equipment it could take a little bit to settle.   Maybe 15 or 20 seconds?  I dont know, I have not imaged at this focal length before. 

You could also reduce your dither scale a little bit.  You can experiment here, but my instinct would be that dithering at a smaller scale would be better than not dithering at all.  I understand that there is a balance between collecting data and spending time dithering and settling, but it is worth the time investment for better data.  (IMO and YMMV).

Pete,
Those spots are small enough to be hot pixels.  The problem is that they clearly repeat the same pattern in each of the groups so that's not the only problem.  If they are hot pixels, something has gone wrong in the stacking process to produce this sort of pattern.   I do agree with Chris that dithering is extremely important.  I also agree with Chris that being careful about setting up a proper settling time for both the guider and for dithering is important to avoid issues like this.  Dithering is trivial so my advice is to figure it out before working on another image.

So, what is it?   A number of things can cause this sort of result (e.g. wind, settling time too short, registration problems) but few that I can immediately think of would produce this pattern randomly distributed over the entire field.  In order for us--or better, for you, to diagnose something like this you have to be methodical about working your way backwards to see where it came from.  Help us out with a little more information:

1) Does this pattern exist exist in the individual subs?  Carefully blink through each sub and examine them very carefully to see if any of these points show up in the registered subs.

2) Are all of the subs all perfectly registered?  Again Blinking through the stack will quickly answer that question.  Integrating a stack with just two frames mis-registered could produce this result from hot pixels alone; although if that's the case, it would also exhibit other obvious symptoms in the main object as well.

John

Update - 

Lesson learned about dithering; after getting it to function properly last night I see a *significant* improvement in the data. Not just hot pixels but almost everything looks better, smoother yet sharper.

It looks as if I'll be axing my current non dithered projects! I hope it clears up! I need new data now!