Flat Field vs F-Number?

[dmcclain]

Let me start by stating that I know enough about optics to be dangerous, but nonlinear intensity imaging is a new area for me.

It appears to me that as the F-number of the optical train increases, the intensity dome of non-flattened fields becomes more gentle and predictable. Sky flats produced at F:2.8 appear to have more extreme curvature than sky flats produced at F:8, for example.

Is this in fact the case? Should I expect to produce separate sky flats for each F-stop that I intend to use for image composition?

[NGC7789]

It is my understanding that the entire optical chain must be identical for a valid flat including f-stop and focus. The only thing that should vary from your light is the exposure and the light source although some may allow ISO to vary too.

That being said, why are you varying the f-stop? Shouldn't you be using the fastest f-stop that produces a sharp result?

[dmcclain]

Hi, Thanks for your reply...

I found that, indeed, one wants to use the fastest possible F-stop for extended background region pickup. But that induces significant coma and chromatic aberration (actually a failure of simultaneous focus in R,G,B) on a Canon EOS 6D with the Canon 200 mm F:2.8 lens. But when I stop it back to the point of pristine images all the way across the plate, I reach F:8. That's great for stars, but picks up almost zero background for any reasonable integration periods.

So I found that I could grab the extended background at F:2.8, remove the stars from that image, and then add back the stars from an F:8 exposure. The results look terrific!

However, it is a bit difficult to create a repeatable workflow. Every time I pair up a F:2.8 stack with an F:8 stack, I end up with slightly different results in background coloration, contrast, and depth. So I'm trying to find a formulaic method of making valid combinations for mosaic creation. (Repeatability is probably an issue for everyone, not just for this kind of combination work)

[dmcclain]

Here's an example composite of the region near AE Aur. A stack of 5x300 s at F:2.8 + 5x300 s at F:8. Stars in the image are clear all the way down to 15th mag. Canon EOS 6D (unmodified) + Canon F:2.8 200 mm FL. All the magic credit goes to PI!

[NGC7789]

Maybe its the jpeg but I'm not seeing the magic. Can you link to a full resolution somewhere? I'd also be curious to compare to  5x300 at F:4. And if the F8 shot is really for stars only are 300 second subs really necessary? Interesting that you've gone to great lengths to include all those stars while others try to minimize stars to emphasize nebulousity.