Flats leaving bright circle in image, Nikon D810 + 50mm 1.8g + PixInsight

[pipnina]

I have been trying to get something to enter for the 50mm lens challenge, and figured my new-old DSLR kit would be a pretty good setup to enter with... However I seem to be encountering issues with calibration. All I've taken is flats and lights, as I thought bias/darks often were not useful for DSLRs due to a lack of temperature control and ADU offset? It seemingly worked a bit better than this in my D3200/Sigma 105 lens but I am not sure what's causing this issue.

As you can see in the stacked and autostretched image above the corners seem quite bright with maybe skyglow, the lower quadrent is comparatively dark and the middle has this odd glowing bow, and between the green and grey it's actually purple. I have to put this down to flat calibration issues as the un-calibrated lights alone produce a very normal (but highly vignetted) image. Sadly the stack below I forgot to turn on debayering but it does show the image coming out as I would expect, no odd shapes in the vignette that might cause the issue above, meaning I must be doing something wrong in capture or setting something incorrectly in pix? My flats should have been at the same focus, aperture, iso etc as my lights but at a very short exposure time due to the fast lens and high ISO (800) of my lights, I think 1/4000 putting the center of the image in the middle of the histogram.

Any tips from more experienced pix + DSLR shooters greatly appreciated! 3 lights and 3 flats in the files below if anyone wants a closer gander!

Many thanks

 

_DSC1087.NEF _DSC1088.NEF _DSC1089.NEF _DSC1060.NEF _DSC1061.NEF _DSC1062.NEF

[happy-kat]

There is lots of helpful info on this by member sharkmelley I think covering same thing

[tooth_dr]

I’ve used my Nikon D800E with my epsilon and the flats are similar.  It’s because it’s full frame and the issue is light reflecting off the edge of the mirror (when it’s flipped up taking an exposure) creating flats that don’t match the lights. Careful processing might help mitigate it a little but only by physically removing the mirror box can you resolve it completely.

Edited January 8 by tooth_dr

[pipnina]

12 minutes ago, tooth_dr said:

I’ve used my Nikon D800E with my epsilon and the flats are similar.  It’s because it’s full frame and the issue is light reflecting off the edge of the mirror (when it’s flipped up taking an exposure) creating flats that don’t match the lights. Careful processing might help mitigate it a little but only by physically removing the mirror box can you resolve it completely.

Now this is interesting... After you brought this up I looked at my flat RAWs in both pix and rawtherapee...

The same file, stretched in a different program seemingly causes a massively different result?

The bottom image seems like it would cause the issue in my final stack, whereas the top one seems totally innocuous!

Could the camera be embedding something in the raw file that programs can tap into to perform "correction"? Why would PIX apply it in one context (calibration) but not another (viewing)?

Most puzzling, I am going to have to delve further into this...

[pipnina]

Yeah this seems to be exactly what's happening...
I got PixInsight to convert the raw files for one light frame and one flat frame to TIFF in order to strip it of most metadata. When I calibrated and processed with the tiff files everything behaved exactly as I expected it should, the flat calibration worked perfectly and background extraction flattened the image completely.
So I guess I need to convert files from raw to tiff if I want to process D810 images for astro... interesting...