Are my flats over-correcting? Other reason for bright glare?

[Moonstruck]

Hello Robin,

I am having an issue with bright glare in a stacked file (it looks like reverse vignetting around the edges). I know this is outside the topic of image capture, but I thought I'd ask anyway in the event you knew of a possible solution. I also posted this to Cloudy Nights, with the hope someone will respond.

I shot Thor's Helmet (NGC2359) the other night under very dark skies with no moon (est skies were bortle 4-5). I used a C11 Edge telescope with 0.7 reducer, ZWO asi 2600 camera, and Celestron OAG with ZWO asi 174mm guide camera. My mount was an EQ8-R-Pro. My stacked image included fourteen 10-min lights, 20 flats, 20 darks and 20 dark flats. I did the stacking in DSS. When I stretched the stacked image in PI, I was surprised to see a bright glare around the edges of the frame. It looked like vignetting, but bright instead of dark.

Am I over-correcting with my flats? If so, how do I correct that? Or do you see another reason for the bright glare?

I have attached a link to a single light frame, and a master dark, flat and dark flat frame below. I have also included the stacked file with all lights and calibration frames. If you can assist with a possible explanation (and hopefully a fix) I'd really appreciate it! Thanks.


Single light frame:

https://drive.google.com/file/d/1EpxJGN ... sp=sharing


Master dark:

https://drive.google.com/file/d/14XO8lP ... sp=sharing


Master flat:

https://drive.google.com/file/d/1dQvVyM ... sp=sharing


Master dark flat:

https://drive.google.com/file/d/16gpMOa ... sp=sharing


Stacked file with calibration frames:

https://drive.google.com/file/d/1__L8Tj ... sp=sharing

[Moonstruck]

Never mind - I received a response from someone on CN who suggested I re-do my darks. I did, and it worked just fine. Thanks anyway.

[admin]

Hi,

just because I downloaded and looked before reading your second post, I'll reply...

The bright areas in the processed image match the darker areas in the flat so yes, they are overcorrecting. This is usually due to a mismatch between either

* lights and darks

or

* flats and dark flats

If the bias level in either pair is different between the two in the pair then you will get problems - things that can cause this are different exposure, gain, offset (brightness/black level) or colour balance settings.

Anyway, good to hear that the re-done darks fixed the issue. It's probably worth noting that if you stretch the image far enough, you will pretty much always see some hints of over/under correction, but when it is done properly you can stretch much more before it becomes noticeable.

cheers,

Robin

[Moonstruck]

Thanks. The person from CN that gave me the advice that my darks were bad also pointed out that the dark frame "offset" was an issue, as it was much lower than the other offset value of the other calibration frames. What is an offset, and how can I see what the offset is for light, dark, flat and dark flat frames? I don't know how to locate or analyze offset information. I didn't see it in the capture files.

[admin]

Hi,

the 'Offset' is a control that you have in the camera controls on the right. Actually the name depends on the brand of camera - for ZWO it is 'Brightness', for other brands it may show up as 'Black Level', but they are all the same thing (I chose a name to match the way that the manufacturers themselves describe it for each brand).

Anyway, whichever of those three names show up in the controls, they all do the same thing - they control what the minimum pixel value for a perfectly dark pixel is. If you turn down to minimum, dark pixels with no light will be close to pixel value zero. As you turn them up, the pixel value for an unilluminated pixel increases. Basically, this adds a constant offset to each pixel value in the image as the image is captured on the camera.

Why not just set this to zero? Because there is also noise in the image, even for unilluminated pixels, which spreads the actual value for each pixel out in a distribution around the average. You need to turn this control up enough that there is *no* chance of any pixels being able to try to generate negative values due to the noise, since negative values get set to zero which then biases the image and makes processing properly close to impossible.

So, the thing to do with this control is set up the gain and other settings you use, with no light on the camera, and adjust the offset to ensure that the peak in the camera histogram isn't touching the left hand side of the graph. Once you have done that, you've worked out your offset value and use it for everything.

If you have a difference in offset between lights and darks (or between flats and dark flats) then when you subtract the darks away from the lights, part of the result will be due to the offset difference rather than it all being due to light falling on the sensor, and the flat correction will be applied to that constant offset part too, making the sort of effect you are seeing.

The value used when capturing an image is recorded in the 'OFFSET' header for FITS files. You will also find it in the CaptureSettings.txt file that is created next to other capture file types.

cheers,

Robin

[Moonstruck]

Thanks for the great explanation! I have a ZWO camera so in my case it's the brightness control. Right now the brightness is set to 113 and the histogram peak is close to (but not touching) the left side of the graph, so it looks ok. I have already made a dark library with brightness set to 113 (the brightness slider is about half-way across), so I won't change it to avoid a mis-match w/lights and my other calibration frames, which I take at the time I shoot. This is very helpful - thanks again.