Monochrome flats being created in color

[Starslicer]

It seems anyway. Lately my flats are being made in color even if the monochrome setting is chosen, see attached file.

Also see files below. Camera settings and the master flat created.

https://drive.google.com/file/d/1s7m2FA ... sp=sharing
https://drive.google.com/file/d/1E7xz9M ... sp=sharing

[admin]

Hi,

thanks for reporting - I can see the line of code that *might* be the cause of this happening, but oddly that has been there since 2018, so there must have been some other cause for this changing recently - I will continue to investigate.

cheers,

Robin

[admin]

Hi,

Ah, I see now - this is actually how it is supposed to be. For various reasons, flat correction works better without debayering the colour images (debayering introduces a blur to the image, so sharp edged dark dust spots would no longer correct as well). So, to make a 'monochrome' flat, SharpCap doesn't debayer the image to a true monochrome, instead it balances the brightness of the four colour channels in the raw image (R, G1, G2, B). This should leave a flat frame that doesn't affect the colour of the image (as long as the flat frames themselves are a consistent colour across the image). This also explains why the code in question has been like it currently is since 2018

Looking at your flat frame in detail, there are some areas where there is no bayer pattern visible, indicating that the balancing of the colour channels did a good job, and others where the pattern is present (at about a 2-3% variation in brightness level). How are you illuminating for the flat frames? The only thing that seems to make sense as the cause for this is a colour variation across the flat frame due to something to do with the illumination or flat frame generation technique.

cheers,

Robin

[Starslicer]

For the one I put on here, it was a white LCD screen placed on the end of the scope. I did another yesterday morning to make sure it wasn't something else and I shot it at the morning sky with a T Shirt over the front of the telescope. Results were the same so I can rule out the LCD being an issue.

I even tried unchecking the monochrome flat option just to see what it would look like. It looks the same. I can show you all of that when I get home tonight.

The stacked image ends up looking ok, but I do worry that the colors are off because the flat is in color. A year or so ago, all the flats I took in sharpcap were black and white in my ASI fits viewer.

https://drive.google.com/file/d/18KkPzw ... sp=sharing

[admin]

Hmm, OK, let me go through that again (and test with an older version of SharpCap to see if I see the change in behaviour too).

cheers,

Robin

[Starslicer]

I finished up my current project so I decided to test a few things out now that I can mess with the optical train.

I took flats with the NBZ that I had before, an Antlia Quadband and also some without filters. They come out the same no matter if monochrome is selected or not, maybe that's just how it works. I just figured they would look different if monochrome was selected or not.

IDAS NBZ
https://drive.google.com/file/d/1wH-MnV ... drive_link
https://drive.google.com/file/d/18H96l6 ... drive_link

Antlia Quadband
https://drive.google.com/file/d/1qRfkSS ... drive_link
https://drive.google.com/file/d/1uwI9lb ... drive_link

No Filter
https://drive.google.com/file/d/1lSqLJ3 ... drive_link
https://drive.google.com/file/d/1uusvfP ... drive_link

[admin]

Hi,

I'm confused, since to me, those files look very different - taking the 'no filter' pair as an example and opening each in FITSLiberator, I get

Screenshot 2025-07-04 215558.jpg (448.22 KiB) Viewed 2211 times

The 'Off' version on the left has three distinct histogram peaks coming from the R/G/B pixels in the bayer matrix grid (the green is probably the right hand most one as it is fatter than the others, and there are 2x as many green as R or B pixels). The 'On' version merges those into a single peak by the re-scaling of the different colour pixel groups to give matching average brightnesses. You still see a hint of the bayer matrix effect, but the smaller range means you can actually see the dust and vignetting in the view of the flat.

I checked the quadband ones with similar results. The NBZ only has 2 peaks in the 'Off' version - maybe the brightness of two of the channels (R/B probably) are close enough for their peaks to overlap in that situation.

cheers,

Robin