Flat colour balance

[Al McLean]

On the occasions that I use ASTAP to calibrate and stack my OSC images, it complains about the colour balance in the flats:

**" Warning flat pixel colour values differ too much. Use white light for OSC flats!!. Will compensate accordingly." **

I use an LED panel normally and, yes, the histogram shows a separation between the colour peaks.

So yesterday, with a fairly uniform gray, cloud-covered sky I took some sky flats, using a white acrylic diffuser and some folded white cotton sheet.

When I came to stack these ASTAP still complains about the colour balance.

My first question is why does the colour balance matter? The same histogram is visible in the light frames so this is fundamental to the camera (an ASI 2600 mc) not to the light source, and if the lights and flats both show the same response (and they do) why won't it calibrate without 'Will compensate accordingly'.

And my second question is what whiter source of light than a cloudy day time sky is there?

[han.k]

Hi Al McLean,

The warning was added because I had bad experience with a local club member using an electro-luminance panel where red was disappearing. This resulted in too much noise in the red channel and an unbalance.

The unbalance is now much better compensated. Now the program uses the red, blue and two green sensitive pixels separately. So as if four separate flats where made.

So in answer to your question it is pretty good compensated. Likely it will make no difference.

Normally daylight should result in no warning. So your camera is somehow an outlier. The warning comes if one of the colours is a factor two weaker then the strongest colour. So for example if blue sensitive pixel has value 50000 and red 24999.

What is the difference between the brightest and faintest response in the flat? Just measure 4 pixel in the center. Or share a flat/raw.

Han

[Al McLean]

Not sure how to measure several pixels at once nor to determine the colour, they only show up in shades of grey.

So I have put three files (a light, an ordinary flat and a skyflat) in a dropbox folder here:

https://www.dropbox.com/scl/fo/u7p6hxnes698d7wt4sx0m/h?rlkey=vhcaml1qgpt1y8qs64zibc1zs&dl=

Thanks for taking a look. Any comments will be welcome.

[Axel Thomas]

Han,
just stumbled over this thread as I have the same problem. I use an astromodified Nikon D750 and the flats (LED panel) definitely look reddish (see attachment). However as Al already pointed out if the camera has a certain spectral response this holds for both lights and flats so there should be no need for special calibration. I do not see any color shift in the stacked image. What exactly does the compensation?

cs Axel

[Axel Thomas]

sorry, here comes the attachment