Flat frames over correcting.

[Crushellon (Tim)]

Hi guys.

I'm having a problem with my flat frames over correcting my light frames, ie lighter corners and dust motes being brighter instead of darker. I'm using bias and dark frames all done with the batch preprocessing tool in pixinsight. The flats are around the mid range of adu count with 6 seconds exposure at twilight. I've done a bit of reading of posts with this same issue and have seen in some cases adding to the adu count with pixel math can fix the issue but I don't know exactly how to do this. Any ideas?

[RickS (Rick)]

Originally Posted by Crushellon

Hi guys.

I'm having a problem with my flat frames over correcting my light frames, ie lighter corners and dust motes being brighter instead of darker. I'm using bias and dark frames all done with the batch preprocessing tool in pixinsight. The flats are around the mid range of adu count with 6 seconds exposure at twilight. I've done a bit of reading of posts with this same issue and have seen in some cases adding to the adu count with pixel math can fix the issue but I don't know exactly how to do this. Any ideas?

Hi Tim,

I've never had a problem with flats over or undercorrecting but it seems like a common thing. If you want to add a pedestal to a flat master you can use a simple PixelMath expression like "$T+0.05" and apply PixelMath to the flat master. PI normalises data values to a range 0.0 to 1.0 so this is equivalent to a 16-bit integer pedestal of 65535*0.05 or approx 3277. Adjust the value until you're happy with the result...

Cheers,
Rick.

[Crushellon (Tim)]

Originally Posted by RickS

Hi Tim,

I've never had a problem with flats over or undercorrecting but it seems like a common thing. If you want to add a pedestal to a flat master you can use a simple PixelMath expression like "$T+0.05" and apply PixelMath to the flat master. PI normalises data values to a range 0.0 to 1.0 so this is equivalent to a 16-bit integer pedestal of 65535*0.05 or approx 3277. Adjust the value until you're happy with the result...

Cheers,
Rick.

I have no idea what any of that means but I'll give it a go, thanks.

[RickS (Rick)]

Originally Posted by Crushellon

I have no idea what any of that means but I'll give it a go, thanks.

Use the PixelMath process. Set the RGB/K expression to $T+0.05. Click the "Create new image" button in Destination. Apply process to your master flat to generate a new version. If it still corrects too much try making the 0.05 bigger. If it doesn't correct enough make it smaller.

Cheers,
Rick.