Something wrong with flats

[RedLeader1]

I do flats each time I do imaging but each time I see what almost looks like flats over compensate. I make sure the white channel gets to around 50% brightness and typically go for 15 frames. I also make sure that none of the channels go above 90% brightness.

One thing to point out which could be the issue is that I use my floodlight to generate bright light with telescope pointed to it about 6ft away and then use a white sheet to diffuse. I then adjust the gain and exposure to get the white channel to the 50% mark. The red channel is most prominent which I get to 85% at the top end of the brightness.

I have actually purchased an A3 LED light panel which seems to be a popular way of doing flats as perhaps the floodlight isn't producing the right kind of bright light?

See attached image for an example and below is what is output to the text file when I do a save during staking:
[ZWO ASI071MC Pro]
CameraSerialNumber=201F480D17010900
FrameType=Light
Debayer Preview=Force RGGB
Pan=0
Tilt=0
Output Format=FITS files (*.fits)(Auto)
Binning=1
Capture Area=4944x3284
Colour Space=RAW16
Exposure/Gain Shift=0
Hardware Binning=Off
Turbo USB=100
Flip=None
Frame Rate Limit=Maximum
Gain=90
Exposure=15.000s
Timestamp Frames=Off
White Bal (B)=50
White Bal (R)=50
Brightness=8
Window Heater Power=On
Cooler Power=17
Temperature=-4.9
Target Temperature=-5
Cooler=On
Auto Exp Max Gain=300
Auto Exp Max Exp MS=30000
Auto Exp Target Brightness=100
Mono Bin=Off
CA Reduction/Synthetic Blue=Off
Mask over exposed pixels=Off
Flip (after dark/flat)=None
Trail Width=3
Minimum Trail Length=100
Trail Detection Sensitivity=9
Remove Satellite Trails=Off
Background Subtraction=Off
Planet/Disk Stabilization=Off
Banding Threshold=8
Banding Suppression=60
Apply Flat=E:\SharpCap Captures\2024-11-11\Capture\flats\MasterFlat_22_18_39.fits
Hot Pixel Sensitivity=5
Subtract Dark=E:\SharpCap Captures\darks\ZWO ASI071MC Pro\RAW16@4944x3284\15.0s\gain_90\MasterDark_30_frames_-4.9C_2024-11-02T16_13_33.fits

I also attach the master flat I used.

Any ideas of where I am going wrong?

[admin]

Hi,

just from looking at your images, the master flat looks very dark - only about 1/4 of full image brightness in the middle, and getting down towards zero near the edges/corners - that's not looking good (especially the very low values near the edges/corners - they imply very large corrections being applied). It's odd that the visible brightness of the flat doesn't tie in at all with your description of how you were aiming to expose the flat frame.

If the flat frame is roughly right in terms of the relative brightness of different areas, then the amount of light reaching the corners and left/right edges of the image is 50 to 100 times lower than the amount reaching the middle - that means that the flat correction is trying to boost the image by 50 to 100 times in those regions, and it's probably unreasonable to expect flat correction to work flawlessly in that sort of situation. You would certainly need to have everything spot on to even get close - you can see various hints on getting good flat frame correction here : https://docs.sharpcap.co.uk/4.1/#Gettin ... Correction

However, if the level of vignetting in your set-up is getting towards a factor of 100 or more in the corners, you may be best off cropping the image down to maybe 75% of full width and height - that would exclude the bulk of the problem areas where the flat is having to do too much work.

cheers,

Robin

[RedLeader1]

Thanks Robin.
I am going to wait (not much choice with all the cloud going over the UK atm) until my LED light panel arrives as I don't think using the floodlight method is helping. However I do think I need to try things out a bit more with flats and get different results and see which one works better.

On the link you sent it mentions the following: "When capturing flat frames either use the option to subtract bias frames or the option to subtract dark flat frames." If selecting either the of the options you mention does it do bias frames or dark flats automatically as part of the flat frame capture if I select either of those options or do I have to do them separately in Sharpcap?

Thanks.

[admin]

Hi,

yes, if you select those options then SharpCap will automatically capture bias frames or ask you to set up dark conditions to capture dark frames - it will all happen automatically.

cheers,

Robin

[RedLeader1]

That's fantastic, thank you.

[RedLeader1]

I have been capturing flats with bias frames as is suggested but as you will see with the stacked image of the bubble nebula there is still vignetting with colour. I capture 30 frames for flats and separately I do 60 frames for darks. I now use an LED light pad and use a couple of sheets of A3 to diffuse the light a bit.

I do 60 second and 30 second exposures with gain set to unity gain of 90 for my ZWO ASI071MC pro.

Is there something I am not doing right?

[admin]

Hi,

I'm not sure if it is the same issue, but discussions in another thread (viewtopic.php?p=46086#p46086) have just reached the conclusion that the simpler ways of using flat frames (just flats + bias, no darks) do not cope with the way that ZWO cameras tend to apply colour balance to the image when the colour balance controls are moved away from their neutral values of 50/50. Perhaps this could be causing the over-correction in the corners of the image?

cheers,

Robin

[RedLeader1]

Thanks again Robin. I will give that a try.

[RedLeader1]

I gave it a try without the darks and interestingly it made things slightly worse but I think I found out how to manage it.
Dialling down the brightness/offset from the default of 8 down to 5 seemed to help an managed to get a couple of good shots - see attached with no post-processing and came out well. Horsehead/Flame nebula shot is 15 x 60 second frames and Orion is 32 x 60 second frames.

I had tried previously with the with the brightness setting but I think it was likely I changed other settings as well but I think it will probably be a setting I will need to tweak ever so slightly depending on the target but for now happy.

[admin]

Hi,

well those images have come out nicely in the end, so well done. The fact that setting the brightness/offset to a lower level helped is a sign that the problem is related to over/undercorrection by the flat due to either

* dark/light frame mismatch (gain, exposure, offset/brightness and colour balance shoult all be the same)
or
* flat/darkflat or flat/bias mismatch (less likely as the SharpCap process captures the flat and the bias together).

Skipping the dark entirely will cause SharpCap to fall back on the estimate of the bias level from the flat frame, which is as mentioned above suspect if the camera colour balance is in use.

cheers,

Robin