Flats are overcompensating for vignetting

[CraigG]

Hello All,

I'm just starting to wade into adding flats to my DSS stacks of light and dark frames and I'm noticing that the resulting image seems to overcompensate for the vignetting in the original light frames. The center of the image is now appropriately dark but unfortunately it
gets gradually lighter moving out to the edges. Its almost the complete opposite of the original vignetting!

I'm using a Samsung tab with a generic lightbox app for the light source and assembling master flat files ranging from 10-30 frames. There's little difference in which master flat that I use. When I capture the flats, the light curves seem to be in the right place in SCs histogram from what I've seen online and I get a 'ok' message from SCs flat capture dialog box.

I guess I need to land somewhere in the middle to have a nice flat frame to work with but I'm at a loss to know what to try at this point. Any suggestions? Thanks!
-Craig

[turfpit]

Craig

Can you post a single flat somewhere on the internet, so that it can be downloaded and inspected. Also the camera capture settings file would be useful.


Dave

[CraigG]

Hi Dave,

I'll work on it this weekend. Thanks!

Craig

[CraigG]

So, I wondered if the attached info would be as helpful. I don't have an internet source established for uploading photos with FITs data so thought I'd try uploading reduced screenshots of my 10 frame master flat with the FITs header info from ASTAP. There are also reduced screenshots of my camera settings in SC that I used to capture the flat master files. Please note that the output file format showing in the camera settings was not SER but FITs files.

Helpful at all?

Thanks Dave..
Craig

[turfpit]

Craig

Looking at the histogram in the ASTAP image - the histogram is over to the right making the flats too bright, see the Minimum and Maximum figures of 45,000 - 56,000.

ASTAP-histogram.JPG (23.6 KiB) Viewed 829 times

Try adjusting the exposure down in SharpCap until the histogram is central (or slightly to the left of central). The (unstretched) small Display Histogram on the right will help with this. Using the histogram will take the guesswork out of the exposure required.

flat-in-display-histogram.JPG (13.3 KiB) Viewed 829 times

I use FITS Liberator from https://esahubble.org/projects/fits_lib ... nload_v30/ for inspecting FITS files. My typical flat will be like this with a Mean of 25,000 - 30,000 ADU. When I am imaging, I always take a single test flat and inspect it with FITS Liberator before committing to capturing my usual run of 50 flats.

mono-flat.JPG (141.14 KiB) Viewed 829 times

It will be interesting (and useful for future reference) to see what min & max figures ASTAP returns with the lower exposure.

Try to create and test the new flats indoors rather than waste any clear skies.

Note that including the camera settings file in a post is faster than posting the settings via screen shots. The text file will contain all the necessary settings to help with problem solving.

Dave

[admin]

Hi,

I don't think the histogram will be a big problem, as you look to still be well short of 90% (and camera response should be linear up to well past 90%).

One thing to be aware of is that flat frames typically also require dark frames to work correctly.

If you use dark flats or bias frames when creating the flat, but do not use dark frames when using then, they will tend to over-correct. SharpCap tries to adjust for this if you *have* used dark flats or bias by adjusting for the bias level automatically, but that relies on the brightness (offset/black level) being unchanged between the flat and the light frames.

Conversely, if you use darks when capturing lights, but no dark flat/bias, then the flat will likely undercorrect.

If you don't use darks/bias for either then you tend to get an even worse mess - the flat will overcorrect in the darker areas of the frame and overcorrect in the brighter.

cheers,

Robin

[CraigG]

Wonderful info gentlemen! That helps a lot..

Dave, I'll follow your suggestions including checking out Liberator. I really thought I had my exposure correct after viewing Cesar Alverez's excellent YouTube bit on capturing flats in SC but next time I'll try to favor more to the left of the histogram and see what Liberator tells me. Wonder why ASTAP showed a higher value than SC's histogram?

Do you think it would help to lower my camera gain from 100 to say 0 before adjusting the exposure?

Robin, I only capture a limited number of darks..2-3 each session since the ASI2600 doesn't need much help. I then add them to the mix in DSS only. I'll have to revisit the flats capture module to see but are you saying that there's a way to capture darks along with flats? Still trying to wrap my head around the relationship between flats/dark flats and lights since I just started playing around with flats earlier this week.

I appreciate the good counsel as always!
Craig

[turfpit]

Craig

My imaging has evolved using FITS Liberator therefore I have an understanding of the effects of the mean ADU (25,000 - 30,000), the position of the histogram and the effect on the final image. I suppose it is whatever tool you get used to that is the most useful. I have never used ASTAP.

As far as capture goes, I do this (for the sake of my sanity and consistency):
- lights at whatever exposure, gain and offset (brightness) have been decided
- darks same settings as lights
- flats same gain and offset (brightness) as lights with exposure reduced to obtain desired histogram
- dark flats same settings as flats but with the cap on the scope
- bias (if needed) same gain and offset (brightness) as lights but with smallest possible exposure

Over complicating exposures in the cold and dark is something I avoid.

As cameras (electronics) have evolved the requirements for bias and dark frames have also evolved. If you read around the internet you will turn up what appears to be conflicting information about calibration frames. CCD cameras need bias frames. Older CMOS cameras need bias frames - with a school of thought that the bias exposures are maybe up to a couple of seconds. The latest low noise CMOS cameras (no amp glow) no longer require bias frames and the dark frames have lower noise. For your camera the read noise drops dramatically at gains above 100 - see the graphs at
https://astronomy-imaging-camera.com/pr ... -pro-color.

The idea with calibration frames is to take many as the creation of the masters takes advantage of the averaging out over many frames. I take 50 each of bias/flats/darks/dark flats (as required). It looks like your camera is cooled so you could take a set of darks at say -10C twice a year for each exposure you intend to use. That way you save valuable time to use on capturing data. I would certainly advise capturing more than 2/3 darks.

I got caught out with flats twice in the last couple of years. Once with LRGB filter. The RGB frames were being captured at bin2x2 which made the flats bright and the histogram shot off at the right hand side. I couldn't turn down the brightness of my EL panel. An emergency fix was a white tee shirt folded double. The other was using a Canon 7D with a f3.5 lens. Again too bright for the flats which I fixed with an ND8 filter to reduce brightness.

What image processing software are you using?

Astro imaging is more of an evolutionary journey rather than a big bang. You will learn more when things don't work. Have a look at my M42 Chronicles viewtopic.php?t=596

Dave

[CraigG]

Morning Dave,

Well I've done some more experimentation since our exchange yesterday. I captured some new master flats indoors last evening with a lightbox app on my tab using your suggested ADU ranges and analyzed them with Liberator (love that program BTW!). I also captured some flats early in this morning's clear blue sky and was able to grab single and masters in and out of the suggested ranges. Unfortunately, all of my sample flats from both of the sessions still are not evenly illuminated. All samples have a brighter core to varying degrees.

Now I noticed that the white light curve in my SC histograms appear to be narrower and sharper than Robin's flat aquisition tutorial. There also not as flat as the bell curve in your Liberator example so I'm guessing that I have something else going on in my setup that's not allowing even illumination no matter what I try. Have to noodle a bit on that issue.

My processing software is primarily Nebulosity but I also use StarTools. The scope is a Stellarvue SVX102T-R with the dedicated field flattener, not the FF/FR. Using a Baader UV/IR cut filter at present.

Yes, this really is a journey and we move forward in ...mostly...small steps. Wouldn't miss it for the world though!

Have a good weekend Dave,
Craig

[oopfan]

Hi Craig,

Dave and I follow the same regimen. We've never had problems with vignetting since we religiously capture all calibration frames: Bias, Darks, Flats, and Dark Flats (only if Flats exposure >2 sec.) In Robin's reply he said as much.

I read your reply to Dave (above). Your Flats will always have a bright center. (How large it is depends on the size of your sensor relative to the Imaging Circle of your optics.) You appear to have a large sensor. So, your Flats are good. What you need to check are the calibration frames you're using. Do you have Darks?

Brian