Back to forums

Why is vignetting so different between RGB and Narrowband filters?

Ritchey-Chrétien Reducer Processing Deep Sky Vignetting
11 replies•769 views
urban.astronomer avatar
urban.astronomer avatar
I recently got my hands on a 10" RC telescope and have completed a couple of astrophotography projects using both narrowband and RGB filters from my light polluted balcony (Bortle 8). Since the telescope is natively F/8, I also installed a reducer, (TS-Optics 2" CCD Reducer 0.67x), which seems to introduce quite some vignetting, actually much more vignetting than I have ever seen using other telescopes.

What I can't twist my head around to understand, are two things:
 
 1. Why is there such a huge vignette difference in my light frames captured using RGB filters and those captured using narrowband filters?
 2. Why doesn't the vignetting disappear after calibration of the RGB frames? 

Here are two streched, but otherwise unprocessed light frames captured using a 2" Blue filter from Baader, and a 2" Ha filter from Antlia:

I captured a series of flat frames using a light panel and made sure the ADU count was close to 20k.  The master flats vary in strength, but with individual streches they both capture roughly the same pattern:

The end result after using the Weighted Batch Preprocessing script in Pixinsight. As one can see, the Master Ha seems to be fine (apart from some field rotation which was my clumsy fault), while the Blue (and also Red and Green) suffer from extreme vignetting not removed in the processing 
Well Written Helpful Engaging
andrea tasselli avatar
andrea tasselli avatar
The answer to the first question is rather prosaic: you capture a LOT less background light with NB filter that you would otherwise do with BB filters. And that is why they are so effective in fighting LP.

The answer to the second question is not easy to give but in general it depends on mistakes you can make while you calibrate the light frames. It could be the worng offset, wrong bias, no bias, wrong master dark and wrong flat technique. I suspect it would essentially affect everything, NB images included but, given that the magnitude of the background with NB is trivial, it does not show in them as well as in the B frame (or G or to a lesser extent R).
Helpful
Arun H avatar
Arun H avatar
While it is true that the magnitude of light admitted by a NB filter is much less than an LRGB filter, the relative magnitudes (i.e., normalized spatial profile of light) between the two should be similar or even identical. You actually see this if you compare the H-alpha flat and Blue flat which show very similar vignetting profiles. 

I agree with what Andrea said in the answer to your second question. There is a recent thread around calibration of CMOS sensors which goes into detail:

https://www.astrobin.com/forum/c/astrophotography/deep-sky/lets-discuss-about-dark-bias-dark-flats/

Rather than do WBPP, I'd suggest manually calibrating for now. A bit tedious, but you have a better shot at identifying the issue.

Edit: I should add for completeness that while the spatial profile of vignetting should be identical between filters (assuming similar placement in the filter wheel, etc), the response of the sensor can be different. For example, the 294MM sensor will show a different spatial response to H-alpha and SII than to OIII or LRGB. Properly taken flats will still correct it just fine - the purpose of flats is not just to correct for vignetting and dust, but also Pixel Response Nonuniformity.
Well Written Helpful Insightful
Andy Wray avatar
Andy Wray avatar
One quick check you can do is to make sure that the camera offsets you used for the RGB dark frames and light frames were the same (I did say darks rather than flats intentionally).  It's caught me out a couple of times recently and gave me similar results.  You can find the offset in the FITS header of the individual frames.

If they are different, then you can do a quick and dirty temporary Pixelmath fix on the master dark to see if that is the root cause.

Disclaimer:  this could be a red herring, but would only take a couple of mins to check.
Well Written Helpful Insightful
John Hayes avatar
John Hayes avatar
The bigger question that you should be asking is, "Why do my flats looks so poor and why isn't calibration working correctly?"

When you see a pattern like you see in your flat data–namely arched/circular light/dark with distinct edges, you've most likely got a problem with stray light.  It's also a dead give away when you see sharp edges in the calibrated results.  Stray light can come from light leaks or from stray internal reflections.  In your case, I suspect a stray internal reflection from the inside of a connector that is causing the outer regions of your flats to be too bright, which will produce a dark region in the calibrated result, as you see in your data.  Remember that flat calibration only corrects multiplicative signal modulation; for additive signals!  Stray light is additive and it will kill your results.  You also could have stray light leaking around an opening in the image train.  Are you taking flats in the dark or during the day in a lighted environment?  Are you 100% certain that you don't have light leaks?

Excessive vignetting can also cause calibration problems.  How much vignetting are you getting and why?  Vignetting more than 30%-40% is where you'll most likely experience problems.  If you are vignetting more than that, you need to fix it or it's going to be very difficult/impossible to work with.  Have you looked at the size of your components and insured that the spacings are correct?

Once you have the problem worked out and you can show a collection of properly taken flats, then we can try to understand why your NB flat data might look different than LRGB flat data.

John
Helpful Insightful Engaging
Andy Wray avatar
Andy Wray avatar
A couple of other thoughts:

* Are you using Local Normalisation within WBPP?  If so, turn that off and try again just to rule out the chance that it's chosen a bad LN Reference frame.
* What pixel rejection algorithm are you using in Image Integration within WBPP?  If it is set to Auto, then try setting it manually to something like Winsorized  Sigma Clipping

Another thing you could try is using PixelMath to manually calibrate a single (uncalibrated) light frame.  I used the following formula:

(LightFrame-MasterDark)/MasterFlat*mean(MasterFlat)

If you wanted to make available a few files for us to look at, we could probably see if we can see anything unusual.  A few uncalibrated lights, the master flat and the master dark would be a good starting point.
Helpful Engaging Supportive
Rob Johnson avatar
Rob Johnson avatar
I don’t have a solution but I’m faced with similar issues after upgrading to a larger sensor with a TSGPU CC which causes significant vignetting . I have found that luminance calibrates well but RGB/Ha bad to varying degrees. I tried calibrating red frames with a luminance flat and that gave a big improvement but I’m not sure what effect that has on the data if any? I also noticed calibration seems to work better at the higher dynamic range (gain 26 on my QHY268M) but can’t be sure, need a few more clear nights to experiment.
Arun H avatar
Arun H avatar
John's comment about light leaks is worth following up on, and I wish I'd remembered my own experience in this area and mentioned it.

In my case too, with my reflector, narrow band was being calibrated nicely, but LRGB was not. I ultimately diagnosed it as reflected light from concrete entering the scope through the gap between the mirror and tube through the back at the scope. It was fixed by using a black shower cap. NB filters will be less sensitive to light leaks because the filter will block most of the light, the LRGB filters will admit a lot more of it. I am also careful not to take flats in direct sunlight; usually I take them before dawn, after dusk, or indoors.
Helpful Insightful Respectful
andrea tasselli avatar
andrea tasselli avatar
My rule N.1 in taking flats is always take them when the environmental light (that might get into the sensor) is just a minute fraction of the light used for taking the flat, which rules out taking flats during daytime. I had feared that using a large open-ended newt would result in difficult flat calibration or worse since I do take sky (dusk) flats but this so far as proved unfounded. Which makes me wonder why people on here find it difficult to protect the imaging train form stray light. Mind you, this newt of mine is nothing special and there is plenty of space between mirror/mirror cell and tube and I do not cover (yet) the rear end with a cap (one is however nearly ready for action, just in case). I do appreciate that cassegrain/gregory configurations are more prone to stray light issues (and not just when taking flats) so appropriate leeway must made for this telescope group. Needless to say, when I used mak-cass I never had any issue either but then YMMV.
urban.astronomer avatar
urban.astronomer avatar
Thanks a lot for many helpful tips in this discussion. 

I went over my equipment again, and I belive that there might be stray light that caused the issue. The open truss telescope was probably not completely covered during when taking the flats, as my flat panel is just a tiny bit too small to cover the full 10" opening of the telesope.

I will have to buy a new panel that completely seals the opening with a uniform illumination.

Thanks again!
-Martin
Respectful
Steve Cooper avatar
Steve Cooper avatar
It looks to me from your flat frames that the vingetting is the same between your filters. It doesn't show up in you light frame image as much because of the screen stretch. 
John Hayes is correct. You are suffering from a light leak. This is a known issue with these scopes, especially when using a reducer.
I am not advertising, but here is some info and a possible solution to the leak:
https://astromart.com/classifieds/astromart-classifieds/telescope-ritchey-chretien/show/gso-ritcheychretien-rc-baffle-extensions
I am in no way affiliated with this product. I just happened to see it on AM recently. 
I had the same issue with my AT16RCT and had to make my own baffle before these were available.
Hope this helps and good luck getting to the bottom of the rest of this. You are getting good suggestions.
Helpful Supportive
Rob Johnson avatar
Rob Johnson avatar
Thanks a lot for many helpful tips in this discussion. 

I went over my equipment again, and I belive that there might be stray light that caused the issue. The open truss telescope was probably not completely covered during when taking the flats, as my flat panel is just a tiny bit too small to cover the full 10" opening of the telesope.

I will have to buy a new panel that completely seals the opening with a uniform illumination.

Thanks again!
-Martin

Let’s know how you get on Martin, I’d be interested to know especially re the vignetting.
thanks, Rob