Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Why my flats have reverse vignetting?

  • Please log in to reply
10 replies to this topic

#1 axlns

axlns

    Vostok 1

  • *****
  • topic starter
  • Posts: 150
  • Joined: 17 Sep 2019
  • Loc: Novosibirsk, Russia

Posted 08 August 2020 - 04:07 AM

I dont know if it is normal or if something is wrong with my setup. Scratching my head as I did not have this with my previous setup (on previous setup I got normally looking flats, with brigher area in the center and darker area on the edges). So I'd appreciate any help...

Here are my 3 master flats for Ha, SII and OIII filters.

 

Ha

flat-HA.jpg

 

SII

flat-OIII.jpg

 

OIII

 

flat-SII.jpg

 

As you can see all 3 shows similar thing, bright edges and dark center.

 

This is my setup

 

Telescope: TS Optics Imaging Start 102mm APO with builtin flattener

Camera: QHY695A camera with OAG; Set of LRGB and NB Optolong filters 36mm unmounted

 

I create flats using USB controlled 520mm Artesky Flat Box 

 

On my previous setup I did not had such problems, that setup had almost all the same stuff, but I changed telescope from Skywatcher 8'' newtonian to TeleskopService APO refractor, and also I changed from noname simple flatbox to that expensive one from Artesky.

And now I'm getting these results. 

 

Do they look normal to you? Maybe its the way that APO with builtin flattener works? So maybe its expected?

 

Also, at this moment I cant really say if my flats correcting my lights fine, maybe you can see it from the stacked Ha image? The problem for me is that there is a lot of signal in this frame and I dont really know where it just background and where it is actual signal.

 

 

light-HA.jpg

 

So what might be wrong here? 

Thanks!



#2 james7ca

james7ca

    Cosmos

  • *****
  • Posts: 8,832
  • Joined: 21 May 2011
  • Loc: San Diego, CA

Posted 08 August 2020 - 05:33 AM

Flats correct more than just vignetting and they will also be affected by reflections and light scatter. Plus, it's obvious that you've stretched these flats and that can make things look worse (or better) depending upon the level of stretch and the variation that exists within the particular frame (thus making it hard to do a direct comparison between two different systems).

 

One thing you can do is to apply the same level of stretch to each system or series of flats (don't rely exclusively on an "auto" stretch or something like PixInsight's Screen Transfer Function).

 

In any case, if the flats seem to be working then why do you care what they look like (particularly in comparison to a different scope)? I don't see anything obviously wrong with your sample image so the flats can't be that "bad." The only comment I'd make is that it looks like you have some large dust spots on either your camera (most likely on the sensor chamber window) or perhaps on the flattener itself since the dust seems to stay put when you change filters.


Edited by james7ca, 08 August 2020 - 05:34 AM.

  • axlns likes this

#3 jdupton

jdupton

    Mercury-Atlas

  • *****
  • Posts: 2,755
  • Joined: 21 Nov 2010
  • Loc: Central Texas, USA

Posted 08 August 2020 - 08:54 AM

axlns,

 

   I have seen the same with narrow-band filters in my system. Those appear rather normal.

 

   I am not positive but think the darker centers has to do with the nature of the narrow-band interference filter itself and how it works. The filter is made up multiple layers of dielectric materials laid down in precise thicknesses. That is how they can be made to pass only (for example) a 3 nM band of wavelengths. They are made to block everything outside of that desired wavelength.

 

   Now, when you use these, the light coming from on-axis will be filtered so that only the 3 nM band of wavelengths is passed. Nearly all other wavelengths are blocked. However, light coming in off axis will see a different relative thickness in the dielectric layers. This means that band-width and center frequency of what wavelengths are passed can change. More light of different wavelengths may now be able to reach the sensor. That means that this narrow-band filter is less effective at blocking off-axis light, making the off-axis areas brighter. Off-axis, the 3 nM filter may operate as a 5 nM filter with a shifted center wavelength.

 

   If this is what is happening, then the effect should be worse as the focal ratio is reduced. That is one reason why these filters are often much less effective when used with very fast optical systems compared to their performance with slower optics.

 

 

John


Edited by jdupton, 08 August 2020 - 09:01 AM.

  • axlns likes this

#4 axlns

axlns

    Vostok 1

  • *****
  • topic starter
  • Posts: 150
  • Joined: 17 Sep 2019
  • Loc: Novosibirsk, Russia

Posted 08 August 2020 - 11:17 AM

thanks for your comments,

just to give more info on my setup

 

focus length of my reflector is 1000mm F/5

focus length of my APO is 520mm F/5.1

 

the reason why I think my flats are looking strange is because with my other scope, SkyWatcher 8'' reflector (and also different flattener, baader mark iii) but with the same camera and same set of NB filters I'm getting rather "normal" flats, for example this is my HA flat done with that scope (yes, its stretched with STF in PixInsight to better show "pattern")

 

flat-HA-newt.jpg

 

as you can see with relfector my flat is brigher in center and darker on the edges, this is what I would expect from any flat..

but with APO I'm getting the opposite - dark center and bright edges

I thought this is strange, but maybe its normal as you saying...

 

John,  after looking at this, I think your theory that it might be caused by how well NB filter blocks off-axis light seems to be not working with newtonian telescope, as newtonian show rather "normal" pattern?


Edited by axlns, 08 August 2020 - 11:18 AM.


#5 RJF-Astro

RJF-Astro

    Viking 1

  • -----
  • Posts: 642
  • Joined: 13 Aug 2018
  • Loc: Zeist, Netherlands

Posted 08 August 2020 - 11:55 AM

Or it could be that your newtonian had more vignetting, and the effect from the narrowband filter is not that visible in a flat. If the apo has less vignetting, the effect of the filter becomes more visible. Just a theory.
  • axlns likes this

#6 jdupton

jdupton

    Mercury-Atlas

  • *****
  • Posts: 2,755
  • Joined: 21 Nov 2010
  • Loc: Central Texas, USA

Posted 08 August 2020 - 11:59 AM

axlns,

 

   Another contributing factor is that it is likely that the reflector has much more field vignetting than the APO. It may well be that the vignetting in the reflector is overwhelming the effects of the narrow-band filter. It may be possible to see if that is the case. You could try this: normalize the average ADU of the flats from both optical systems and then subtract off the APO flat from the reflector flat with a small pedestal. Then use ABE with a function degree of 2 to process the result. If the result then looks more or less flat, it was just an excess of vignetting that masked the differences. I may take your screenshots and see if I tell anything from this or a similar experiment.

 

 

John



#7 axlns

axlns

    Vostok 1

  • *****
  • topic starter
  • Posts: 150
  • Joined: 17 Sep 2019
  • Loc: Novosibirsk, Russia

Posted 08 August 2020 - 12:22 PM

RJF-Astro, John, thats good theory that newtonian had more vignetting overhelming effect of NB filter, yep that makes total sense.

I shall see how Lum flat would look like on APO.

and, if John you would want to amke that experiment, you could use that zip file instead of screenshots - https://www.sugarsyn...9685078_6646990

it contains 2 FIT files, one HA flat from newtonian and another from APO 



#8 jdupton

jdupton

    Mercury-Atlas

  • *****
  • Posts: 2,755
  • Joined: 21 Nov 2010
  • Loc: Central Texas, USA

Posted 08 August 2020 - 01:43 PM

axlns,

 

   I am now convinced that the reflector's greater vignetting is simply hiding the patterns in the HA filter.

 

   I ran the experiment I outlined above. I took the Reflector HA Flat and simply subtracted the APO HA Flat (while adding an offset of 25000 ADU to keep the same rough level in the result). I then ran the FlatContourProfile script against the original Reflector Flat and the result of the subtraction. 

 

   You can readily see that the original Reflector HA Flat contains a rather complex set of gradients. The result of the difference between the two flats shows reasonably clean pure vignetting. Once we remove the complexities of the relatively flat APO flat, the Reflector's flat show the vignetting over and above any present in the APO.

 

flat_Ha_Newt_contourPlot.png

Original Reflector HA Flat

 

HA_Ref_minus_HA_APO_contourPlot.png

Result of Original Reflector HA Flat minus the APO HA Flat

 

   In my interpretation of this experiment, the odd patterns you see are due to the narrow band nature of the HA filter and off-axis light effects. The pattern is much less obvious in the Flat Frames from the reflector because they are being masked by a greater degree of vignetting. They are visible as shown in the contour plot for the raw frame but you must look closely.

 

 

John


Edited by jdupton, 08 August 2020 - 01:46 PM.


#9 james7ca

james7ca

    Cosmos

  • *****
  • Posts: 8,832
  • Joined: 21 May 2011
  • Loc: San Diego, CA

Posted 08 August 2020 - 11:04 PM

While an interesting and potentially useful exercise, I'm not certain that the most recent conjectures have provided a definitive explanation on what IS or MAY be going on with the OP's flats. However, I think it does illustrate the problem that I mentioned at the very begin of this thread. That being that you have to be very careful about how you interpret two different sets of flats that have both been subject to PixInsight's ScreenTransferFunction (i.e. auto stretch). In fact, I would NOT use a ScreenTransferFunction (alone) in order to do a definitive comparison between two different systems.

 

If you want to do a direct comparison you really need to do a statistical analysis looking at both the average and the variation across the field (and how that actually affects your calibrated subs). A simple auto stretch MAY wildly distort any differences that exist between two different sets of subs and in the end that may be a difference that is no difference.

 

When looking at the OP's results (based upon an auto stretch) all that one can conclude is that the patterns are different, but the relative differences between areas in each individual flat (i.e. the "true" variability) is really hard to judge in an auto stretched sample. In addition to appearing different, I guess you could conclude that the refractor is showing less vignetting which might be expected. However, the brightening along the edges and corners of the refractor MAY or MAY NOT be that significant in comparison to the variability shown on any other system.

 

For example, look how "clean" the Ha flat from the reflector looks in comparison to the Ha flat on the refractor. This should tell you that the auto stretch process is greatly exaggerating small differences in the refractor which may or may not be that significant in terms of what is happening on the reflector (or on any other system).



#10 sharkmelley

sharkmelley

    Soyuz

  • *****
  • Posts: 3,602
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 09 August 2020 - 02:25 AM

Given this anomalous behaviour, is it possible that a sky flat will perform better i.e. that a sky flat will be a better match for the night-time lights?

 

Mark



#11 axlns

axlns

    Vostok 1

  • *****
  • topic starter
  • Posts: 150
  • Joined: 17 Sep 2019
  • Loc: Novosibirsk, Russia

Posted 12 August 2020 - 12:35 AM

John, thanks for your analysis, I tend to agree with your conclusion. Also it seems that these "strange" flats seems to be working fine correcting my lights.

 

James, yeah, lesson learned. one should not directly compare flats from 2 different systems (even if camera is the same in both).

 

Mark, thanks for your suggestion. But I tried sky flats for Ha filter, and the pattern look about the same as with flat box.




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics