First LRGB image. Channels shifted or what happened?

scadvice

Member
I seem to have some sort of misalignment once the channels are combinded. You can see the red and blue along the star edges in the close-up image and a lot of small red shifted stars in the main. I have some idea of what may be caused this but wanted your thoughts here on the PI site.

As I inferred this is my first effort going from a DSLR to the mono camera (ASI1600-Pro). I attempted to clean up the background noise with MultiscaleLinearTransform and later TVGDenoise. I even tried ChannelMatch with little improvement.

As this is more or less a test imaging session I’m more interested in what is wrong not so much improving the image.

Things I know:

I only focused the scope with the Lum filter ( The filters are all Chroma and using a EFW).

The scope does not have any aberration issues. DSLR images do not have this problem.

I used the AutoColor script.

I only focused twice. At the start and after the meridian flip. (I need to get auto focus setup)

The weather is not cooperating right now but it would be nice to have some idea what went wrong before going out again. Thanks for any help!

Steve
 

Attachments

  • M81_M82_temp.jpg
    M81_M82_temp.jpg
    692.4 KB · Views: 93
  • crop1.jpg
    crop1.jpg
    68.1 KB · Views: 93
Hi Steve

I'm not sure what scope you are using so can only provide a general comment that changing temperatures throughout an evening's imaging will affect the focus position. As an example, I tend to re-focus my scopes every 0.7 deg C change in temperature. Focus is critical to image quality. As well as temperature affecting focus, whenever you change filter you need to re-focus or have your imaging software move the focus position at filter change to a defined off-set that you have previously measured. By only focusing with one filter before switching to a different filter without applying an off-set the focus with the new filter will not be optimal. You need to focus more frequently than you have described.

Further to this, the focus position will most likely vary with altitude as well as temperature (atmospheric effects due to the path of the photon travelling through more atmosphere when altitude is lower). Even at the same altitude, atmospheric diffraction effects each colour channel differently, thus shifting star position minutely between the subs of different filters; it's always best to capture your blue sub-frames at the highest altitude for your given target for example. You need to set your capture programme for the lights and filters accordingly.

The net effect of these factors, plus other optical and capture/equipment related factors, eg your optics, focus precision and reproducibility, seeing/transparency among others, will effect the star profile of your combined RGB image and can result in the artefacts you are observing (among other artefacts). Post processing techniques can improve channel alignment however it is always better to attempt to maximise sub-frame quality by honing and improving your capture techniques and AP practise. It can help in Star Alignment if you ensure you have distortion correction enabled and I have found that varying the colour channel of the sub used as the master for alignment can sometimes result in improved alignment of star centroids, eg using a sub from a red filter rather than a sub from a sub-optimally focused, slightly star-bloated blue sub.

I am not clear how much saturation you have applied to your image; there appears to be a lot of chromatic noise in the background. Nor the details of your sub-exposure lengths or number of subs per channel. Calibration of subs is also very important to overall image quality and there are very experienced users of CMOS cameras on the forum who can help if you think this would be of benefit (I only use CCD cameras).

Hope that's a 'starter for ten' to help you along the AP journey . . .
 
Last edited:
I only focused the scope with the Lum filter ( The filters are all Chroma and using a EFW).

Unless you are using all reflective optics this will not give the optimum result. When filters are advertised as parfocal, it means exactly that - the FILTERS are parfocal with each other. They will not magically make the telescope colour corrected.

The scope does not have any aberration issues.

All scopes have aberration issues - the better ones just minimize them by their more complex optical design or the addition of correctors. Even high end refractors have some amount of colour dispersion - usually the higher the cost the smaller the dispersion :)

Blinking through the channels of your image shows some misalignment so the first thing to check is that StarAlignment worked properly.

Knowing what optics were used will help if it is distortion rather than mistranslation that is the issue. Access to the raw master version of all the channels, posted somewhere like Dropbox, may help someone to diagnose the problem if neither of these is the issue.
 
Last edited:
Thanks Cosmic and Berry, Sometimes when I struggle to improve or try to find a solution in image processing I think I tend to end up over correcting and even getting lost in where I'm at in the flow. I suspect that is what has happened here. This is the first mono LRGB image I've done. I've processed a fair amount lot of OSC images but they are a whole different ball game it seems.

Obviously as Berry suggested waaay oversaturated. What has me perplexed is the the blue and red channels being shifted or appearing to be. Further confusion added when ChannelMatch seemed to have little if any effect as I was thinking any improvement there would lean toward filters not adjusted to the each other in position. Maybe I just expected too much out of it.

Lack of focus correction certainly is one of the issues as I'm using a Stellarvue SVA130T which is a 910mm FL. Cosmic my wording was imprecise as to aberration issues but likely that is not my problem the last Zygo interferometric test report was .992 Strehl and I have no reason to doubt it. I also agree that I was under a misimpression that it was not necessary to add to offset corrections between filters for this test image. I like your comment that blink suggest that StarAlignment may not have worked properly.

I used BPP but had moved the Image Registration -> Noise Reduction to 3 because I was getting and error at 1. I'll look at doing that all separately.

Again thanks for your comments and suggestions.
 
Last edited:
I'm using a Stellarvue SVA130T which is a 910mm FL. Cosmic my wording was imprecise as to aberration issues but likely that is not my problem the last Zygo interferometric test report was .992 Strehl and I have no reason to doubt it.


Nice scope. I notice that there is a field flattener and reducer/flattener available for it. Are you using either of these?
If you are you may not have the spacing quite right as the field appears to have slight curvature. If not using a flattener the curvature seems about normal for a telescope of this type. This can be mitigated a little by focusing on a star about a third of the way from the center towards the edge.
I have a 120 Esprit with a Feathertouch focuser that comes with a dedicated flattener and it took 3 attempts before I got the spacing exactly right and I only have an 18mm diagonal chip - much less demanding than yours.



ercl9cfk_eccentricity.png



It also looks like the camera is slightly tilted in the focuser. This should probably be corrected first to get an accurate curvature reading.

ercl9cfk_FWHM.png



The above plots are the support images from the FWHMEccentricity script run on the large image you posted. To really get accurate results I think you need to run this on your greyscale linear data.
 
Last edited:
I like your comment that blink suggest that StarAlignment may not have worked properly.

I always run my registered masters through Blink, with the image zoomed to at least 100%, and check that all 4 corners and the centre are properly aligned.
 
Last edited:
Did you also measure the FWHM of the individual R,G,B photos compared to each other and to L? If R or B are consistently above G, search for chromatic aberration or focusing issues.
The strange thing is that the blue is ALWAYS on the left and red on the Right. If it was chromatic aberration it should have swapped sides on the left and right side of the frame. Also the "L" stars are considerably smaller than the R,B, especially below M81
First of all did you use "L" filter or just R,G,B and left the "L" slot empty?
Secondly, did you buy the filters as a "set" that is declared to be "parfocal"?
 
I have not measured the FWHM of the R,G,B images and compared to the Lum. (yet)

"The strange thing is that the blue is ALWAYS on the left and red on the Right. If it was chromatic aberration it should have swapped sides on the left and right side of the frame...." That is an interesting observation.

Yes, I used the "L" filter. Approximently 50% more images of L than RGB.

All the filters were bought as a parfocal set and from Chroma. I checked with their technical support about how to face the filters relative to the sensor. Mine are all facing the same direction. They said that it doesn't make any difference which side. However, if I felt there was an issue they would replace one or all of the filters anytime I wished.

Adam Block was nice enough to do a Zoom conference with me and the conclusion was that the star alignment in BPP may be the culprit. I'm going to redo pre-processing manually without BPP. Also, that I lost focus during the later imaging sets so I'll be pulling all the images made after a certain point. I can't thank him enough for doing that!
 
Yes, just to put a bow on this. After looking at the data- the registration routine did it best to align both in and out-of-focus stars. The optical distortions of the out-of-focus stars resulted in radially dependant aberrations that I think were the problem.
The homework was to take all of the in-focus stars and just align those. Very likely if only in-focus stars are used- the alignment will be great without the flaring distortions.
-adam
 
Back
Top