RGB star problem

[Carole]

I am not sure if any-one will be able to help with this as most of you use OSC, but I thought I'd ask anyway.

I don't know why but just recently I combine my RGB filters into one and all the stars are aligned over each other just fine with no artifects, but if I try stretching them any further, even when I have already stretched them before combining, I end up with weird artifacts in some of the channels like this:

In this case the Red and Blue channels seem to be the worst.  This is also preventing me from using Mark Arcsinh method as the same thing happens.  Up until now it didn't seem to happen if I processed the RGB channels separately, but now it is happening there too. 

All channel stars are perfectly fine on initial stretch it seems to happen once I combine them and stretch a bit more. 

[MarkS]

Hi Carole,

Your "final combining" image looks like straightforward chromatic aberration to me.  If you look at the corners of a typical image do you consistently find the blue is on the outside of the star (or that it's on the inside)?

The channels can be aligned slightly better by using a green sub as the reference frame fro aligning and stacking your red channel and blue channel.  Maybe you are already doing that?  In the end there is not a lot you can do if you are experiencing chromatic aberration.

By the way, when you do a typical stretch on an RGB image, it will mix the channels and this will lead to weird effects if you separate the channels again afterwards because now, for instance, the green channel will have bits of both the red channel and blue channel added in.  The same for the other two channels.  A colour preserving stretch is the only way to prevent this mixing.

Here's my article on colour preserving stretches:
http://www.markshelley.co.uk/Astronomy/Processing/Colour_Preserving_Stretch/colour_preserving_stretch.html

Mark

[Carole]

I'll have a look that that tutorial Mark, but my problem is I am getting "sticking out bits" where none previously existed making coloured sticking out bits looking like the channels are not aligned properly.

It happens where I use your Arcsinh colour preserving stretch as well. But not always - no idea why sometimes its OK and sometimes not. 

It's very annoying.

Maybe I need to do all the stretching before combining and do none after that. 

Carole

[Carole]

Chromatic aberration.  This is using a Newtonian, I thought you never got CA with them.

Quote from Wikipedia

Advantages of the Newtonian design

Newtonian optical assembly showing the tube (1), the primary mirror (2), and the secondary diagonal mirror support (also called a "spider support") (3).
They are free of chromatic aberration found in refracting telescopes.

Carole

[MarkS]

Chromatic aberration.  This is using a Newtonian, I thought you never got CA with them.

Carole

Mirrors on their own have no CA,  But does it have a coma correcting lens?

Mark

[Carole]

Oh yes it does.  But I really don't think it is chromatic aberration.  It is something that is happening to the data during stretching.

Carole

[MarkS]

Oh yes it does.  But I really don't think it is chromatic aberration.  It is something that is happening to the data during stretching.

The red, green and blue colour components of the stars are slightly displaced.  The stretching simply reveals what is already there by brightening up the coloured blue and red fringes of the star which would otherwise go unnoticed.

Image stretching is very effective at revealing all types of optical aberrations 

Mark

[Mac]

This is not a lesson in sucking eggs..
Mac.

It might be due to the fact that RGB filters focus at slightly different positions, i.e. Chromatic aborations
If you have focused using the R filter first  then the Green and Blue will be out of focus, if you dont choose to focus inbetween

now im not talking about mm's here its sub mm focus.
i'e maybe 1/5 of a turn on your micro focuser.

The green might look in focus
as does the blue when your looking at a small image on the back of the screen
even using the laptop screen,

here you have to use 1:1 imaging.

i'e pixel peeping.

and view each image at full resolution.
one pixel on the screen representing one pixel on the ccd.

then each image can be focused exactly.

Now is this achievable,
yes but you need to look at the added time constants.

Each filter will now need to have a separate image taken, at full resolution,
you cant cheat this and use 2*2, 3*3 or even 4*4 binning.

So if you taking an image which is 60 seconds,
you might be taking 10 separate images before its in focus, repeat this for each of the filters and you see the problem.

You could start with the green channel and focus that absolutely spot on and then do the red and blue.
these will be out of focus as well, but the difference in focus between the red and blue will be closer then
if you focused on the red and didnt change.

https://www.edmundoptics.co.uk/resources/application-notes/optics/chromatic-and-monochromatic-optical-aberrations/?site=UK&countryid=231



Mac.

[Carole]

Hi Mac.

Well the different filters do tend to produce different sized stars, but I check all the filters before I start and all are in focus, so I do a sequence.  If I had to stop and re-focus in between every filter, I would give up imaging.

Unfortunately its not the difference in star size that is causing the problem, but your post has reminded me that I didn't register the RGB in Registar before combining, and I will go back and re-do that. 

But its the artifacts in the RGB images after stretching that are causing the problem more than anything else and I don't understand why this is happening, and I never had this problem in the past. 

Carole

[Carole]

Well thanks to Mac's post, it jogged my memory that I should have registered the individual RGB filters, and I have tried that and it seems to be much improved, it was just a part of the routine I forgot.  Never happened with narrowband as I always register them to the Ha individually anyway.

What I was doing here was combining the RGB without registering to each other and then just registering the combined RGB to the luminance and Ha.  I guess the miniscule difference that Mac mentioned made the colours slightly misaligned.

Carole

[Mac]

I check all the filters before I start and all are in focus

You said that each filter produces slightly different sized stars, to me that screams of only one filter being in focus.
i.e. the one that produces the smallest star, the other two are probably slightly out of focus.

Also one other thing that needs to be included / though about but is very rarely is temperature change.
As the temperature of the scope changes over the period of the night, then its focal length will also change,
again were not talking huge amounts, but it might be enough to slightly alter the length of the light path.

It might be interesting to compare the first NB image against the last one (same filter) to see if there is a shift of focus.

http://www.astropix.com/wp/2011/06/20/focus-change-due-to-temperature-variation/
Just a quick search found quite a few results.

Mac.

[Carole]

Star size is not necessarily to do with focus Mac, its the different amounts of light that some of the filters let through.  For instance Ha stars are much smaller than any broadband stars.  Oiii seems to produce bigger stars, but they are all in focus.  There used to be a tutorial in Starizona which talked about this, but annoyingly it seems to have been removed. 

If one filter is different to the others I will re-focus that, but it rarely happens that they are different and I don't understand why sometimes they could be different but not always.

Carole

[Mac]

For instance Ha stars are much smaller than any broadband stars

Just a thought, are all your filters narrowband?
If not then there is your answer.

Personally I still think its down to either a focus problem or a temperature shift causing a focus problem.

Have you measured the FWHM of your focused stars to actually see if you are in perfect focus?

Most imaging software have this built in.
The sharper your image the smaller the FWHM number.

https://en.wikipedia.org/wiki/Full_width_at_half_maximum

Also the seeing will also cause the star image to spread as well.

So if your RGB are taken of different nights then the focus could also be affected by the quality of each nights image.
Also as the night progresses the quality may increase or decrease.

Mac.