Colour Halo Reduction

[avandonk]

Some of you may have noticed that with modded DSLR's that there can be a red or orange ring around even perfectly focused stars when you use even a very high quality lens. This is because lenses were not designed to accurately focus far red all the way to HA. Lenses are designed only for visible light.

Fortunately ImagesPlus has a star reduction function. I used IP to first star reduce only the red layer until there was no sign of red colour bleed. I then used star reduction slightly on all colour channels equally and then a bit of RL enhancement.

Result here 11MB

http://d1355990.i49.quadrahosting.com.au/2009_09/SMCmosenhc2.jpg

Here is an animated gif showing the enhancement on a 100% crop 1MB

http://d1355990.i49.quadrahosting.com.au/2009_09/smcenh.gif

Below are the same two images


This is also very handy for narrow band combination to get rid of any colour haloes.

The purple ring you get with achro scopes is a combination of red and blue haloes.


Bert

[[1ponders]]

Great!!! Thanks for that Bert. I've been collecting comparison images for a new thread for exactly this problem. Thanks for the solution, I'll try it out first chance I get.

Time to do some reprocessing. ;)

[[1ponders]]

So how do you apply the star reduction to only one colour? Do you split the image into separate r, g, b images reduce the stars in R and then recombine or is there a way to do it with just the star reduction function? I'm using IP3.8

[[1ponders]]

found it under "Star Size and Halo Reduction" of course :P

Confusing when they also have "Star Size Reduction"

[jjjnettie]

I've used the Chromatic Abberation, Star Repair and Star Rounder plugin's in CS3 to combat this problem. Both worked quite well on the brighter stars but produced a streaky effect (for want of a better word) between the fainter back ground stars.
I can see that the Star Reducer function in IP does the same thing to a certain extent.

[avandonk]

Jeanette this is another classic result of the First Rule of The Universe.

There is no free lunch or you do not gain without an equal loss somewhere else!

It is all about compromises. It is most probably something to do with Fourier wavelets (coefficients).

bert

[[1ponders]]

I've had a quick play with one of my 135mm shots. I'll definitely need to spend some times tweaking the settings of the Halo reduction and especially the LR iterations and point spread function size.

Do you have a standard set that you start with Bert and then adjust as needed?

[jjjnettie]

TANSTAAFL yep, I hear you.

[avandonk]

Paul I have found upsizing the image first with IP (x1.5 or even x2.0) and using bicubic interpolation does wonders as the corrections are then statistically more meaningful. Especially RL enhancement. There is no recipe as it depends on too many factors. I am afraid it is trial and error.

Bert

[[1ponders]]

Thanks for the tip Bert. I'll give it a go.

[avandonk]

please explain!

Bert

[jjjnettie]

It's an acronym coined by Robert A Heinlein...
There Ain't No Such Thing As A Free Lunch :)

[[1ponders]]

Isn't it TINSTAAFL?

"there is no such thing as a free lunch"

[avandonk]

The A threw me, thanks.

Bert

[avandonk]

As you all know the Canon 300mm F2.8L is one of the finest lenses ever built. The 135mm F2L is almost as good. There is a cigarette paper between them.

It took me a while to figure out what was happening. Usually unless I am sure of my facts I keep quiet.

Now I am almost sure I thought I should explain what took me far too long to figure out.

I look forward to better images from all that this factoid can help.

Bert

[[1ponders]]

I noticed this effect out at Duckadang. I had imaged M31 with my WO72 FD and with the 135mm. The 72 image was fine while the 135 showed the halos. At first I thought it might have been smoke haze causing the issue, but I shot a few images on the 135 night and no matter their altitude they all showed a similar halo. I was starting to think there was something wrong with the lens. This issues is particularly problematic in dense starfields, eg Nth American Neb/Deneb area and inner tail of Scorpius. I was planning on shooting the same area (M31) with my 85 f/1.8 (admittedly not an L lens) to see if it produced the same effect as the 135. I'll still do it, but at least now I won't be so concerned with the lens. :thumbsup:

[jase]

Blinking the two images Bert, it certainly is a marked improvement. Stars take on a tighter appearance, but not over cooked so aesthetics are maintained. Reviewed the star profile of a same star on both images in MaximDL. Perhaps a slight increase in saturation post reducing the halos is all that would be needed. Thanks for sharing this info.