Rainbow in star "cross"

[bitli]

Hi,
When I combine the RGB or LRGB images containing very bright stars, the star "cross" produced by the Newton secondary mirror holding pieces (sorry, I do not know the english name of that artifact) has a strong gradient of colors, like a rainbox. I have attached a jpeg of a small section of the image (the only processing was automatic STF copied to histogram and 2x resampling to make it easy to see).

Is this an optical artifact or a digital processing artifact that I should get rid off?

 -- bitli

[Nocturnal]

Some people pay money to buy software to add spikes like that on purpose

It's an optical artifact. Easily verified by checking if it's present in an unprocessed image.

[Philip de Louraille]

This is a known diffraction effect due to the vane spider mount which holds the secondary mirror.
More knowledge here: http://www.astronomyhints.com/spider.html

[bitli]

Thanks,
The spikes were expected, it was the rainbow effect that I surprised me.

I was suspecting another possible cause for the coloration because in the LRGB image there is a yellow rectangle that appears on the right of the bright star (see image below). I did not see any obvious reason for the yellow rectangle while inspecting the raw image. Nothing like that is visible in the L, but it is present but very dim in the combiend RGB. It becomes very visible when L is added (or LRGB build directly). It is quite visible after other processing is applied.

I assume nobody is paying to get small rectangle along their stars  . Did not found the cause. An alien spaceship? 
How to get rid of that?  Photoshop? Phasers ?
-- bitli

[Philip de Louraille]

I just spent 2 minutes looking at my collection of APOD pictures and quickly found 3 showing (bright!) stars with the same diffraction grating as yours, including the rainbow effect. Now... I can't say for sure if this is due a subtle effect due to sub-pixel alignment of the RGB frames or whether you would get the same effect out of a one-shot color camera.

[Philip de Louraille]

I checked some of my own one-shot color camera (SBIG ST2000C) of the Pleiades, and the rainbow effect is seen. Of course, now, can I say for sure that is not due to the de-bayering of the data...? There is always another layer to the onion of complication when trying to get to the final answer...

[vicent_peris]

Some people pay money to buy software to add spikes like that on purpose

That's the downside of painting. Even the simple act of adding spikes can result in a bad knowledge transmission.

These spikes, being a diffraction effect, have color gradients because diffraction angle depends on wavelenght of the incoming light. Thus we see the typical rainbow on the spikes.

Read the article at Wikipedia (http://en.wikipedia.org/wiki/Diffraction). Then, think that you can teach diffraction to people through your images.


And last, think about the use of software generated spikes.


Regards,
Vicent.

[vicent_peris]

I assume nobody is paying to get small rectangle along their stars  . Did not found the cause. An alien spaceship? 
How to get rid of that?  Photoshop? Phasers ?
-- bitli

I don't see any problem with these artifacts, as they are not intrusive into the objects of the image. They don't distort the objects, like a dust sopr or vignetting pattern can do.

Leave them in the image.


V.

[Nocturnal]

Bitli, I've seen similar artifacts to the sides of my stars as well. I think they're caused by the brightness of the core and are a type of bleeding where the high ADU count of the previous pixels carries over into the next. Just a theory though.

[Carlos Milovic]

Yes, it seems like blooming.

[Simon Hicks]

Bitli,

Vincent and Philip are right. It is part of the complex diffraction pattern obtained as light diffracts past the vanes on your scope.

The main structure, and the easiest to understand, is the big cross. If you only had one vane running vertically, then you would get a horizontal diffraction line. Basically the light passing the vertical vane wants to spread out slightly in the horizontal direction as it passes the vane. Much like watching ripples on a lake changing direction as they pass an obstruction.

This is how those Bahtinov masks work....they put a bunch of parallel 'vanes' in the path to give a strong diffraction spike at 90 degress to that bunch. Then they put two more bunches of parallel vanes, but ay different angles to the first bunch, to create the other two spikes. Each bunch of vanes covers a different sector of the mirror (or lens) and therefore the spikes only cross perfectly when the three sectors of the mirror (or lens) are collimated. Very cool.

The colour effect within the big spikes are the fine detail of the diffraction pattern and, as Vincent points out, they are due to the fact that the differing wavelengths diffract by differing amounts, so you get constructive and destructive interference taking place. If you were imaging at one wavelength (say Halpha) then you will not see the various colours, but you might see light and dark bands radiating out along the spikes.

The fact that you can see all of this detail with all of the colour banding means that your scope is well collimated.  8)

Cheers
         Simon

[Philip de Louraille]

Simon says... "The fact that you can see all of this detail with all of the colour banding means that your scope is well collimated."

Excellent point!

[bitli]

Thanks to all,

Sanders, I am afraid you may be right. I have another picture which had the same 'bleeding' effect but stronger (and in blue). Carefully examining the singel channel images, I can see the effect on the RGB channel (taken with a binned 2x, left column), but not on L (top right).

I tried to use the Chrominance Noise Reduction on the image, but the result is less than convincing (bottom right).  Maybe I will resort to the stamp tool after all...

The spikes are cute, but they make using a star masks or erosion about impossible, or did I miss something?
 -- bitli

[Carlos Milovic]

Hi Bitli
This effect is called "blooming" and is caused by pixel saturation. Google it to find more info and sample images.