Author Topic: On adding "diffraction spikes" post-aquisition  (Read 9300 times)

0 Members and 1 Guest are viewing this topic.

Rick

  • Administrator
  • Galaxy Cluster
  • *
  • Posts: 6545
    • http://www.orpington-astronomy.org.uk
On adding "diffraction spikes" post-aquisition
« on: Nov 27, 2007, 08:47:04 »
Hmmm... Think I prefer the first version myself, probably because it has fewer saturated areas, and thus shows a bit more of the subtle detail. (Refers back to this thread.)


(Note: This discussion sub-thread started with my asking the following question...)

Are the crosses through the stars diffraction spikes from the scope's spider, or are they an artefact of the receiver or the processing?
« Last Edit: Nov 29, 2007, 12:26:06 by Rick »

Fay

  • O. A. S.
  • Galaxy Cluster
  • *
  • Posts: 9210
    • Faysastroimages
NGC 7635 Bubble Nebula
« Reply #1 on: Nov 27, 2007, 09:09:32 »
Processing, Rick
It is healthier to be mutton dressed as lamb, than mutton dressed as mutton!

Rick

  • Administrator
  • Galaxy Cluster
  • *
  • Posts: 6545
    • http://www.orpington-astronomy.org.uk
NGC 7635 Bubble Nebula
« Reply #2 on: Nov 27, 2007, 11:47:05 »
They are artificially added in using Astronomy Tools.
Erf. :( Not good. If they're an un-avoidable side-effect of the processing, then choosing options to minimise them is to be preferred. If they're added deliberately then I'd go further, and say it's a bad tool.

Fay

  • O. A. S.
  • Galaxy Cluster
  • *
  • Posts: 9210
    • Faysastroimages
NGC 7635 Bubble Nebula
« Reply #3 on: Nov 27, 2007, 12:26:56 »
I don't agree with too much processing, Rick. I mean, people like Nik Syzmanek & Hubble, also, nearly recreate their images with their great knowledge & application of effects & enhancements with programs like PhotoShop.

I think a spike is very minimal, I would rather not use false colour, & prefer that the image is 99.5% natural.

 
It is healthier to be mutton dressed as lamb, than mutton dressed as mutton!

Mike

  • Observing Consultant
  • O. A. S.
  • Galaxy Cluster
  • *
  • Posts: 8501
    • Electronicle
NGC 7635 Bubble Nebula
« Reply #4 on: Nov 27, 2007, 12:34:52 »
I'm with Rick. Processing is essential to get a good image, but diffraction spikes are an anomoly due to your particular optics and are to be removed where possible. Sure, they can look nice if used in the right way, but to me, unless you are specifically doing an image of a single star, they distract from the main object in the image.

Processing is a dark art that I have yet to get involved in heavily myself and a lot of images are ruined by over processing. i've seen some great images that are noisy or have artifacts that have been ruined by over-processing to get rid of the noise or artifacts or in an attempt to sharpen the image, when it would have been better of left alone.

We live in a society exquisitely dependent on science and technology, in which hardly anyone knows anything about science and technology. Carl Sagan

JohnP

  • Guest
NGC 7635 Bubble Nebula
« Reply #5 on: Nov 27, 2007, 14:24:08 »
umm... I could see this thread developing into a very long one......!

My thoughts on this are that if you want an accurate image as possible then obviously the addition of any artificial data such as diffraction spikes is a NO-NO. However, in some instances the addition of a diffraction spike adds to the overall image (it doesn't work for all subjects but certainly helps with some) - it kinda gives it the wow factor to an otherwise flat image.

Fay - you need to be very careful when adding the spikes - I think on your image just spikes on the bright star above the bubble would have sufficed. If you look closely you can see spikes on other stars & the actual centre of the bubble as well. If after you run the script in Astronomy tools & add as a layer on top you can go in & manually delete all the spikes you don't want.

It's a great image by the way - colour would really enhance it it. Good to see you posting so many images.

John

Mike

  • Observing Consultant
  • O. A. S.
  • Galaxy Cluster
  • *
  • Posts: 8501
    • Electronicle
NGC 7635 Bubble Nebula
« Reply #6 on: Nov 27, 2007, 15:51:25 »
I like it, I also remember John and Mike posting images with spikes????

Guilty! I used a cross of cotton thread over the end of my ED80 method. Though the spikes were small and subtle and I mainly did it to obtain better focus.
We live in a society exquisitely dependent on science and technology, in which hardly anyone knows anything about science and technology. Carl Sagan

Ian

  • Administrator
  • Galaxy Class
  • *
  • Posts: 3399
NGC 7635 Bubble Nebula
« Reply #7 on: Nov 27, 2007, 16:08:07 »
I quite like diffraction spikes, but they really are a sign of a coffee table image. Trouble is, if I see a refractor in the kit list and spikes on the stars I get suspicious...

I don't get that sort of spike from my newt as the secondary is held on a bar, I think I might put another bar across the other way to even it up. When I start imaging through it, anyway. (need a guidescope with an FL over a metre I think...)

Rick

  • Administrator
  • Galaxy Cluster
  • *
  • Posts: 6545
    • http://www.orpington-astronomy.org.uk
NGC 7635 Bubble Nebula
« Reply #8 on: Nov 27, 2007, 17:02:05 »
Genuine diffraction spikes are a fact of life. Certain optical arrangements will produce them. They're real optical phenomena, and, to mis-quote Scotty, "We canna' change th'laws o' physics, Capt'n!". ;)

If you have an accurate idea of exactly what diffraction signature your particular optical system produces, and you have sufficient compute power and the right software, then you might be able to remove most of the diffraction effects from an image. To do this accurately you'd probably have to do it frame-by-frame before stacking, though.

So for the most part we're going to be stuck with them whether we like them or not.

Now, if a processing system were to add fake diffraction spikes to make an image look "better", that would be bad science.

JohnP

  • Guest
NGC 7635 Bubble Nebula
« Reply #9 on: Nov 27, 2007, 17:44:17 »
Rick - I definitely wouldn't remove them if my optical system added them i.e. due to design of secondary holder & I have never heard of any software that can remove them...! I think that trying to remove diffraction spikes that are added due to optical system would be even 'LESS SCIENTIFIC' than adding diffraction spikes to images that didn't have them.....!

So Rick do your images have diffraction spikes or not...?

Cheers,  John.

Tony G

  • O. A. S.
  • Galaxy Class
  • *
  • Posts: 3050
NGC 7635 Bubble Nebula
« Reply #10 on: Nov 27, 2007, 19:34:25 »
Personally I like the diffractions ;)

Tony G
"I'm normally not a praying man, but if you're up there, please save me Superman." - Homer Simpson

Rick

  • Administrator
  • Galaxy Cluster
  • *
  • Posts: 6545
    • http://www.orpington-astronomy.org.uk
NGC 7635 Bubble Nebula
« Reply #11 on: Nov 27, 2007, 22:32:19 »
I think that trying to remove diffraction spikes that are added due to optical system would be even 'LESS SCIENTIFIC' than adding diffraction spikes to images that didn't have them.....!
Hmmm... You'd have to take that up with the Hubble folks, then. The algorithms they used to sharpen up the original camera's images (before they fitted COSTAR) are pretty much the same as the algorithms you could use to remove diffraction effects. They're all algorithms which take information that's been spread out one way or another and move it back to the position from which it came. They're not inventing fake stuff.

Rocket Pooch

  • Non-OAS
  • Galaxy Class
  • *
  • Posts: 5027
NGC 7635 Bubble Nebula
« Reply #12 on: Nov 29, 2007, 10:38:33 »
I think that trying to remove diffraction spikes that are added due to optical system would be even 'LESS SCIENTIFIC' than adding diffraction spikes to images that didn't have them.....!
Hmmm... You'd have to take that up with the Hubble folks, then. The algorithms they used to sharpen up the original camera's images (before they fitted COSTAR) are pretty much the same as the algorithms you could use to remove diffraction effects. They're all algorithms which take information that's been spread out one way or another and move it back to the position from which it came. They're not inventing fake stuff.

Hello,

Me again, the pre COSTAR images used de-convolution algorithms to remove as much as the optical effect as possible caused by the original mirror miss figuring.  That is now fixed and has nothing to do with diffraction spikes.

The post COSTAR image we’re fixed due to the corrective lens (Hubble glasses) which was put in front of the camera, or in their words “compensates for spherical aberration in Hubble's primary mirror”.

All the images produces by Hubble contain diffraction spikes, see below, however depending on the brightness scaling used they are either really easy to see or quite difficult to see.  Personally I don’t think they are removed.

If the “the Hubble guys” did remove the diffraction spikes then they are about as scientific as Fay’s image (nice image Fay, I now love the spikes, I like controversy).  Hubble people bad, very bad people, evil in fact and should be stoned (rocks not grass) for producing non spiky images.  What will they do next name teddy bears!

Anyway have a look here; you will see they don’t remove the spikes from images, well unless they are for public release for pretty pictures, and only rarely.

http://hubblesite.org/

See http://hubblesite.org/newscenter/archive/releases/2005/36/image/a/ for a good example of Diffraction spikes on Hubble, post COSTAR.  Also if you search for Diffraction spikes and you will see all the examples of them.

Also here’s a good one http://imgsrc.hubblesite.org/hu/db/2004/28/images/a/formats/print.jpg


Chris

Mike

  • Observing Consultant
  • O. A. S.
  • Galaxy Cluster
  • *
  • Posts: 8501
    • Electronicle
NGC 7635 Bubble Nebula
« Reply #13 on: Nov 29, 2007, 10:42:02 »
FIGHT! FIGHT! FIGHT! FIGHT!  :twisted:
We live in a society exquisitely dependent on science and technology, in which hardly anyone knows anything about science and technology. Carl Sagan

Fay

  • O. A. S.
  • Galaxy Cluster
  • *
  • Posts: 9210
    • Faysastroimages
NGC 7635 Bubble Nebula
« Reply #14 on: Nov 29, 2007, 11:58:43 »
I was watching a Hubble/ESA disc last night & noticed the spikes on some of the images.
It is healthier to be mutton dressed as lamb, than mutton dressed as mutton!