Author Topic: Sunday, 4 May, 2014: Diffraction Spikes, of the SDSS Kind  (Read 5969 times)

JeanTate

  • OotD posters
  • Hero Member
  • *****
  • Posts: 2976
    • View Profile
Sunday, 4 May, 2014: Diffraction Spikes, of the SDSS Kind
« on: May 04, 2014, 04:14:12 pm »
We Galaxy Zoo zooites have threads, here in the forum, for many different kinds of galaxy - barred, merging, ellipticals (and many more) - catalogs (e.g. Kiso Ultraviolet Galaxy Catalogue, NGC Catalogue, Flat Galaxy Catalogue), nebulae (e.g. The Nebulae Collection (formed from many threads)), comets, asteroids, Hubble stars, conjunctions of stars and galaxies, faces, oddballs, .... *



That's the DR7 image of SDSS J013308.42-000412.3; yes, it's a nice galaxy/star conjunction, and the star itself is rather cool, but I think one of the most spectacular aspects is the 'compass points', the four diffraction spikes. Or 'diffspikes', as I call them. Given our extreme enthusiasm for making collections, I was slightly surprised to discover that there is no specific thread devoted to diffspikes!  :P At least, none that I could find. Of course, there's a very long thread which has many excellent images of SDSS diffspikes, The best Stars!, and all of the diffspike images in today's OotD come from that thread^ (unless otherwise noted).

What are diffspikes?

Almost all reflecting telescope designs have something in front of the main mirror, an obstruction. For the main 2.5m SDSS telescope, that obstruction is the 1.08m secondary mirror:



That secondary mirror needs to be supported; its support is four 'spiders', vanes which extend from the main (exterior) trusses to the secondary mirror and its support structures. This main SDSS telescope has an altitude-azimuth mount, and the spiders are at ~45° to the horizontal. You can just make these spiders out in this image:



And in this image, you can see more clearly just how much hardware there is, 'inside' the telescope (not just the secondary mirror and its spiders!) (source):



Enough about telescopes, more diffspikes!



SDSS J095206.32+540343.6, SDSS J021856.66+283822.7, and SDSS J124420.22-084017.0 (the last one is a repeat of SDSS J013308.42-000412.3).

Although the spiders are fixed, and as the telescope has an alt-azimuth mount, the angle the diffspikes make with North (up in all images; East to the left) is not always ~45°. As is clear in these four images.



SDSS J023605.49+065321.3, posted by AlexandredOr, with the caption "A star reflected in a broken mirror". There seems to be at least two sets of diffspikes here, can you work out why?



That's SDSS J132107.00+471631.2, posted by EricFDiaz, as the only post in the Old Red thread. He wrote, "Look at the diffraction spikes on this baby. You can see spectral colors in the spikes."

Spectral colors? As far as I could find, this is the only reference, here in the GZ forum, to diffspikes having "spectral colors", or similar (if you come across any others, please let me know!). Once it's pointed out to you, you'll see these spectral colors in many, perhaps most, SDSS images of diffspikes.

And indeed, diffspikes are low grade spectra; do you think it's possible to analyze the data in the five FITS files that are the digital record of the five SDSS images (one for each band, u, g, r, i, and z), and produce a spectrum?

If you're interested in learning more about the optics, how obstructions cause diffspikes (and more), etc, I found this Beugungsbild.de webpage (by René Pascal, "... Aesthetics and Physics of Imaging Methods") very well-written and easy to follow (it also has cool images): Diffraction Pattern of Obstructed Optical Systems. And DuckDuckGo (a cool search engine) will find you lots of good sites which explain the physics of diffspikes in as little, or as much, detail as you'd like.

* A Classifier's Guide To The Objects Topics (Or Where To Post That Great Galaxy) is a great resource for finding many of these thread, but not all of them
^ the ones I used were posted by Hofi (0133-0004), elizabeth (0952+5403), Mjtbarrett (02182838), and Mukund Vedapudi (1244-0840).

AlexandredOr

  • Hero Member
  • *****
  • Posts: 25423
    • View Profile
Re: Sunday, 4 May, 2014: Diffraction Spikes, of the SDSS Kind
« Reply #1 on: May 04, 2014, 04:17:03 pm »
Yes !!! Original OotD  :D and instructive too  :D :D
Many thanks Jean !

Baby star opening its eyes on the Universe.

elizabeth

  • Hero Member
  • *****
  • Posts: 33783
    • View Profile
Re: Sunday, 4 May, 2014: Diffraction Spikes, of the SDSS Kind
« Reply #2 on: May 04, 2014, 04:39:01 pm »
very 8) OOTD  ;D

djj

  • OotD posters
  • Hero Member
  • *****
  • Posts: 24195
    • View Profile
Re: Sunday, 4 May, 2014: Diffraction Spikes, of the SDSS Kind
« Reply #3 on: May 04, 2014, 05:40:12 pm »
Very interesting, JT 8) – and I love the use of the term 'spider' here :D.

Can we take it that it's only stars that can be bright enough to cause this photographic effect?

JeanTate

  • OotD posters
  • Hero Member
  • *****
  • Posts: 2976
    • View Profile
Re: Sunday, 4 May, 2014: Diffraction Spikes, of the SDSS Kind
« Reply #4 on: May 05, 2014, 06:47:35 pm »
Thanks everyone.

This was quite fun; not only is it rather interesting, but as I said, it seems only EricFDiaz thought the 'spectral colors' aspect of (most) diffspikes worthy of being commented on.

@djj: you might get a further kick out of some others; 'vanes' surely doesn't describe the glorified wires very well, does it? Also, my instinct was to write about the 'telescope tube', but clearly those robust interlocking pipe-like things aren't! (I'm not sure 'truss' is both accurate and the term which astronomers actually use)

"Can we take it that it's only stars that can be bright enough to cause this photographic effect?"

I think it's rather more subtle than that. And the Beugungsbild.de webpage (link in the OP) does a pretty good job of showing that it's clearly not just brightness (Mars is brighter than almost all stars, yet shows no such effect, in the images displayed).

For diffspikes - and ignoring all the other optical effects, including what goes on inside the camera itself - a 'star' is simply a 'point source'. So if the SDSS camera images a star, and it shows no diffspikes, in principle you simply have to increase the exposure time until it does. From this you might conclude that there should be lots more diffspiked stars in Stripe 82 images; after all, they are produced by 'stacking' lots of separate images, in effect increasing the exposure time.

But have you ever seen such a diffspiked star, in a Stripe 82 image?

For example, here's a DR8 image (star is SDSS J011615.77+001001.4, seems on the verge of having diffspikes), and next to it a Stripe 82 version (AHZ600072n, much more real estate taken up by the star, but no diffspikes):



A galaxy is just a whole lot of point sources, so the center of a really bright galaxy - especially a Seyfert - should show diffspikes, right? But SDSS images of such galaxies - and there are indeed some - don't. How come?

(There are lots of examples of stars with diffspikes and galaxies without in the Conjunctions galaxy/star thread.)

zutopian

  • Hero Member
  • *****
  • Posts: 2852
  • REOPEN this forum!!!
    • View Profile
Re: Sunday, 4 May, 2014: Diffraction Spikes, of the SDSS Kind
« Reply #5 on: May 05, 2014, 07:04:15 pm »
Here is the SDSS image of the famous Quasar 3C273.: It looks like a star with diffraction spikes and a jet is visible.:


http://cas.sdss.org/astro/en/tools/explore/obj.asp?id=587726014535237707

PS: It was presented as an "Object of the Day" before.
« Last Edit: May 06, 2014, 07:35:01 pm by zutopian »

djj

  • OotD posters
  • Hero Member
  • *****
  • Posts: 24195
    • View Profile
Re: Sunday, 4 May, 2014: Diffraction Spikes, of the SDSS Kind
« Reply #6 on: May 05, 2014, 07:17:02 pm »
JT — I'm not well-enough educted in physics to understand all thet René Pascal says on the beugungbild.de site, but I note that he too speaks of 'these diffraction figures that often show up in rainbow colors'.

JeanTate

  • OotD posters
  • Hero Member
  • *****
  • Posts: 2976
    • View Profile
Re: Sunday, 4 May, 2014: Diffraction Spikes, of the SDSS Kind
« Reply #7 on: May 05, 2014, 07:46:49 pm »
Thanks zutopian!  :)

That makes the point that it's the optical nature of the source, not its physical nature, which determines whether there are diffspikes or not (3C273 is, for SDSS, a point source, no different from a star (ignoring the jet)).

JT — I'm not well-enough educted in physics to understand all thet René Pascal says on the beugungbild.de site, but I note that he too speaks of 'these diffraction figures that often show up in rainbow colors'.

Oops, sorry.  :(  :-[

The images of Mars I found particularly interesting; the 'perfect' one has many features which differ only slightly in terms of their surface brightness, and so if obstructions reduce contrast - which they do, because they 'spread' the light further - then it should show up in the images. And indeed it does, but it's pretty minor, even for significantly large obstructions (yes, it's mostly about the Airy disks, but still ...).

The point - well, one point - about diffspikes is the fraction of the total incoming light scattered (diffracted) into the diffspikes is very small. The reason we 'see' them so clearly is that there's such a huge 'dynamic range' in the images (and our eyes have the ability to deal with dynamic ranges far greater than can be produced on the usual computer monitors/screens we are likely to have).

To get more technical: the SDSS images we see are JPEGs, which have 128 different intensity levels (or is it 256?). If you ever get to see the cleaned "raw" data in an SDSS image displayed as a JPEG with a linear scale*, you'll surely think everything is washed out; change the scaling to "Logarithm" and more features are visible (this scale greatly increases the dynamic range displayed; the data are unchanged). And even this understates the huge range ... the core of a star with diffspikes is usually 'saturated' in at least one band.

Hope this helps ...

* and you still can, when the "View raw data in different wavelengths" tool, in GZ is working ... select an image, a band, and "Linear"; AGZ0003no3, a star with diffspikes, for example

djj

  • OotD posters
  • Hero Member
  • *****
  • Posts: 24195
    • View Profile
Re: Sunday, 4 May, 2014: Diffraction Spikes, of the SDSS Kind
« Reply #8 on: May 05, 2014, 08:09:54 pm »
Thanjs, JT. There was absolutely no need to apologize, and forgive me if you've already covered this question, but why (in a nutshell, please ;D) don't all point sources (stars or quasars) cause diffspikes? In other words, what is the extra something that they need to do this? Or is it still a bit of a mystery?

JeanTate

  • OotD posters
  • Hero Member
  • *****
  • Posts: 2976
    • View Profile
Re: Sunday, 4 May, 2014: Diffraction Spikes, of the SDSS Kind
« Reply #9 on: May 05, 2014, 09:00:07 pm »
Thanjs, JT. There was absolutely no need to apologize, and forgive me if you've already covered this question, but why (in a nutshell, please ;D) don't all point sources (stars or quasars) cause diffspikes?
They all do!  :o

Ignore, for the moment, complications to do with how things differ between bands and other effects which cause a star to appear to be bigger than a single pixel (and so on), and look at how diffspikes 'grow' as a star gets brighter.

I'll use this field (DR7 image) to illustrate; it's centered at ~(19.0,0.10)*:



At this scale, three stars have obvious diffspikes (all images have the same scale and orientation):



And so do some fainter stars, but you have to look more closely (well, a bit more closely, in the zoomed-out image):



Going fainter still, the diffspikes start to disappear:



Fainter still and they've vanished, in these images (though your eyes may be sharp enough to see some tiny traces, at the base):



As stars, these are all quite bright (by SDSS standards!), and likely saturate at least some pixels in at least one band at their centers. But now the diffspikes are so faint they are 'lost in the noise'. It's possible that some really clever image processing could bring them out, in the stars in the last row, but in even fainter stars? I don't think so.

If these stars were imaged by the Hubble, with a sufficiently long integration time, they'd all show diffspikes (but somewhat different ones, as the Hubble's 'spiders' are different from the SDSS's, and the orientation different too). Ditto if imaged by an 8m telescope (the main SDSS telescope is only a 2.5m one). And so on.

Does this help?

Quote
In other words, what is the extra something that they need to do this? Or is it still a bit of a mystery?

Yes, there are additional 'extra somethings', but they are pretty minor, for regular SDSS images ... (no mysteries, that I know of)

* and I'm going to be a bit lazy and not identify each star, by name

zutopian

  • Hero Member
  • *****
  • Posts: 2852
  • REOPEN this forum!!!
    • View Profile
Re: Sunday, 4 May, 2014: Diffraction Spikes, of the SDSS Kind
« Reply #10 on: May 06, 2014, 05:50:35 am »
Stripe82: Star with diffraction spikes:


8647475119813952220

PS: previously posted

djj

  • OotD posters
  • Hero Member
  • *****
  • Posts: 24195
    • View Profile
Re: Sunday, 4 May, 2014: Diffraction Spikes, of the SDSS Kind
« Reply #11 on: May 06, 2014, 08:59:09 am »
That was a brilliant explanation, JT. Thanks very much indeed. You're a star (as they say) and your diffspikes are superb ;D!

JeanTate

  • OotD posters
  • Hero Member
  • *****
  • Posts: 2976
    • View Profile
Re: Sunday, 4 May, 2014: Diffraction Spikes, of the SDSS Kind
« Reply #12 on: May 06, 2014, 12:56:32 pm »
Thank you zutopian, and djj!  ;D

Stripe82: Star with diffraction spikes:


8647475119813952220

PS: previously posted

The star is SDSS J212237.03-010112.3, and here's its DR8 image (centered on the star, not the 'galaxy'; the scale is approximately the same, as is the orientation):



This star is so bright that when the many separate images are stacked the many separate diffspikes remain visible (it's also possible - likely? - that some of the brighter Stripe 82 diffspikes are actually composed of several separate ones, not-quite overlapping ... some diffspikes seem thicker, and have a fuzzier appearance). Notice that the two other very bright stars, with obvious diffspikes in the DR8 image, have only very faint ones in the Stripe 82 image. And some fainter stars which show small DR8 diffspikes are obviously brighter in the Stripe 82 image, but have no diffspikes. For example (these are the yellow stars, with short horizontal lines in their centers, in the top-right corner in the Stripe 82 image):

« Last Edit: May 06, 2014, 12:58:28 pm by JeanTate »

c_cld

  • Hero Member
  • *****
  • Posts: 4948
  • "In some fields, I am at home; in others we are fellow tourists" (Bill )
    • View Profile
Re: Sunday, 4 May, 2014: Diffraction Spikes, of the SDSS Kind
« Reply #13 on: May 06, 2014, 02:55:35 pm »
Messier 34 (cluster in Perseus )

bunch of diffraction spikes  8)
http://dr10.sdss3.org/sas/dr10/boss/photoObj/frames/301/3643/1/frame-irg-003643-1-0019.jpg


c_cld

  • Hero Member
  • *****
  • Posts: 4948
  • "In some fields, I am at home; in others we are fellow tourists" (Bill )
    • View Profile