Starship Asterisk*

At the Limit of Diffraction

Explanation: Did you ever want to just look through the eyepiece of a large telescope in space? If you could, you would see a sharp view that was diffraction limited. Unaffected by atmospheric blurring that ultimately plagues earthbound observers, the angular resolution of your diffraction limited view would be determined only by the wavelength of light and diameter of the telescope lens or mirror; the larger the diameter, the sharper the image. Still, in this working earth-based snapshot a new active adaptive optics system (MagAO) is being used to cancel out the atmospheric blurring in a visual observation of famous double star system Alpha Centauri. Testing the system at the eyepiece of the 6.5 meter diameter Magellan Clay Telescope at Las Campanas Observatory, astronomer Laird Close is enjoying a historic diffraction limited view (inset) and the wide apparent separation of the close binary star system ... without traveling to low earth orbit.

<< Previous APOD

This Day in APOD

Next APOD >>

[/b]

Well, this might be the nerdiest thing to be jealous of, but I'm definitely jealous. It might be useful to compare the view they saw to an actual space telescope's view of the pair, and there just happens to be such a view available in Hubble's archive. It's not the prettiest thing because of the charge bleeds spilling out from the stars, but you get the idea. The Order's select few had a pretty good view.

Here's a lecture by Laird Close where he talks about this system and how he hopes this technology will ultimately enable us to image light coming off of exoplanets:

https://www.youtube.com/watch?v=SVeTncz ... e=youtu.be

Some of the talking points from this lecture:
- By analyzing Kepler data and extrapolating for what we believe the typical exoplanet population is like in the Milky Way, as many as 20% of stars in the MW may well have an Earthlike world in the habitable zone... that's one in five stars.
- NASA is working on a concept for a next generation large space telescope after James Webb whose goal is to be able to image Earthlike worlds in the habitable zone of sunlike stars. This will enable atmospheric analysis by means of spectroscopy, which would enable us to detect methane and oxygen in planet atmospheres. A detection of both of these gases is a strong indicator of a biosphere.

Upshot: We could have evidence for extraterrestrial life within the next couple decades. Really exciting!

This is from a series of lectures at University of Arizona covering various new developments in Astrobiology. The whole series was fascinating:

http://cos.arizona.edu/connections/life-in-the-universe

-s

Just imagine all the planetary nebulae that could be observed!

The sad thing that occurred to me is that the photo of this piece of beautiful cutting edge technology will cause most Americans to think those two stars are 4 inches apart.

geckzilla wrote:It's not the prettiest thing because of the charge bleeds spilling out from the stars, but you get the idea.

I take it "Charge Bleed" is what is causing the circles around the stars? Perhaps you have a link where I can read up on that?

Craine wrote:

geckzilla wrote:
It's not the prettiest thing because of the charge bleeds spilling out from the stars, but you get the idea.

I take it "Charge Bleed" is what is causing the circles around the stars? Perhaps you have a link where I can read up on that?

Geck is referring to the over exposed Hubble picture.

The small circles around the APOD stars are primary mirror Airy disks.

The large circles around the stars are secondary mirror Airy disks.

I wonder, are they using a 2" eye piece? I've only ever seen one of those once, and also had the chance to look through the telescope -- in the store.

Spif wrote:Here's a lecture by Laird Close where he talks about this system and how he hopes this technology will ultimately enable us to image light coming off of exoplanets:

https://www.youtube.com/watch?v=SVeTncz ... e=youtu.be

Some of the talking points from this lecture:
- By analyzing Kepler data and extrapolating for what we believe the typical exoplanet population is like in the Milky Way, as many as 20% of stars in the MW may well have an Earthlike world in the habitable zone... that's one in five stars.
- NASA is working on a concept for a next generation large space telescope after James Webb whose goal is to be able to image Earthlike worlds in the habitable zone of sunlike stars. This will enable atmospheric analysis by means of spectroscopy, which would enable us to detect methane and oxygen in planet atmospheres. A detection of both of these gases is a strong indicator of a biosphere.

Upshot: We could have evidence for extraterrestrial life within the next couple decades. Really exciting!

This is from a series of lectures at University of Arizona covering various new developments in Astrobiology. The whole series was fascinating:

http://cos.arizona.edu/connections/life-in-the-universe

-s

When I said the same thing back in the seventies.... I just got .... Deer in the headlights looks, and Cricket sounds...

Now they say the samething,... And I was just college kid.....

:---(===) *

Craine wrote:The sad thing that occurred to me is that the photo of this piece of beautiful cutting edge technology will cause most Americans to think those two stars are 4 inches apart.

Though I am not one who thought they are 4-inches apart...... ......the question did come up in my mind as to how far apart they really are. Do you know? Thanks.

SarbaGuha wrote:Though I am not one who thought they are 4-inches apart...... :D......the question did come up in my mind as to how far apart they really are. Do you know? Thanks.

The two orbit each other with an 80 year period, and an orbit of fairly high eccentricity (0.5), so the distance between them varies quite a bit, from about 11 AU to about 36 AU. And from our perspective looking at this system somewhat from on-edge, the apparent separation range is even greater, from 2 to 22 arcseconds.

I thought the "four inches"... Was the diffraction limit...or something...

Maybe someone should enlighten us, and indicate what it is???

:---(===) *

neufer wrote:

Craine wrote:

geckzilla wrote:
It's not the prettiest thing because of the charge bleeds spilling out from the stars, but you get the idea.

I take it "Charge Bleed" is what is causing the circles around the stars? Perhaps you have a link where I can read up on that?

Geck is referring to the over exposed Hubble picture.

The small circles around the APOD stars are primary mirror Airy disks.

The large circles around the stars are secondary mirror Airy disks.

Hah! I am learning new stuff.
Thanx!

Tho...with modern technology can't it be compensated for? From what I read it is something that can be anticipated, and so it may be possible to let a computer reduce that effect?

Clear days ahead, fairly warm.... Might get out and take some pics....


:---(===) *

Boomer12k wrote:Clear days ahead, fairly warm.... Might get out and take some pics....


:---(===) *

Nice. But the warmer weather will cause more atmospheric blurring, so you'll need one of these big ones with adaptive optics.

Craine wrote:Tho...with modern technology can't it be compensated for? From what I read it is something that can be anticipated, and so it may be possible to let a computer reduce that effect?

Only to a degree. The position of diffraction artifacts is predictable, as is the average intensity. But you can never know for certain if a given recorded photon was diffracted or represents a source present behind the diffraction structure. There's an inherent uncertainty on the relationship between the actual intensity of a source and the measured intensity (called "noise"), and that cannot be eliminated by any technology. So while you can theoretically subtract out most of the light from the diffraction artifacts, this is rarely done since anything behind them is probably below the noise threshold of the artifacts themselves.

That said, all modern professional telescopes have rigorous models of their optical performance, and those are used to correct data in various ways, as deemed relevant for any particular investigation.

Boomer12k wrote:
I thought the "four inches"... Was the diffraction limit...or something...

Maybe someone should enlighten us, and indicate what it is???

4 arc seconds of separation

Is it just a blip in the pic, or is that a planet at ten o'clock from Alpha Centauri A?

Craine wrote:

neufer wrote:

Craine wrote:I take it "Charge Bleed" is what is causing the circles around the stars? Perhaps you have a link where I can read up on that?

Geck is referring to the over exposed Hubble picture.

The small circles around the APOD stars are primary mirror Airy disks.

The large circles around the stars are secondary mirror Airy disks.

Hah! I am learning new stuff.
Thanx!

Tho...with modern technology can't it be compensated for? From what I read it is something that can be anticipated, and so it may be possible to let a computer reduce that effect?

The charge bleeds are those bright white lines coming off the stars. Each pixel of the CCD is like a bucket which can only hold so many electrons. Once that bucket is saturated, they start to overflow into adjacent pixels in the row. In this case the stars were overexposed because astronomers were looking for planets rather than at the stars themselves. I guess they were hoping to find a very large planet orbiting quite distant from the star since the star's point spread function and the charge bleed artifacts cover up closer ones. They didn't find any planets that time, but that was 20 years ago. It was only in that same year that the very first exoplanet discovery was confirmed. Exoplanet observational strategies were being formed and evolving rapidly during that time and soon after we saw a boom in exoplanet discoveries. These days so many are discovered that they can barely make the news.

To add on to what Chris already answered (while I was writing this post!): Airy disks, diffraction spikes, and all that stuff around the star are caused by the nature of light. We can't get around it any more than we can avoid anything else in nature, like gravity. As Art said, the stars are overexposed; a shorter exposure will avoid the charge bleeds. Another useful tool is a coronograph which blocks much of the incoming light of the star from reaching the detector. Often, the point spread function is subtracted from the image. For Hubble at least, its point spread function is very stable over time so it is possible to do this. Here is an example of combined use of a coronagraph and point spread function subtraction, allowing us to view circumstellar disks around several stars: http://hubblesite.org/newscenter/archiv ... web_print/

The James Webb Space Telescope will feature some entirely new technology invented specially for it occluding the light of bright sources. They're called microshutters and they're amazing. Astronomers will be able to use them like coronagraphs to selectively block out points of light they don't want to see, but with much greater precision and with as many sources as can fit on the detector instead of just one.

Oh, if you are confused about what a point spread function is, just think of it as everything you see when looking at a star. The star itself is a point smaller than a single pixel but its light spreads out from there because of light's wavelike behavior and how it interacts with a telescope's parts or even your eyes. When you see glare from a bright light, some of that is your own eye's unique point spread function.

geckzilla wrote:Often, the point spread function is subtracted from the image.

This isn't something that would be useful, nor even something that makes mathematical sense.

The PSF describes what a point source looks like on the image (to really make things complex, it typically varies with position on the image). If you know the PSF- which you generally do, because you can compute it and partly measure it directly by imaging stars- you can use that information to partially unscramble the distortions that can end up in a final image. In essence, you compare an actual star image to the PSF, and then try to figure out distortion filters that could change what you expect into what you observe. Once you find a filter that does this, you can make an educated guess about what the extended (non-point-source) parts of the image looked like before they were distorted. This process is called deconvolution, and while it uses knowledge of the PSF to sharpen things, it doesn't involve any subtraction of the PSF.

Deconvolution has to be used with great care. Not only does it typically introduce new artifacts, but convolution (applying a filter) is a trap-door operation. Except for some very special cases, you can't deterministically reverse a filtering operation (once something is blurred, you can never get back to the original sharp image). That's why I said deconvolution represents an educated guess about the unfiltered source data. But there are an infinite number of other sources that could produce the same results, with different filters. The technique is very useful, but it can also bite you in a big way.

bcaydavid wrote:Is it just a blip in the pic, or is that a planet at ten o'clock from Alpha Centauri A?

It's present in multiple Hubble images, too. Not sure what it is though I find it hard to believe that multiple observations by trained astronomers failed to notice a planet so I suspect it is something mundane.

Chris Peterson wrote:

geckzilla wrote:Often, the point spread function is subtracted from the image.

This isn't something that would be useful, nor even something that makes mathematical sense.

Alright. I'm not using my own words here, though. Take it up with the people who wrote the paper I got it from: http://adsabs.harvard.edu/cgi-bin/bib_q ... .148...59S

geckzilla wrote:

Chris Peterson wrote:

geckzilla wrote:Often, the point spread function is subtracted from the image.

This isn't something that would be useful, nor even something that makes mathematical sense.

Alright. I'm not using my own words here, though. Take it up with the people who wrote the paper I got it from: http://adsabs.harvard.edu/cgi-bin/bib_q ... .148...59S

It's not entirely clear from the paper what they're doing, but I wouldn't call it PSF subtraction. They appear to be generating from empirical data a template image of a coronagraphically occluded star and using it as a mask to subtract out some of the remaining background. They call it "coronagraphic point spread function template subtraction (CPSFTS), which is clearly something much more complex than simply subtracting the PSF from a star image.

In any case, the value of the PSF is mainly in its ability to back out some of the aberrations that get introduced during imaging.

Chris Peterson wrote:It's not entirely clear from the paper what they're doing, but I wouldn't call it PSF subtraction. They appear to be generating from empirical data a template image of a coronagraphically occluded star and using it as a mask to subtract out some of the remaining background. They call it "coronagraphic point spread function template subtraction (CPSFTS), which is clearly something much more complex than simply subtracting the PSF from a star image.

In any case, the value of the PSF is mainly in its ability to back out some of the aberrations that get introduced during imaging.

Well, multiple sources I have read (here's another I referenced before making that post) refer to it as PSF subtraction and while that may be a simplification, I don't know any better than to use the same terminology that I see used most frequently.

You guys realize that when I hold a camera all I do is point-n-click right?
k...not entirely true. But almost.
Once I had to write some code that would take images of really blurry text and try and reconstruct the text. So yeah...I am in kindergarten as far as manipulating image data goes.

Craine wrote:You guys realize that when I hold a camera all I do is point-n-click right?
k...not entirely true. But almost.
Once I had to write some code that would take images of really blurry text and try and reconstruct the text. So yeah...I am in kindergarten as far as manipulating image data goes.

Well, now you know what to expect next time you ask a bunch of imaging nerds a question. Give yourself some credit, though. Kindergarten is using a web app to apply meme text to cat pics.