Slashdot Mirror
The Deepest Photo Ever Taken
Astroturtle writes "Astronomers using the Hubble Space Telescope's powerful new Advanced Camera for Surveys (ACS) have taken the deepest visible-light image ever made of the sky. The 3.5-day (84-hour) exposure captures stars as faint as 31st magnitude, according to Tom M. Brown (Space Telescope Science Institute), who headed the eight-person team that took the picture."
Imagine a Beowulf... um. Seriously, how do you cope with reciprocity failure in a 3.5 day exposure. I would have thought that stray heat or electron flow would turn the whole image to static with such a long exposure. HST must consist of unfathomably cool (literally and figuratively) electronics.
--
BitTorrent in C -- LibBT
http://www.sf.net/projects/libbt
3.4 day exposure? Even for a space-based platform, that has to be really stable to produce a good image. Does anyone out there have any info on how they maneuver the telescope to keep it pointing at the same point while minimizing shifts in the field?
Derek
Don't Panic...
It is still pretty incredible...pointing an object the size of a bus and accurately focusing it on something the size of a spec of sand...really, really, really far away. All while moving at a relative 26,000 miles an hour or whatever to keep it up in the sky...Not to mention the orbial speed of the earth itself... Only took 8 guys, several computers, and millions of dollars worth of equipment. Oh yeah, and that one maintenance run made a few years back to keep it pointing straight.
Who is this that even the wind and the waves obey Him? Surely this computer must submit also!
And, to think I used to complain about having to get the tripod out for exposures that were longer than 1/8th of a second! I'll never comlpain about slow film or lenses again!
Yeah, and you'd think NASA could afford 1600 ASA film for the price they paid for hubble...
I mean geez!
Direct link to the full-resolution JPEG. (~4.9MB)
/ a/formats/full_jpg.jpg
http://imgsrc.hubblesite.org/hu/db/2003/15/images
See also the press release with tons of photos. Enjoy your new wallpaper ! :)
Here is a link to a higher resolution image.
Hubblesite.org
The image is not actually a single exposure of 3.5 days in duration, but is actually made from 250 separate exposures taken from Dec. 2 to Jan. 11, 2003. The total exposure time was 3.5 days.
For those who are interested, the original hubble press release is located here.
The site includes the image in a variety of different formats, including a 123 MB tiff file.
For the love of all things scientific, have mercy on their 122MB TIFF image.
And to think that we've turned servers into slag by Slashdotting a 43kb page.
It all goes downhill from first post
As they spin, the momentum from their motion causes the telescope to move.
Well, it's techincally a litter different than that. The wheels don't actually cause hubble to translate within a plane. Instead they rotate hubble. By turning the spinning wheels, a torque is exerted on hubble, causing it to rotate.
neurostardesktop background ever created :) Its sure worth the effort, however!
Modern optical/IR/UV telescopes typically have a large primary mirror, which reflects light back to a smaller secondary, which reflects the light through a small hole in the primary to the detectors. The secondary is supported by little rods. It is diffraction of light by those supports which cause stars to have distorted shapes.
(Astronomers understand the diffraction issues very well... it's usually not a problem; it just looks weird.)
- A friendly neighborhood astrophysicist
The grandeur of such an image almost forces one to reasses their place in the world. To think that the area in the photograph is equivalent to the area covered by a grain of sand at arms length is mindnumbing. The universe is unbelieveably amazing.
The best way is to download the processed HST images and see what the count rate is for a faint star. Then multiply by the gain (in the header of the image) which will give you the number of photons detected. A way to guestimate the number of photons is to compare the flux of the faintest star with the Sun. At the Earth's distance the Sun has a flux of 1.36x10^6 erg s-1 cm-2 and the apparent mag of the sun is V=-26.8. If we assume that we have a star with V=31 mag (the 50% completeness level is V=30.7 mag) then the flux recieved from the star is given by: F2/F1 = 100^((m1-m2)/5) where F1 and m1 are the flux and magnitude of the sun and F2 and m1 refer to the star. This gives 1.03x10^-17 erg s-1 cm-2. Convert the ergs into photons by the de Broglie frequency (E=hv) where we assume that a V-band photon has a wavelength of 550nm or a frequency of 5*10^14 s-1. Thus, each photon carries 3.61x10^-12 ergs which gives a rate of 2.85x10^-6 photons s-1 cm-2. So a 3.5 day exposure is 302400 secs and HST has an aperature of 240 cm so we get about 50000 photons at the entrance of the telescope. Remember.. detection of these sources means having a low background so that these photons are not lost in noise! I should also point out that HST does not leave the shutter open continuously for 3.5Hs, instead it takes a series of short exposures that are co-added. I hope this helps (and doesn't freak people out!)
I'd hate to have to hold my finger on the button for that long without shaking the camera.
*This is a lame joke*
it is only after a long journey that you know the strength of the horse.
The "streaks" centering on stars are diffration spikes from the secondary mirror support. The colour alternates as different wavelenghts cause different diffration spacings.
The big bright cluster is actually a member of Andromedae (M31). Very impressive! The appearance of fuzziness is because the CCD oversamples the resolution of the telescope - which is necessary for good photometry - if you want it "sharp" then just bin the pixels by 2x2 or 3x3 or whatever looks best!
http://wuarchive.wustl.edu/users/tom/mirrors/hu
is a mirror of the full JPEG - about 5M. Enjoy.
Something I've wondered for a while... what's up with the points coming off the stars?
As was mentioned in another post, those are diffraction spikes from the supports for the secondary mirror.
Newtonian reflectors and classical Cassegrain telescopes support their secondary mirror with "spiders" that produce diffraction spikes. There have been various efforts over the years to eliminate these from that type of telescope. One method is to seal the tube with an optical flat (a flat piece of optical glass) which supports the mirror. The trade-offs include longer times for the scopes to reach temperature equilibrium, distortion from imperfections in the optical figure of the flat, and slight light loss. Other attempts have included the use of spiders with curved support arms, which reduce or eliminates spikes at the cost of slightly degraded overall image contrast.
Other telescope types, such as refractors, Maksutovs, Schmidt-Cassegrains, and Yolo reflectors have no diffraction spikes, but they are all more optically complex (Yolos, for instance, require toroidal mirrors) and are more difficult to produce as a result. Refractors have the added problem of chromatic abberation, which is the fringing of color on the edge of bright objects. Various complex, multi-element objectives have been developed to reduce, or even practically eliminate, this problem. The problems are optical complexity, cost, and light loss. Figuring a 3-element objective lens for a refractor means grinding six optical surfaces with precise curves. Compare that to a Newtonian which has a single parabolic primary mirror and a flat optical secondary.
There are many other telescope types than the few popular types I mentioned here and each have their proponents. Most designs that have survived the test of time can be made to perform well, but each has trade-offs.
Does anyone know if there is a BitTorrent file out for the 128mb TIFF? the nasa servers are a bit slow and I feel my hardware cycles and bandwidth could be of use...
Unretouched excerpt from full-resolution image.
--
I don't want to rule the world... I just want to be in charge of mayonnaise.
Actually, the really unique thing about this image is the stellar populations. The stars you see in the image are almost all in the Andromeda galaxy (aka M 31), seen here.
M 31 is 2.2 million light-years away. This is the galaxy that Hubble originally resolved into stars, thereby settling the Shapley-Curtis debate on the true scale of the Universe. However, the stars Hubble saw were the very brightest supergiants in M 31. In this HST image, we see stars 2 magnitudes fainter than the ancient main-sequence turn-off; i.e., stars which are intrinsically fainter than our Sun! This lets us learn a lot about the ages and chemical composition of M 31's halo stars, which turn out to be quite different from the stars in our halo (our halo is entirely composed of ancient, metal-poor stars; M 31's halo contains stars that are only 6 Gyr old, and much more metal-rich than our halo).
I heard Tom Brown give a talk on this work last week; very cool stuff.
Liberal (adj.): Free from bigotry; open to progress; tolerant of others.
You're right, if you take deepest image to mean "image of most distant objects" instead of "faintest objects". However, the Universe is 13.7 Gyr old, not 20 Gyr.
8 _ilc_64 0.jpg
Here's your deepest image then:
http://map.gsfc.nasa.gov/m_ig/020598/02059
That's from the recent WMAP mission, which mapped the cosmic microwave background in exquisite detail, pinpointing the age of the Universe (and many other cosmological parameters) to high precision. You're looking at an all-sky image of the Universe as it looked when it was 100,000 years old, and became transparent for the first time. IOW, you are literally seeing the fires of creation.
Liberal (adj.): Free from bigotry; open to progress; tolerant of others.
Ok, here's the calculation for you curious types, regarding how many photons arrived from the faintest star in the picture:
Let's suppose that the picture was taken in the "V" filter. I just happen to have the number of photons per second per meter squared that arrive from a star of 20th magnitude: 86.157. (taken from here ).
So the faintest stars in this picture are 31st magnitude? That's 11 mags fainter than 20, which by the handy old formula
mag1-mag2 = -2.5 * log(flux1/flux2)
which means that the 30th magnitude star puts out about 4x10^(-5) times as much flux.
Using the reference star's flux from above, this means that 0.0034299 photons per second per meter squared arrived at Hubble. The exposure was 84 hours, and the area of Hubble is (2.5m)^2*pi, so tada:
The total number of photons in the picture from the faintest star is: 20365.83
Still not too shabby. They probably could have found even fainter stuff.