Hubble's hidden treasures
August 23, 2012 7:21 PM Subscribe
Hubble's hidden treasures "Hubble has made over a million observations since launch, but only a small proportion are attractive images ... but the vast amount of data in the archive means that there are still many hundreds of beautiful images scattered among the valuable, but visually unattractive, scientific data that have never been enjoyed by the public. We call these pictures Hubble’s hidden treasures, and a few months ago, we invited the public to look through Hubble’s science archive to help us find them."
Makes all those questions over on AskMe seem a bit insignificant, eh?
posted by HuronBob at 7:38 PM on August 23, 2012
posted by HuronBob at 7:38 PM on August 23, 2012
This is great, a nice companion to my ongoing re-watch of Carl Sagan's Cosmos on Netflix, which I have not seen since childhood. I highly recommend it if you haven't seen it lately - 5 minutes into the first episode I was already freaking out about the existence of the universe.
posted by something something at 7:47 PM on August 23, 2012
posted by something something at 7:47 PM on August 23, 2012
This. This. This.
What's up with the strange aliasing/artifacts in the "hidden treasures" image? Otherwise, awesome.
When you see a tiled image with a quarter of it missing it's because that array has another type of camera in that quadrant. Gap artifacts are the result of two images being tiled together. There are other optical artifacts that occur at different resolutions.
posted by clarknova at 8:01 PM on August 23, 2012
What's up with the strange aliasing/artifacts in the "hidden treasures" image? Otherwise, awesome.
When you see a tiled image with a quarter of it missing it's because that array has another type of camera in that quadrant. Gap artifacts are the result of two images being tiled together. There are other optical artifacts that occur at different resolutions.
posted by clarknova at 8:01 PM on August 23, 2012
Recently I went to see the Hubble iMax. It was amazing and worth the pain those seats cause me.
posted by cjorgensen at 8:08 PM on August 23, 2012 [1 favorite]
posted by cjorgensen at 8:08 PM on August 23, 2012 [1 favorite]
You know, I was kinda into astrophotography when I was a kid; processing my 35mm black and white film in the school's darkroom. I had a wooden device with a screw hooked up to my tripod to track the stars without star trails.
To be honest, after Hubble started turning up images like this, I just kinda gave up.
posted by Jimbob at 8:25 PM on August 23, 2012
To be honest, after Hubble started turning up images like this, I just kinda gave up.
posted by Jimbob at 8:25 PM on August 23, 2012
What's up with the strange aliasing/artifacts in the "hidden treasures" image?
clarknova has the seams explained, but you're also seeing diffraction spikes on the bright stars.
posted by Wemmick at 8:29 PM on August 23, 2012
clarknova has the seams explained, but you're also seeing diffraction spikes on the bright stars.
posted by Wemmick at 8:29 PM on August 23, 2012
Awesome initiative. I hope they also publish the raw metadata on volunteer's ratings.
Hidden gems within hidden gems.
posted by porpoise at 8:59 PM on August 23, 2012
Hidden gems within hidden gems.
posted by porpoise at 8:59 PM on August 23, 2012
What's up with the strange aliasing/artifacts in the "hidden treasures" image?
Some of the artifacts are electronic - the same stars that show diffraction spikes look funny (dark) in the middle in this image as more photons are being captured than the electronics can count. In the versions of Hubble images made for press releases (i.e., the ones you've generally seen), such artifacts are generally photoshopped away.
In most cases, any gaps you see are probably gaps between CCD chips in the detector array rather than due to tiling (I'd expect most of these images to come from the WFPC2 camera, which had 4 chips in a chevron pattern, or ACS, which has 2 separate chips as you can see clearly here). Generally when we make mosaics of images with Hubble we try to make sure there are no gaps (partially by overlapping tiles, and partially by shifting between each image of a single tile to fill in the chip gaps).
In many of these images, the seams are apparent not because there is no data there, but because the image is noisier in regions where there was a shorter exposure time (e.g., because for half the exposure time there was a chip gap there) - the eye picks up on that easily.
posted by janewman at 9:52 PM on August 23, 2012 [1 favorite]
Some of the artifacts are electronic - the same stars that show diffraction spikes look funny (dark) in the middle in this image as more photons are being captured than the electronics can count. In the versions of Hubble images made for press releases (i.e., the ones you've generally seen), such artifacts are generally photoshopped away.
In most cases, any gaps you see are probably gaps between CCD chips in the detector array rather than due to tiling (I'd expect most of these images to come from the WFPC2 camera, which had 4 chips in a chevron pattern, or ACS, which has 2 separate chips as you can see clearly here). Generally when we make mosaics of images with Hubble we try to make sure there are no gaps (partially by overlapping tiles, and partially by shifting between each image of a single tile to fill in the chip gaps).
In many of these images, the seams are apparent not because there is no data there, but because the image is noisier in regions where there was a shorter exposure time (e.g., because for half the exposure time there was a chip gap there) - the eye picks up on that easily.
posted by janewman at 9:52 PM on August 23, 2012 [1 favorite]
harff...phlarpfff whoo um....oooooh.
whew. Bends the mind. I'm glad I had an hour to kill.
Thanks.
posted by mule98J at 10:39 PM on August 23, 2012
whew. Bends the mind. I'm glad I had an hour to kill.
Thanks.
posted by mule98J at 10:39 PM on August 23, 2012
These are breathtaking.
posted by Cocodrillo at 1:24 AM on August 24, 2012
posted by Cocodrillo at 1:24 AM on August 24, 2012
Jimbob: You can get breathtaking images without Hubble still! Coincidentally the Astronomy Photographer of the Year competition entries have been released, and you can see from previous years just how amazing the images some people get are. There's a non-trivial outlay of cost to start getting images like that of course, and it helps to be in a good location.
(I enjoy the challenge of fighting off the glow from streetlights and dealing with poor conditions, but that's probably just me...)
posted by edd at 2:37 AM on August 24, 2012 [2 favorites]
(I enjoy the challenge of fighting off the glow from streetlights and dealing with poor conditions, but that's probably just me...)
posted by edd at 2:37 AM on August 24, 2012 [2 favorites]
How many images are there? I find it odd that this magnificent telescope needs crowd sourcing to find interesting images. Is it on autopilot? It's like me asking the Internet to find pictures of my kids in my archive. Don't they know what they're aiming at when they take the image?
posted by pashdown at 6:26 AM on August 24, 2012
posted by pashdown at 6:26 AM on August 24, 2012
pashdown: When astronomers get their images they're not usually aiming to get a pretty result. An image can be scientifically very useful even when it has a few flaws in it, and in addition to that actually turning a data file with what can be a tremendously large dynamic range into a pretty picture is a genuine art, and not one that many astrophysicists have dabbled in. HST employs people to spend a lot of time producing the gorgeous shots you usually see, from multiple exposures (you have to combine multiple bands into colour images, in addition to the tiling mentioned above).
As for how many images there are - in 1998 there were 150,000 images with 5TB of data although not all of that was astronomical imaging (a lot of calibration etc needs to be done too). Now there'll naturally be vast amounts more.
posted by edd at 6:39 AM on August 24, 2012 [1 favorite]
As for how many images there are - in 1998 there were 150,000 images with 5TB of data although not all of that was astronomical imaging (a lot of calibration etc needs to be done too). Now there'll naturally be vast amounts more.
posted by edd at 6:39 AM on August 24, 2012 [1 favorite]
pashdown: To build on edd's comment - Hubble generally takes at least 2 exposures of data per 90 minute orbit (too many more is inefficient, as it takes several minutes to read out the detectors or even longer to move the telescope, while with only 1 image it is difficult to remove cosmic ray hits from your data -- even with 2 images there will often be a few pixels that got hit in both frames, making those bits of the data useless). That gives you a minimum of 11000 individual images per year (or >250,000 over the lifetime of the telescope). This FAQ says that Hubble has taken >1,000,000 observations to date, but I expect that includes a lot of calibration data (which would often be taken repeatedly with shorter exposure times).
posted by janewman at 11:54 AM on August 24, 2012
posted by janewman at 11:54 AM on August 24, 2012
after Hubble started turning up images like this, I just kinda gave up.
These images should include the disclaimer: Cosmetically doctored photos. These are not what you'd see in a telescope. Or maybe "Pimping the Hubble."
Live views of the unvarnished universe have the natural beauty of reality ... unlike magazine covers with "artistically enhanced" images of impossibly fantastic subjects. Let's hope the originals aren't lost.
posted by Twang at 12:53 PM on August 24, 2012
These images should include the disclaimer: Cosmetically doctored photos. These are not what you'd see in a telescope. Or maybe "Pimping the Hubble."
Live views of the unvarnished universe have the natural beauty of reality ... unlike magazine covers with "artistically enhanced" images of impossibly fantastic subjects. Let's hope the originals aren't lost.
posted by Twang at 12:53 PM on August 24, 2012
Twang: Fortunately astronomy is big on archiving - the 'artistically enhanced' versions are of no use for research. The raw data from missions going back to the 1970's are available for use.
For HST, not only is all the old data online (from here and elsewhere), but you can even apply for funding to analyze it in new ways. At this point, most papers using HST data are making use of these archives (see here).
posted by janewman at 1:50 PM on August 24, 2012
For HST, not only is all the old data online (from here and elsewhere), but you can even apply for funding to analyze it in new ways. At this point, most papers using HST data are making use of these archives (see here).
posted by janewman at 1:50 PM on August 24, 2012
« Older Manufacturing Company Becomes a Design Firm, Rips... | How to have a career: advice to young writers Newer »
This thread has been archived and is closed to new comments
posted by Joakim Ziegler at 7:32 PM on August 23, 2012