Hubble is a cyclops

By Daniel Holz | April 13, 2010 10:06 pm

A few days ago the following headline on the New York Times website caught my eye: Seeing What the Hubble Sees, in Imax and 3-D. There are two reasons this headline is worthy of note. First, it is amazing that an IMAX movie about the Hubble Space Telescope exists at all, and is worth mentioning on the front (web)page of the NYT. NASA-Apollo8-Dec24-EarthriseHubble is a part of the popular imagination, and may be the object most closely tied with Science in the eyes of the general public (even more than the LHC). Furthermore, it is absolutely astounding that NASA launched hundreds of kilos of camera equipment and film into orbit, and spent valuable astronaut time (both on the ground and in space) to pull off the filming. I would claim one of the lasting legacies of the Apollo missions to the Moon are the photographs, and in particular Earthrise. That single photograph of our home as a small blue marble against the vastness of space put our planet into proper perspective for the very first time. NASA is well aware that part of its mission is to light up the public imagination, getting us to peer past our limited horizons, and out into the vast Universe beyond. This film is part of that tradition.

The second interesting aspect of the headline is that it’s nonsensical. Hubble has only one eye. It has one mirror. It can’t perceive depth, and therefore can’t see in 3-D. We see slightly different images in each of our eyes, and then a fairly impressive difference engine (called a “brain”) figures out the depth to everything we are looking at, and whether that rock is about to bonk us on the head and we need to duck NOW! 3D movies (such as Alice and Avatar) use circularly polarized light, and glasses with different filters in each lens, to produce the different images for each eye. (The light is circularly polarized so that, if you tilt your head, it all still works; the old linear polarization approach didn’t do this, and had a tendency to make one feel motion sick [at least, it did for me]).

In general astronomical sources are too far away for us to discern distance using parallax. That’s why the night sky looks “flat”, even though the planets and stars and galaxies are at a tremendous range of distances. If you wanted to be able to directly “see” the distance to the nearest star, in the same way that you ascertain the distance to an approaching lion, your eyes would need to be separated by roughly 10 billion km. (Eye separation = Distance*Angle. The human eye has an angular resolution of roughly 1 arcmin = 0.0003 radians, and the nearest star [Proxima Centauri] is 4 lightyears = 3.8e13 km.) The way we figure out distance in Hubble images is by using color information (and, in particular, the spectra) to discern recession velocity (redshift), and thereby distance (using Hubble’s law). This is not something we do with our eyes (although we do use color information to discern temperature; you’re unlikely to grab something that is so hot it’s glowing blue). Hubble sees a purely two-dimensional Universe. So “seeing what the Hubble sees …in 3-D” is a contradiction in terms. Was the headline carefully crafted to see if we were paying close attention?

ADVERTISEMENT
  • http://mengbomin.wordpress.com/ Meng Bomin

    I believe that in the 3-D IMAX, they did some image manipulation allow for a stereoscopic effect in displaying the images. I haven’t seen the show and probably won’t so I can’t comment on how it turned out, but the enterprise of manipulating Hubble images to give them a stereoscopic effect makes me a bit uneasy for some of the very reasons you list above: it further distances our perception from reality.

    Now, that’s not to say that this sort of image manipulation can’t be educational. Many objects that Hubble observes have a shape that a casual observer would have difficulty translating into their three dimensional forms without facilitation. Of course, I feel that it would be best to present the original image and only use the stereoscopic manipulation for the sake of illustration, rather than presenting the image as “what Hubble saw”.

  • Walter

    I just went to see the movie a couple of weeks ago and it is very impressive. Are you sure about the circular polarization? Before the movie we were playing with the glasses and I got the impression it was linear, at least for the movie they have at the California Science Center.

  • http://sacrilicio.us Matunos

    You neglected to mention that the actual film is of the Hubble repair mission, not images from the Hubble itself (at least, not its main focus, maybe they’ll throw some of those in). They apparently flew out an IMAX 3D camera during the most recent mission, so that footage at least will be real 3D like we humans see it.

    Even if it was a movie of Hubble footage, (a) the images would probably mostly be false color, and (b) the images would be so distant, what good would human-style depth-perception provide? IMO, even if it’s 3D recreated by other information and converted into images our vision can understand, it’s still 3D. The objects out there have depth, you’re calculating the depth, and projection those calculations to our eyes, right?

  • http://janineweshould.com Justin

    Also, you might note that it is possible to create (possibly synthesized) stereographic images from “flat” images. The recent remake of Alice and Wonderland was marketed as a 3D movie, and was a beautifully done one, art-direction wise. However, the entire movie was shot on a “cyclops”, 2D IMAX camera. All of the stereographic enhancements were done in post-production, and mostly likely synthesized. It’s very possible that any 3D Shots of space-ly sights were done the same way — shot in 2D, made 3D in post-production.

  • Nonnormalizable

    Hubble (with astronomers, software, cosmological theory, etc.) sees 3D via redshift. As it happens, humans see 3D via two eyes and parallax. Just translating one way of detecting 3D into another: no contradiction.

  • Brian Siana

    As Matunos points out, the primary theme of the movie is the servicing mission, though I think they only shot ~10 minutes of IMAX footage on the shuttle. Having said that, by far the best parts of the movie were the “fly-throughs” of the Hubble images. The trip through the Orion Nebula was spectacular. Every single person around me in the theatre was in complete awe. I know it’s not scientifically accurate, but this movie beautifully motivates our research to the general public.

  • http://www.astro.multivax.de:8000/helbig/helbig.html Phillip Helbig

    “not images from the Hubble itself (at least, not its main focus, maybe they’ll throw some of those in)”

    Pun intended?

    “The recent remake of Alice and Wonderland was marketed as a 3D movie, and was a beautifully done one, art-direction wise. However, the entire movie was shot on a “cyclops”, 2D IMAX camera. All of the stereographic enhancements were done in post-production, and mostly likely synthesized.”

    First, it’s not really a remake so much as a sequel. Second, knowledgeable reviewers prefer the 2-D version, since the 3-D stuff is tacked on (obviously jumping on the Avatar bandwagon).

  • beanfeast

    From what I have read it seems that IMAX 3D uses linear polarisation. However non-IMAX 3D films seem to use RealD which utilizes circular polarisation. I have only seen IMAX 3D films, Avatar & AiW, and I can confirm that in both of these the stereoscopic effect is lost if you tilt your head.

  • dk

    Not all depth cues are binocular. Close an eye, does your perception flatten?

  • http://www.astro.multivax.de:8000/helbig/helbig.html Phillip Helbig

    “Not all depth cues are binocular. Close an eye, does your perception flatten?”

    True, but 3-D films use only the binocular clue. You can’t choose which depth to focus on, you can’t move your head to look around something etc.

    There was some discussion at Ted Bunn’s blog on 3-D technologies. Here is something from a reply by me there:

    n the latest Physik Journal (magazine of the German Physical Society), there was a two-page article on 3-D techniques. You mentioned three: colour, linear polarisation and the quarter-wave-plate model (your hypothesis was correct; that’s how it works). The last is definitely the best of these three, but shares this problem with linear polarisation: the reflected image has to be polarised, so the screen has to be mirror-like, not just a white screen.

    There are two other techniques. One projects frames at twice the normal rate, altenately for each eye, and the glasses contain infrared-controlled LCD shutters which alternate at the appropriate rate. Probably the best system, but the glasses are more expensive.

    Another one is quite interesting: for one eye, use three primary colours, and for the other eye, use three OTHER primary colours. The filter for each eye only lets through the primary colours intended for that eye. (For a given perceived colour, there are many ways of mixing it out of narrow-band “primary” colours”.

    See also more discussion in another thread there:

    http://blog.richmond.edu/physicsbunn/2010/01/20/how-3d-movies-really-work/

  • James

    I mean, come on. Also — in addition to that salient point — I think the comma in the headline itself makes your insight less needed. The second clause is descriptive of the way that the viewer will be seeing it, not the way Hubble sees it. The headline isn’t “Seeing what Hubble sees in 3D, in 3D.”

  • http://lighthouseinthesky.blogspot.com/ Anne

    Just to be picky, the Hubble space telescope can actually see in 3D using, effectively, stereo vision: parallax measurements take almost the same image six months apart, and for close enough objects, they see exactly the same parallax shifts the human eye uses to see depth. Effectively, since the scene does not change, the Hubble is acting as two eyes separated by two AU.

    That said, I don’t think there’s any single object that shows any appreciable depth in parallax; while it would let you see the differences in distance between (nearby) stars, it’s not going to let you see any depth in something like the Orion nebula. And in fact, I’m not sure that there are two stars that show appreciable parallax in the same Hubble field-of-view.

  • Trevor

    dk hits a very important point. binocular vision isn’t even the most dominant component to depth. it works, yes, and so 3d technology makes use of it, but your brain is much more sophisticated at making assumptions about how the world is constructed, such that walking around with one eye closed is not a problem at all, and you could still easily avoid a rock thrown at you with one eye closed. it’s easy to undermine this processing — just look at some escher prints! but in the real world it works much better than relying on binocular cues.

    nowadays it’s pretty straightforward to emulate the procedure the brain uses and figure out which things are in front of what other things in a 2d image, and bring Hubble images of the Orion Nebula to 3d life in a spectacular way (it really was amazing). Indeed 3d tv’s on the very near horizon will be able to do this in real time to 2d pictures.

    saying your brain simply differences the two eyes’ images to determine depth is totally oversimplified, to the point of being wrong, as dk’s simple experiment shows. clearly the article was written by a theorist! :-)

  • Trevor

    right on James: “Seeing What the Hubble Sees, in a Movie Theatre”

  • Low Math, Meekly Interacting

    Ditto to what Trevor said. Stereopsis does play a significant role in depth perception, but it’s just one of many means the brain uses to perceive depth. For distant object’s, especially if they are moving, it plays almost no role at all, and motion can provide plenty of parallax information to the one-eyed, too, at any distance. If Hubble takes a picture of Saturn, my brain “knows” its an oblate spheroid surrounded by concentric rings. Probably the billows and voids of a nebula provide some sense of what’s closer and further from the foreground. While it’s gimmicky and not entirely honest, I don’t think it need be terribly deceptive to extrapolate depth from the same sorts of cues our brain uses to “see” depth in 2D photographs, and synthesize anaglyphs accordingly. So long as there’s disclosure about the manipulations, and effort made to respect real spatial relationships (to the best of our ability to measure them, anyway) in the synthetic 3D images, seems to me like a valuable way to help people visualize celestial objects in their true glory.

  • ian

    I think the NY Times title is ok, the key point is the comma. Hubble sees in 2d, but we see in 3d. Therefore if we look at what Hubble sees, we see the 2d images using our 3d vision. If I’m looking at a 2d image on a piece of paper or tv screen, I’m still seeing it in 3d. It’s just that I’m seeing a 2d image embedded in 3d space.

    I’d like to see Hubble pictures in imax. They could use some lasers and stuff though.

  • Toiski

    Imax 3D uses LCD shutter glasses in some theaters (http://en.wikipedia.org/wiki/IMAX#IMAX_3D). The other method is linear polarization, as Walter and beanfeast said.

  • Kaleberg

    The first space stereogram dates from 1858. It consists of two images of the moon and takes advantage of the moon’s libration, the wobble in its point closest to the earth. Thanks to libration, we can see more than 50% of the “light side of the moon”. It also means we can make a stereogram and see the moon in 3D as if we were looking through a giant’s eyes. (I think the reference is to De La Rue.)

    We can definitely do 3D iMax astronomy. We just need to the eyes of giants. We could use the parallax from images six months apart, though a 3 1/4 light year distance yields only a second of an arc difference. More interesting would be to synthesize our own giants using data from Hipparchos or red and blue shifts to provide a 3D sense of nearby stars and other objects.

    When 3D fails, iMax can always exploit visual dominance. Look at the opening to the original Star Wars. That was almost 3D in its own right. Clever composition can give a great sense of distance, proportion and distance. Combine that with some actual 3D iMax footage of a shuttle repair mission, and I’ll bet we’ll have a hit on our hands.

  • http://www.astro.multivax.de:8000/helbig/helbig.html Phillip Helbig
  • http://www.astro.multivax.de:8000/helbig/helbig.html Phillip Helbig

    What’s the deal? Why can’t I post the URL to Ted’s blog?

    • http://blogs.discovermagazine.com/cosmicvariance/sean/ Sean

      For some reason, the link to Ted’s blog was alarming our spam filter. Maybe it has something against Bayesians?

  • http://www.astro.multivax.de:8000/helbig/helbig.html Phillip Helbig

    Would that be your prior assumption?

  • Hubbleuser

    ‘The way we figure out distance in Hubble images is by using color information (and, in particular, the spectra) to discern recession velocity (redshift), and thereby distance (using Hubble’s law).”

    Talk about one-eyed cosmologists! Hubble spends much of its time looking at things other than very distant galaxies. And the Hubble Law is hardly the appropriate way to figure out the distances in these cases.

  • spyder

    Hubble is a cyclops, and takes very long looks at the past.

  • Rohan Mehra

    I just got back from seeing Hubble 3D at the Science Museum in London.
    Their cinema uses circular polarized glasses, tilting your head maintains the illusion and putting them on upside down reverses the depth which is realllly weird.
    They claim the imagery is accurate and based on real measurements.
    Whatever the case, it’s stunning, I recommend for all, even people who are likely to spot if a star is 10nm too blue or 5pa too close to some other star…

  • Rick Shafer

    If you’re going to pick nits, Hubble wasn’t launched with much in the way of `film’. All digital detectors, thank you.

  • Brian Too

    Perhaps stretching the topic a little far here, but I just read an interesting article discussing the emerging consumer level technologies supporting 3D in computers.

    They made the interesting point that many (most?) computer games are already fully 3D internally and have to add logic to provide a point of view, then project this point of view on a (typically) 2D computer screen. The games have to subtract out the third dimension in order to support the display technology that’s available.

    As a result it was possible to achieve 3D viewing on many such games even when such capability was entirely unofficial and unsupported by the vendor!

    I didn’t really get how they reverse-engineered the games to do this, though.

  • G Driver

    One image I would like to see is Earth, with possibly the Moon, against a background of stars and galaxies.

NEW ON DISCOVER
OPEN
CITIZEN SCIENCE
ADVERTISEMENT

Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Cosmic Variance

Random samplings from a universe of ideas.

About Daniel Holz

ADVERTISEMENT

See More

ADVERTISEMENT
Collapse bottom bar
+