Spitzer Visualization Scientist

"If only you could see the things I've seen with your eyes." - Roy Batty from "Blade Runner"

Posts by Robert Hurt

08.02.10

Phony Colors? Or Just Misunderstood?

In a recent review of a space imagery exhibit at the National Air and Space Museum, the New York Times reviewer goes on an odd tangent half-way down the page, seemingly taking exception to the way astronomical images from Hubble (and presumably other missions like Spitzer) are produced. Referring to the colors as "phony" and stating that "Many are colorized far more radically than any 1930s movie," he makes it sounds like something dodgy may be going on...

Trust me, it's not.

It seems that astronomy is still being haunted by misconceptions about how color is managed when dealing with telescopes that see light differently from the human eye. It doesn't help that so often we read misleading terms like "false color." I personally dislike that term a lot because it implies something is being misrepresented.

Everything we ever learn about color in school, or even art school, is really tied to the long and semi-aribitrary evolution of the human eye. The only fundamentally special thing about the "primary" colors of light (red, green, and blue), is that they correspond to the parts of the light spectrum our human eyes happen to be able to distinguish. Lots of animals on the planet see different slices of the light spectrum than we do, even into the infrared and ultraviolet. That falls beyond the human-centric definition of "visible" light!

Entire books can (and have) been written about how we see colors across the electromagnetic spectrum, and how we can shift information from other parts into the familiar red, green, and blue so that we can experience them with our own eyes. A good analogy might be translating text from a language you can't read into one you can. The translation isn't "phony" because it isn't the original text (which would be meaningless to you), it has merely remapped the ideas into something you can wrap your brain around.

In the same way, the colors we see in astronomical images like the ones here on the Spitzer website are very much real colors, representing fantastic complexities of light across the spectrum. They've just been translated for your eyes, remapped into red, green, and blue. 

That's very different from the task of "colorizing" a black and white movie. An artist does that by arbitrarily picking colors for everything in the scene and painting them on. Those colors may have nothing to do with what colors were actually there... that process truly could be described as "phony."

The Spitzer website would be kind of dull if we just showed you infrared light the way your eye sees it. There would be a lot of black, blank pictures, since it's not visible. Well, visible to us anyway. Somewhere on some other planet, there might be some infrared-viewing aliens debating over the colors in images that their astronomers observed in our visible light spectrum. Let's hope they don't conclude that the colors we can see are phony!

By the way, the exhibit in the review really does sound spectacular. It may be worth checking out if you are in the Washington D.C. area!

Comments

August 3, 2010 at 05:05 AM

memphismike said —

these pictures are..for lack of a better word..badass.Have some of these as my screensaver.

August 3, 2010 at 06:58 AM

Kelle said —

Do you have any ideas for a term to replace "false color"? Maybe "color mapped"?

August 3, 2010 at 10:59 AM

Robert Hurt said —

Kelle, I think the best going term now is "representative color." Lots of syllables but it's a lot more descriptive of what's going on. And instead of "true color" for actual red-green-blue images I tend to say "natural color." It's still specific to human nature, but again I think it helps avoid misunderstandings.

August 3, 2010 at 01:21 PM

Superstringcheese said —

I think 'human-visible' color sums it up nicely. 'Wide-spectrum' color is everything else.

August 3, 2010 at 02:34 PM

Zolt Levay said —

Nicely put, Robert.

August 6, 2010 at 11:04 AM

Christopher said —

This is a very timely article for me, thanks. Just got in a discussion with a visitor to our observatory on how she doesn't like that colors she sees in astronomy photos are "fake". I tried - not sure if I was successful - to explain that they aren't fake, they're just frequently chosen to highlight certain structures, etc, but the structures they show are very real. I like the language translation analogy....I'll have to keep that in my back pocket.

August 12, 2010 at 05:00 PM

Alan Stockton said —

I think that the confusing thing is that we have used the term "false color" in two different ways. One is the way described above, where we map other wavelengths to the (human) visible spectrum. The other is where we map some other property (usually intensity) to colors, so, for example, we can show more dynamic range than we can with a monochrome image (e.g., inner structure in a galaxy without losing faint outer structure). We need at least two different terms for these uses.

August 17, 2010 at 10:58 AM

Mario Lessard said —

Where can I find more information about this color translation? Particularly Hubble effect information on speed - distance of stars? In other words, can we tell how far a star is by its rep color? Sorry if this is posted elsewhere on this site, but it is so vast and awesome that I haven't found it yet. I discovered this site only yesterday and I'm thrilled!

August 24, 2010 at 05:36 PM

Robert Hurt said —

Alan, you are absolutely right. When we apply a gradient of color to a grayscale image then those colors are arguably "false". I personally tend to stick with the term pseudocolor for cases like that. It still means "false color" but without the associated hint of illegitimacy. You will find a number of such images on the Spitzer site when we present a single wavelength of data.

October 16, 2010 at 10:20 AM

Joshua C said —

I have two questions on this. 1) Is there a universal or standardized method used when translating the non-visible colors? For instance, will let's say, 32nm ultraviolet light always be translated to the same color between telescopes or is it really up to the maker of the telescope and/or person(s) processing the collected data? And 2) What happens with visible color? Does it just get left as is and essentially bleeds in? Is it removed first? Etc.

Add Your Comment

Name
Email
URL optional
Comment