Image with intrinsic color relationships.	Same image roughly convolved with the response of the human eye.

What is a true color astronomical image? Is it what an astronaut would see if there was an eye piece on, say, the Hubble Space Telescope? Or is it one that captures the intrinsic colors emitted by the stars, nebulae, and gas clouds in galaxies? For the Hubble Heritage image of globular cluster, M15, we constructed an image which attempted to captured these intrinsic colors as closely as possible given Hubble Space Telescope (HST) data. The detail on the left above, highlighting Planetary Nebula K648, is an example of this approach.

Our goal was to maintain the relationships between the intrinsic energy output at a selected wavelengths. HST Wide Field and Planetary Camera 2 data consists of black and white exposures taken through filters. (Please visit "Is that what they really look like?" by Zolt Levay, for an in-depth explanation of image construction.) Each filter is centered on a certain wavelength and selects a range of wavelengths in the electromagnetic spectrum. For M15 we used black and white exposures taken through 5 different filters.

For all our images we assign a color to the image in each filter and combine the color images. If the color assignment increases with increasing central wavelength, then we call this assignment chromatic order. Usually we don't worry if the assigned color exactly matches the filter's central wavelength nor if the difference in brightnesses between filters is maintained. But in the case of M15 we did worry about these things.

First we calibrated the black and white data so that we knew how much energy, on average, came through each filter and was detected at each pixel. (For those who are curious, the units were erg/sec/cm*cm/angstrom.) Then we made sure that when we mathematically manipulated the brightness levels, to bring up the faint light, that we applied the identical manipulation to each filter. This way we could maintain the intrinsic relationships between the different parts of the electromagnetic spectrum.

Next we assigned a color to each filter's image. Two filters capture the light from ionized gas; ionized hydrogen appears magenta and ionized oxygen is green to the human eye. Three of the filters capture the light from the stars. One filter has a central wavelength that is blue, another has a central wavelength that is a green which is just a tad yellow. And the last filter, ah well, please forgive us... it has a central wavelength that is in the infrared! That's from part of the spectrum that the eye does not see. Since there were no data through a filter which matched the red color seen by the eye, this one would have to be our substitute.

Combining these color images results in the image on the left above. (For the Dying Star in Globular Cluster M15 image we de-emphasized the oxygen filter so that you could more easily find the pinkish planetary nebula. Here you can see that the pink (hydrogen) and green (oxygen) mix to form a yellowish color.) Clearly some stars are more blue than others, indicating that they have higher surface temperatures. But a surprise was that the planetary nebula appeared more pink than green. Most people looking through a large telescope report that, in general, these nebulae appear green. However independent measurements of this specific nebula demonstrate that it does emit more light due to ionized hydrogen (magenta) than ionized oxygen (green)! (Adams et al. 1984). Our image has captured the intrinsic relationship between these 2 colors emitted by the nebula.

So what would be the "true" color, if instead that was defined to be the color seen by the human eye? Roughly, it could be the image seen on the right above. At low light levels the human eye is most sensitive to green light, then blue, and lastly red. But this brightness relationship is not in color! For example, small faint blue stars simply appear white and brighter than similar red stars. The color sensitive part of the eye responds most to the color yellow and slightly more to red than green. Mixing the brightness relationship with this color response we would expect to see a ghostly green nebula, as reported.

Thus trusting our eyes we would never know that there was more emission from magenta, ionized hydrogen than green, ionized oxygen. And we can barely see a difference between hot (bluer) and red (cooler) stars.

So consider the Hubble Heritage image of M15, a representation of what you could see if astronomical objects were brighter and if your eyes were sensitive to a slightly larger range of the electromagnetic spectrum. And besides maintaining the intrinsic relationships between the colors emitted by the stars and the nebula in M15, don't you think this image has a striking glow?

Enjoy!
Jayanne English and Forrest Hamilton

For another rendition of Planetary Nebula K648, see Howard Bond's Homepage at Space Telescope.