Nearly every photo of space you see, especially from Hubble, is usually a composite, falsely colored, or otherwise altered in ways that bring out the important features of the photo. Note that the idea of "true color" for these sorts of astronomical photos is itself a bit ill-defined.
> Note that the idea of “true color” for these sorts of astronomical photos is itself a bit ill-defined.
Is it? I always figured “true colour” would be whatever you’d see with your naked eye from an approaching or orbiting spacecraft, or the nearest approximation that (say) a consumer digital camera would give you. Can we get anything like that from our current compositing techniques?
> Is it? I always figured “true colour” would be whatever you’d see with your naked eye from an approaching or orbiting spacecraft, or the nearest approximation that (say) a consumer digital camera would give you. Can we get anything like that from our current compositing techniques?
For many astronomical observations, the wavelengths of light observed are ones our eyes are not sensitive to. In this particular example of aurora, it's the ultraviolet. For others, it's infrared, millimeter/radio, or X-rays. For those, there's no way to way to make an image that our eye would see, because it wouldn't see anything. Some of the color images from those wavebands are described as "representative color" because the images are created to show certain features linked to the underlying physics or properties of the objects that are being shown.
For images that purely in the visible portion of the spectrum, in principle one could construct an image "as the eye would see it". However, as @dakr noted[0], color astronomical images are composites of several black and white images taken through different filters. to create an "as the eye would see it" image requires combining those black and white images in a way that considers the relative sensitivity of the different types of color-sensing cones in our eyes.
But, if you really want to get it right, it's more complicated than that. The shapes of the frequency-response of the cones probably differ from the transmission of the filters used in the astronomical observations (not to mention the quantum efficiency of the CCDs used to take the observations). So one would also need to correct for the difference in filterobject versus eye_coneobject (both of those as a function of wavelength). This is typically referred to as a "color correction" (meaning the "color" of the light emitted by the source), and requires some knowledge of the intrinsic spectrum of the source.
And there's even another layer after that. If emission (or absorption lines) from atoms lie within the observation filters, those can contribute a non-negligible amount to the signal in an observation with a given filter. Those wouldn't be properly corrected by a color term (for a broadband filter) because they only emit within a relatively narrow wavelength range. So that's another, second order, "color term" to consider.
Astronomical cameras don't capture color information in the sense that our eyes or consumer digital cameras do. They are "black and white" only detectors. To get color information, several images are taken with different filters in front of the detector. Afterwards, those images can be stacked and each layer assigned a color to create a composite image.
Making this a "true color" image that exactly mimics what your naked eye would see ranges from difficult to impossible (like in the case of UV data). It's also unimportant, because we mostly use color images as a visual aide (of course!). The "science" is in comparing and contrasting the individual "color" images.
This is exactly how CMOS/CCD image sensors (used in digital cameras) work [0]. Each "sub-pixel" is just a photocell which detects light intensity. A bayers filter on top of the sensor array filters light colors so different "sub-pixels" detect colors (light intensity of different parts of the visual spectrum) [1]
Enthusiasts sometimes remove the filter to get higher optical resolution or better contrast B&W images. You can even buy consumer cameras with this feature off the shelf [2]
NASA is doing the equivalent of instagram filters, making photos appear more appealing to the general public. NASA says (via. Space.com [3]):
"Creating color images out of the original black-and-white exposures is equal parts art and science," NASA said.
For example, Hubble photographed the Cat's Eye Nebula through three narrow wavelengths of red light that correspond to radiation from hydrogen atoms, oxygen atoms, and nitrogen ions (nitrogen atoms with one electron removed). In that case, they assigned red, blue and green colors to the filters and combined them to highlight the subtle differences. In real life, those wavelengths of light would be hard to distinguish for humans."
I'm not sure exactly what you mean by "ignore". The colors, though somewhat arbitrarily chosen, do convey physical information. It may not be a "what your eye might see" image, but it does tell the viewer about the relative emission from hydrogen, oxygen, and singly-ionized nitrogen (for the Cat's Eye Nebula). So, the coloring isn't something that could be ignored.
One way to describe those types of images is "representative color".
"Cat's Eye Nebula ... In real life, those wavelengths of light would be hard to distinguish for humans."
The context of the my comment was how the photo compare to the naked eye.
I used to take IR photos for fun (consumer camera with IR filter replaced with visible light filter), and used thermal and night vision technology at my job. They all relay useful information, but are not photo-realistic in anyway.
An IR photo of Earth for example, useful but not photorealistic.
> The context of the my comment was how the photo compare to the naked eye.
I agree that is true for your example of the Cat's Eye nebulae. I guess my response was to my (and presumably other's possible) mis-interpretation that "ignore the color" was aimed at color astronomy images in general.
In some sense, digital cameras do the same. They have a different channel for red, green and blue, that are composited to provide something similar to true color. (Even old film cameras do something similar.)
Sometimes you can get surprises. Most cameras can see some of the UV and infrared light.
And it isn't a very good one at that. The aurora overlay seems to be on a different rotational axis, as if the Jupiter image was taken while the planet was at a very different angle. It makes me wonder whether I can trust the height of these things, which in the image extend well above Jupiter's visible atmosphere. Is that reality, or a poorly-aligned overlay.
I am so happy HST's mission was extended for another few years [0]! I would have funded a Kickstarter to keep it going, to be honest. If there's one thing Hubble does well, it's that it capture the hearts and minds of muggles like myself. HST is an amazing piece of machinery.
