Thoughts on Color in Astrophotography

The Atik 16 is a monochrome camera. It records all of the visible spectrum and some beyond it, especially into the infrared. I like that about the camera, but only one familiar with the camera's range knows what light I am capturing, actually. I have been trying to decided what to do about color. I want to do some experimenting soon. In response to a great suggestion by Anjal Sharma that I start with simple color filters for visual observing (of which I have a few), I responded as follows:

I still haven't decided what to do with color. Don [Taylor, also a member of the Huntsville club,] uses an interesting palette with narrowband filters. I understand why folks would do that. Knowing where the hydrogen, oxygen, and sulphur are allows one to pinpoint where the radiation is that is exciting the gas. The detail possible in a narrow wavelength is also interesting. I can also understand colors for reflection nebulae, in part because dust is a different color than the gas. Where there is information in the color, I can learn something from it. H2 regions in M33, for instance, will be a different color than the young stars in M33's arms.

When it comes to regular old RGB, I'm not sure. Aesthetically, sometimes the colors look very nice, but so does an impressionist painting. Why not red, yellow, and blue? Why not orange, green, and purple? Is it because RGB roughly corresponds to Ha, Oiii, and some other things? So it's sort of broad narrowband? What about Hb and other less prominent lines? In fact, many cameras pick up in the infrared as well, and that records in luminance and perhaps in some red filters, but that isn't even visible light to our eyes. It has no color. Our eyes respond only weakly to Ha, too, so what does a bright red Ha coloration mean? One thing I like about the Atik 16 is its response in Ha and its strength in infrared. It's significantly more than Oiii, for instance, so I am getting a very different view than I would with filters. I am also getting a different view with a reflector than I would with a refractor, which would probably not focus the infrared. I'd have to cut it off.

Anyway, it seems to me that each camera, scope, and filter set are different enough that no one but the photographer really knows what we are looking at---unless there is some attempt to correlate everything, ... everything! But then we would have to justify why we are correlating it the way we are, and I'm not sure I've ever heard anyone justify that persuasively, particularly when our eyes cannot see the kinds of colors in the individual spectra of deep sky objects and our cameras record wavelengths beyond what our eyes can see.

Also, aesthetically, I like black and white about as much as color. Color strength is such a subjective thing, and everyone's monitor is different, anyway, so even the displays of the images will be different. Color seems like more or less a dice roll to me, in the final analysis---too random to matter much.

Actually, my bride would like my images better if they were in color. I may add color that she likes and be quite satisfied with that.

I'd be very interested in others' thoughts on this.

And in response, Don Taylor, whose excellent work can be seen at http://www.theatomiccafe.com/, responded as follows:

There are definitely (at least) two schools of thought on the topic of color.

There are, shall we say, the conservatives.

Those who would only ever use broadband.

They will weight their colors to precisely match an analog star.

Then there are the color liberals.

Those who know humans can never get their natural eyes to ever see such deep sky object in color.

So they reassign certain wavelengths to one of the human eye perceptible frequencies.

I suppose a third school might be the monochromatic enthusiasts.

But even they must break into either the camp that uses the spectrum from 400nm to 700nm or the camp that selects a narrow slice of the magnetic spectrum.

Perhaps even a slice above or below the spectrum visible to the human eye.

In my mind it comes down to one’s personal goals in astro-imaging.

I feel that no method is either right or wrong, better or worse than any other method.

Each has its challenges and rewards.

For me, as long as my images depict the form and natural processes happening in nature, I’m satisfied.

I will correct a fault in the equipment; remove the bloom from a star.

I will readjust the weight of a color channel to emphasize an area of interest.

That’s the extent of the artistic license I’ll take.

I won’t “paint” in features I wish I had been able to capture.

But that’s my self imposed limit.

It doesn’t make my image any better or worse than an image which has undergone more aggressive processing.

Astro-imaging can certainly enter into the scientific arena if that is the direction one wants to go.

However, for me, astro-imaging is an art form, albeit a highly technical art form.

There are vast quantities of information on the many disciplines of astronomy, and I enjoy studying much of it.

Astro-imaging is simply my way to actually be able to go out and touch the universe that I would otherwise feel detached from.

-Don

PS: The technical reason for using RGB is because the color receptors in our eyes (presuming one has “normal” color vision) are tuned to accept red, green and blue light frequencies. As you know any (human eye perceivable) color can be synthesized by combining these 3 colors in varying intensities. (This is the same reason all of the pixels in our monitors are either red, green or blue.) I would love to be able to use one of the pallets you mentioned like orange, green, purple, but my software only allows for RGB or CMYK. At least if I expect to get a “color” image out of it. I have tried using the CMYK pallet with mixed results. I’ve also used the HFC pallet. But Hubble popularized the RGB=Sii, Ha, Oiii pallet and people are used to seeing it. When I first started using it, I got a lot of negative comments about my stars all having a pink glow around them. Funny thing is, I’ve never heard anyone criticize a Hubble image for that, and all narrowband Hubble images have pink halo stars.

Anjal Sharma also commented:

I agree with yours and Don's comments. Here's my take on color; although we do not see color visually in the faint fuzzies, that does not mean that there is no color in them. The issue of color is a very subjective one, but there is some consensus amongst users as to what information color adds to a monochrome image - namely what type of stars comprise that cluster or galaxy, or what type of molecules comprise a nebula. So, I myself cannot be satisfied with just monochrome images and want color in them to see a more interesting view of these objects. That being said, I would like to point out that in the end what we're doing is an art form, not a science. So, the colors depicted in my images are purely an artistic expression of what I think the color of these objects should be. I take a lot of liberties then in the way I depict color in my images. There are certain things I will not do such as add stars or remove stars or take the paintbrush tool and add in a patch of nebulosity if it does not exist or use the clone tool to remove an existing patch of nebulosity. What I will do however is emphasize certain aspects of the object by using layer masks, star masks or selective curving of the histogram in those areas. In my mind this is no different than when a photographer uses a different f-ratio in the camera lens or a selective physical lens filter for example to de-emphasize the background while drawing attention to the subject. So the point I'm trying to make is this - almost all aspects of image processing are quite subjective in my opinion and as the imager all you're trying to do is present YOUR view of the object to us, the audience.

I find much to agree with in both Don's and Anjal's comments. Any others?