What is “true color”? Is it what our eyes see, or is it something else entirely? In the most rigorous academic sense, it is related to how a display renders colors on a screen, but it is often used in the qualitative description of images. What constitutes “truth” in color? We may go through life assuming that other people perceive colors the same way we do, but of course if you stop and think about it, that is obviously not true. Approximately 1 in 12 men and 1 in 200 women worldwide are affected by color vision deficiency or color blindness (https://www.colourblindawareness.org/). Clearly, their perception of color differs from the average person’s. However, even among people able to perceive the full spectrum of colors, how can we be sure that we all perceive colors in exactly the same way?
Every few years this question seems to bubble up to the top of public consciousness in the form of internet memes about the color of some item in a picture. Whether it’s a dress or a shoe, people seem to get in very heated arguments about the actual color of these items. Much has already been written about this phenomenon, so I’ll leave that out of this discussion. Suffice it to say that there is a lot of behind-the-scenes interpretation of color handled by our brains, which leads some people to see things one way while others see it entirely differently. What’s more, if we stare at it long enough, we can convince our brains to swap the colors! Is there any truth to color?
My good friend and avid astrophotographer Vicent Peris recently raised similar questions in regards to imagery coming from the Perseverance Mars rover. Those images are all tuned to give a color representation similar to what the human eye would experience on the surface of Mars…what would normally be called true color. However, we know the Martian atmosphere is very red, due to iron oxide dust that makes up a large part of the Martian soil. This tends to give images from the Martian surface a very strong red cast. If we can divorce ourselves of the notion that the response of the human eye is what constitutes true color, we open ourselves up to more relativistic interpretations of color.
As Vicent rightly points out, there is much to learn about Martian geology, some of which can be inferred from the colors of different rocks, which are all extremely muted in the imagery that we see due to the strong red cast of the Martian atmosphere. If we instead use the light of the whole image as a white reference, we can reduce or remove the red glare and allow the natural colors of the rocks to show through. Incidentally, we see a similar effect right here on Earth with any pictures taken under water, where the colors skew towards blue as light diffuses through various depths of water. This effect can be mitigated in much the same way as Vicent’s approach to the Mars images. In both cases, this may not be what our eyes would see were we standing on the surface of Mars or swimming in the ocean, but it is informative nonetheless. This same white balance adjustment process was applied to the image seen at the top of this post, taken recently by the Perseverance Mars Rover, and Vicent’s image can be seen below.
Hubble images are processed with this philosophy of color, where the constituent filters that make up an image are balanced such that the full combination of all colors in an image sum to white. This technique, while it may not be matched precisely to the color response of the human eye, reveals more details and extracts the maximum amount of information from the image. This approach extends to multi-wavelength astronomical imagery as well, where wavelengths of light well beyond what the human eye can perceive are assigned colors and blended together to give a more complete picture of the object in question. Consider this image of Centaurus A below, which combines “Visible” light (roughly the light we perceive with our eyes) with X-ray, Infrared, and Radio light, which obviously cannot be seen with the human eye, yet reveal new information about this galaxy’s bi-polar jets.
One of Hubble’s most iconic images, the Pillars of Creation in the Eagle Nebula (a frequent feature of this blog), is a great example of this treatment of color as relative. The image is composed from three narrow-band filters (see this post for more information on filters), which isolate the light of ionized oxygen, hydrogen, and sulfur gases. These ionized gases emit photons in the light blue (cyan), red, and “more” red portions of the electromagnetic spectrum. If the composite color image were assigned those exact colors, the image becomes dominated by a red cast as two of the contributing images are assigned basically the same color. If we instead shift the color space such that the shortest wavelength (oxygen) is assigned blue, the hydrogen green, and the sulfur red, we arrive at a much more informative and dynamic image. The combination of all three channels, as seen in the cores of the bright stars, sums to white, while all variations of color are represented within due to the additive color-mixing process combining individual red, green, and blue pixels.
Certainly, we cannot and need not entirely throw out the concept of true color. For cameras and displays, there needs to be some quantitative metric by which color representation can be measured, and true color as it relates to the average person’s response to color is the best way to do so. However, I would advocate that we be less rigid about how things should look and more open to identifying and moving past the limitations of human vision in favor of expanding our visual palette. We know our brains do a lot of work to give context to color, and they don’t always get it right. We can use our perception and interpretation of color as another tool in the toolbox of understanding our colorful universe! And for the record, the dress in that first picture is white and gold, even though the dress in real life is blue and black.