# White balance color space [duplicate]

How were the colors for white balance chosen? Why is it specifically temperature going from blue to yellow and tint going from violet to green?

• Does this answer your question? What is color temperature and how does it affect my photography? May 5, 2020 at 5:13
• The answers to the duplicate are, at best, incomplete when talking about the totality of WB, which encompasses much more than color temperature. The magenta ←→ green axis is also a component of WB. Oct 20, 2020 at 18:04

Commonly, when white balancing the temperature and tint parameters are mapped to a position on Planckian Locus, i.e. Blackbody temperatures, and the Delta uv on the corresponding iso-temperature line normal to the Planckian Locus, respectively.

In the following CIE 1960 UCS Chromaticity Diagram, the Planckian Locus is the curved line in the middle, and the iso-temperature lines are the lines crossing it and perpendicular to it.

If you imagine the temperature and tint parameters defining coordinates in that diagram, the resulting colours should make sense: temperature makes colours varying from blue to orange, tint from green to magenta although this varies with temperature.

The colors along the color temperature axis were "chosen" because they are the colors black bodies radiate as they increase in temperature. This includes everything from heated metals to the surfaces of stars, including our Sun. Almost all of the strong light sources found in nature emit light somewhere along or very close to the color temperature axis. We designate the colors along this line based on the temperature to which a black body must be heated in order to glow at that color.

We use the temperature scale created by Sir William Thomson, 1st Baron Kelvin, OM, GCVO, PC, PRS, FRSE. Addressed by his royal title as Lord Kelvin, he was a mathematician and scientist who create a temperature scale that places the "null" mark at absolute zero, the theoretical temperature point where all molecular motion will cease, and uses units the same size as the Celsius scale. 0 K is equal to -273.15°C. 0°C is equal to 273.15 K.

For most of human history, even primitive artificial light sources like torches and oil lamps emitted light along the Kelvin temperature scale. This is because most of the fuel sources don't actually glow very brightly as they burn, but the gasses and vapors that are emitted from the fuel source as it is heated burn and glow very brightly. Early electric light sources used metals heated to glowing temperatures to produce light. Since the metals used are black body radiators, they also emit light along the color temperature scale. These are what we call incandescent light sources.

In the modern world, though, we deal with a lot of light sources that are not natural and do not fall on the Kelvin scale. The axis that is more or less orthogonal to the color temperature axis is the Magenta ←→ Green axis. This is often called a "tint" or "hue" adjustment. This is represented in the illustration above by the lighter grey hashes along the color temperature axis. Many artificial light sources, particularly those designed primarily to use low amounts of energy, are quite a distance away from the colors emitted by black body radiators on the Blue ←→ Amber color temperature axis.

So in addition to adjusting color temperature to compensate for our light source, we must also compensate along the tint axis. Many cameras call this white balance correction.

For instance, in addition to having a color temperature of about 3700 K, traditional fluorescent bulbs also emit a green tint along the green←→magenta axis and need correction in the magenta direction. On the other hand, many of the popular LED stage lights found in small clubs are also at about 3700 K but have a decidedly magenta tint that requires compensation in the green direction along the green←→magenta axis. Both types of light are the same basic color temperature but look very different without compensation on the green←→magenta axis that is approximately orthogonal to the blue←→amber color temperature axis.

Beyond doing color correction in two dimensions instead of just one, there's also the issue that many artificial light sources don't emit the full spectrum of visible light.

Most natural light sources do. The sun, as seen from the Earth's surface on a clear day, may be centered on about 5500 K, but there is at least a little bit of the entire visible spectrum in sunlight. The Sun emits even more electromagnetic radiation than what we can see and measure from the Earth's surface. The Earth's atmosphere reflects and absorbs some of the energy radiated by our sun, and lets what we call visible light pass more easily. (Of course the reason we call it "visible light" is because we have evolved to be visually sensitive to the wavelengths of electromagnetic radiation that the atmosphere we evolved in allows to pass most easily!)

When using only artificial light sources that do not give a broad spectrum of light, there are certain colors we can't reproduce. This is because there is no light the proper wavelength(s) to reflect off our subject for that color. Custom White Balances come in very handy for such lighting. In such cases, we might also have to use an HSL (Hue-Saturation-Luminance) tool in post processing to remove a color cast.

• "there is at least a little bit of the entire visible spectrum in sunlight. That's generally because the black body radiator is not the exact same temperature over the entire surface" No, the full visible spectrum is present even in constant temperature black body radiation. See the spectrum: en.wikipedia.org/wiki/Black_body . Otherwise awesomely clear answer! May 6, 2020 at 12:15
• @JereKupari Thanks for the heads up. I've edited the answer. It should be noted, however, than no individual star is a perfect black body radiator, just as no metal or other material is. As the linked Wiki article points out several times, we model stars as if they are black body radiators. It also notes that in the case of supergiants and main sequence stars, it's rather remarkable that we can get away with treating them as such. "It is perhaps surprising that they fit a black body curve as well as they do, considering that stars have greatly different temperatures at different depths." May 7, 2020 at 10:15

Craftsman working metals, glass, ceramics, and the like, heat their materials until they glow with luminosity. These artisans have for centuries linked temperature with the glowing color. Heated objects first glow faint red at 930° F = 500°C = 770K. As the temperature upraises the glowing color shifts blood red, cherry red, salmon, orange, yellow, white, and finally blue-white. Science uses the Kalvin temperature scale to make this relationship to honor Lord Kalvin a British physicists, Nobel Prize Laureate.

Sunlight is specified as 6400K. Photo films fashioned for use under sunlight conditions and lamps that replicate sunlight, are attuned to this color temperature. Films and lamps to be used in studios were set for 3200K and 3400K.

The short answer is: Photography attempts to make faithful images. To accomplish, films and now digital imaging, fabricate their devices so that they replicate the response of the human eye/brain mishmash.

How is that blue (-) to yellow (+) is the a axis, and green (-) to red (+) is the b axis of the Lab color space (see https://en.wikipedia.org/wiki/CIELAB_color_space ). The third axis is "lightness", black to white, or intensity. It can make accurate white balance corrections.

Can't say exactly why it was chosen for White Balance Temperature and Tint, except I think there is no better choice. It approximates human vision, and is device independent (RGB is device dependent).

It's called the color balance:

This allows an image to be adjusted in the shadow, midtone, or highlight regions for the red-cyan, green-magenta, or blue-yellow balance.