# Why are high white balance temperatures redder when warmer objects are bluer?

I know it's dumb to ask this type of questions. But, I still don't get this.
I've read the "Black body" story and know that when it's heated, it gives off colors; 'warm' colors (red→orange→yellow..) that have low color temperatures, at low temperatures, (I mean, when it is cooler than afterwards) right? And, the black body gives off 'cold' colors (white→blue..) that have high color temperatures at high temperatures (when it is hotter than before), right? Correct?
So there're two, yes, very dumb, questions and the latter one is dumber..

1. Why are that 'warm' colors said to have low temperatures? And 'cold' colors vice versa?
2. So the redder colors have lower c/t and bluer colors have higher c/t, then why does the camera's color temperature - adjusting slider has the bluer colors on the left, and the redders on the right?

Please, forgive me for my ignorance...

The Why Different is just a difference between physics and art. The "temperature" simply is the actual color of black bodies when heated high (Like steel in a furnace). They do first glow red, and much higher temp turns white or blue. Just does, in physics. And the color does represent temperature.

But the art world has reversed notions, because mere humans seem to perceive red and orange colors (fire) as warm, and blue colors (ice) as cold. The red hot metal is an unusual situation, rarely actually encountered day to day.

Not sure which slider you see, but generally, the color from incandescent light is orange (maybe 3000K), or the open shade sky light is blue (maybe 8000K). The direct Sun is more 5000K-5500K, which we call white. That's physics. Then the temperature slider in photography is correction in the opposite direction, seeking white balance. More orange warms blue, more blue cools orange (which is the human perception of Art). The slider is often about correction of color, instead of about measurement of color.

What is interesting to me about White Balance is that our WB tools (Adobe) match the Lab color space axis. The WB Tint slider is just the Lab -a to +a axis, and the WB Temperature slider is just the Lab -b to +b axis. Center of both is the neutral color, no color cast. The Lab color L axis is Lightness, or brightness, which in Lab, is isolated from color.

• So, when my camera's temperature-adjustment slider is slid to the left, the display (image) turns blue, when slid to the right turns it red. So, the slider is functioning according to "the human perception of Art" as you said? (It adds red to the existing for now blue?; as the slider is being slid to the right (to the 10000k (to the "human perception of Art"'s)) side?) – user152435 Dec 25 '16 at 15:17
• It is Temperature (physics) vs Color (art). In art, orange is a warm Color. But in physics, orange is relatively low Temperature numerically, and blue is high, and we are seeing a Temperature scale, degrees K. The slider moving Left turning blue is a Correction adding blue (the opposite) to correct too much orange (on the left). The resulting color temperature number is higher temp, but a cooler color. In physics, that result is higher Temperature. But artists speak of warm or cool "Color", which is the opposite notion (think fire and ice). It is about Temperature (physics) vs Color (art). – WayneF Dec 25 '16 at 15:48
• i would say it is even more fundamental than art, for thousands of years the hottest thing we were able to touch was fire and the coldest thing was ice or snow - this must have some impact our human languages and thinking – szulat Dec 25 '16 at 15:54
• In the real world the coldest and the hottest things are white. The orange things are in between. The hottest flame is white. The coolest flames are orange. Metals are white when they are heated to just below their vaporization point. So is snow and ice is white. It's what is in the middle that has more color. – Michael C Dec 25 '16 at 16:55
• The hottest stars are blue/white dwarfs. The coolest stars are red giants. Yellow stars like our sun are in between. – Michael C Dec 25 '16 at 16:58

When you slide the color temperature on your camera (or image editing software) up to 10000K you don't actually change the very cool blue/white light in the real world to make it more orange. You change the way your camera makes the very cool blue/white light that is at 10000K look more orange in the picture. If the light is very blue/white then you need to amplify the inverse color of blue light, which happens to be orange, to make it look like normal light centered around 5200K in the picture. But you're not changing anything with regard to the actual light in the real world, it's still very cool blue white at 10000K.

Likewise, when you slide the color temperature on your camera down to 2500K you don't actually change the light to make it more blue. You change the way your camera makes the very warm orange light that is at 2500K look bluer in the picture. If the light is very orange then you need to amplify the inverse color of orange light, which happens to be blue, to make it look like normal light centered around 5200K in the picture. But you're not changing anything with regard to the actual light, it's still very orange in the real world. Your picture is just making it look bluer in the picture.

Another way to look at it is to think of the color temperature setting on your camera or in your editing program as a filter. If the light is tinted towards orange you need to use a blue filter to make the light look more normal. If the light is very blue then you would need to use an orange filter to remove the blue tint. Since 2500K is very orange, one needs to use a blue filter to compensate for it. Since 10000K is very blue, one needs to use an orange filter to compensate for it. If we used an orange filter under orange light it would make the picture even that much more orange!

The change in color that you see when you move the color temperature slider is due to the change in the color of the filter you are applying using the color temperature setting. It is not due to a change in the color of the light that entered the camera when the image was captured.

• Your answer untangles my question well. So, as I understood from your answer, 1) blue light's temperature is at 10000K.. 2)And one 'enters' the existing (blue) light's c/t by sliding the slider to where the existing c/t is mentioned (to 10000K side, in this case) and, adds more inverse color (orange) ... .. Am I correct? – user152435 Dec 26 '16 at 6:49
• And, the 'inverse color' story eh.. captivated (is it the word?) me long before the color temperature thing bemused me.... .. I first noticed the after-image of a single colored shape is the inverse of its color.. .. After that I kept a paper painted with several color squares that I used to 'white balance' (it's not the word for my case, more like 'change the color') my camera by pointing and clicking at the inverse color of what color I wanted to change to. – user152435 Dec 26 '16 at 7:01

As you know, metal heated in a fire soon begins to glow. First the metal takes on a dull red glow, then cherry red. As the metal’s temperature rises, the color changes to white hot, then blue-white. It’s these observed color changes with heating that inspired the color temperature system.

Also, you know that most of the world uses the Celsius system. This establishes water freezing at zero (0) and water boiling at 100. The unit degree translates to “step”. Early experimenters discovered that a hydrogen thermometer is super accurate. This is a hollow tube, filled with hydrogen with a float atop the column. The float uniformly moves up and down with temperature changes. Other substances like mercury and alcohol lack this uniformity. Further, as the surroundings cool, the float drops to near the bottom of the tube. It was calculated that the lowest possible temperature is absolute zero, and if achieved, the float would hit bottom. Thus the Absolute temperature scale was born. This scale was favored by many, as all temperatures are positive, no confusing +20 with -20. This temperature scale was renamed the kelvin scale after the scientist Lord Kelvin’s paper of 1848, on absolute zero.

Now many disciplines use the color of glowing hot substances to gage temperature. To name a few: blacksmiths, iron workers, steel making, ceramics, glass blowing etc. Experiments proved that the glowing color and its related temperature was approximately the same for all materials. Key here is the lighting industry that initially was carbon arc and glowing tungsten, adopted the kelvin scale to relate the color output of lamps.

Some selected kelvin temperatures:

Candle flame 1850K

75 watt household tungsten electric bulb 2820K

200 watt general service electric blub 2980K

500 watt photo-flood electric bulb 3200K

500 watt movie photo-flood electric bulb 3400K

Flash bulb 3800K – 4200K

Caron arc lamp 5000K

Photographic Daylight 5500K

Sunlight standard US Bureau of Standards Noon 5500K

Blue sky 12000K -18000K various times of day

Color films were manufactured to operate under specialized conditions.

Color balance Daylight

Color balance Tungsten movie lights

Color balance Tungsten photo flood

Color films for scientific work – other kelvin temperatures

Note: custom is to write the word kelvin scale as lower case k and omit the degree sign °.

Digital camera makers logically adapted the photo film industry using their color balance notations.

• "Sunlight 1200K - 1800K?" – Michael C Dec 25 '16 at 21:15
• Sorry about that, omitted a zero! Should be 12000K thru 18000K. I made the edit. Many thanks to Michael Clark. – Alan Marcus Dec 25 '16 at 21:55

The artistic/subjective model of warm versus cool colors predates the scientific model based on Planck's Law by about 100 years. Perhaps red->warm and blue->cool reflects a psychometric property of the human vision system. And perhaps not, for the same reasons that photography is or isn't an art or a science.

On a cold wintry night, if I trust Goethe's color sense I wind up warmed in the orange light of a hearth fire though scientific color suggests laying upon a starlit snowbank.

• I one were as close to one of those stars as one would be to the hearth one would be quite warm... – Michael C Dec 26 '16 at 17:00
• @MichaelClark Goethe's wisdom strikes again. – user50888 Dec 26 '16 at 18:50

Why are warm colors said to have lower temperatures?

This is a physics issue. I will try to explain it with as little math as possible so it may be vague so bear with me. Recall the color wheel and that mixing all colors equally would give you white. Now look at this spectrum of radiation emitted by a body, straight from Wikipedia. I will skip for now what a black body means. I will also urge you to ignore the black curve marked "Classical Theory" as it is not valid.

The x-axis is the temperature and the y-axis is the Spectral radiance or in simple words intensity of light at that wavelength/frequency. The "visible light" corresponds to a wavelength range of 400 nanometers (0.4 μm) to 700 nanometers (0.7 μm).

As the temperature decreases, the peak of the black-body radiation curve moves to lower intensities and longer wavelengths. At round 5000K you have a nice mixture of all colors so the mix appears more white. But at a lower temperature, red color is more intense so a lower temperature corresponds to a warmer color. A higher temperature would peak more towards the left and hence cooler. Many modern detectors use this to determine the temperature of an object including night vision.

• Ohh! I don't get it, clear me up first,.. Does the x axis (the horizontal axis) represent the temperature/wavelength (µm)? And the y axis represents the spectral radiance, does it? And what is $\mu$ m? And what is the 'black-body radiation curve'? – user152435 Dec 26 '16 at 6:27
• @user152435 $\mu$ is LaTeX markup for "μ", which isn't enabled on Photo.SE. I edited wander95's answer to make the units clear. – scottbb Dec 26 '16 at 6:37
• @user152435 Clarification: the x-axis is not temperature, it is the wavelength of emitted electromagnetic radiation. It is not an independent variable like you would normally see on cartesian plots. The "independent" variable on this plot is the color temperature lines. To read this chart, pick a color temp, say the green "4000 K" line. This chart says that for a 4000 K black body radiator, the peak spectral radiance (intensity of light) is a bit over 4 kW / (sr·m²·nm) (the y-axis value), and that peak occurs at a little less than .7 μm (x-axis value). – scottbb Dec 26 '16 at 6:44
• @user152435 and the "black-body radiation curve" is the the curve following the color temp you picked (in this case, the green line). The entire green line is the black body curve for a 4000 K radiator. It shows the intensity of radiation for all wavelengths of radiation. – scottbb Dec 26 '16 at 6:48
• @scottbb So in simple words, what is meant by "for a 4000 K black body radiator, the peak spectral radiance (intensity of light) is a bit over 4 kW" (related to that 'black-body gives off colors' thing?) – user152435 Dec 26 '16 at 7:12