The Kruithof Curve – Color Temperature VS Illuminance



Have you ever heard of the Kruithof Curve?

Back in the early 1940’s when fluorescent sources were beginning to affect the way we thought about light and color rendering, a scientist that worked for Philips named Arie Andries Kruithof performed some informal tests on how the human eye relates the amount of light in a given time of day to the color temperature of the light source.  Typically, human beings like higher color temperature light sources during the daytime hours, and lower color temperature sources once the sun goes down.  People in warmer climates tend to favor cooler color temperature sources, and people in colder climates like warmer light.  It seems pretty intuitive, yes?

Is this an official guaranteed works-for-every-human-on-earth standard?  Of course not.  Everyone is different.  Eastern societies have different preferences than Western societies.  But – and this is a general but – there is a correlation between the amount of light from a light source (lux) and the color temperature of the light source (degrees Kelvin) that seems to be fairly common among us all in most situations.  This is the research that culminated in A. A. Kruithof’s color temperature VS illuminance curve, as seen above.  Kruithof was working on visually pleasing light sources, and was interested in how adjusting the amount of light altered the amount of illumination needed to maintain a pleasing sense to the human eye.

The rods and cones in the human eye work together, and once the amount of illumination reaches a certain low or high point, the rods (intensity sensors) lead the visual information to the brain.  At night, when dusk conditions occur, you might notice that most of the colors in your view tend to be monochromatic, usually blue – this has to do with the low level of illumination, and a phenomenon referred to as the Purkinje Effect.  The Purkinje Effect tries to explain why our brain switches to scotopic vision at dusk when illumination levels are very low, and color rendering is poor – as the brightness of the day decreases, the vibrancy of reds goes away a lot faster than the vibrancy of blues in our vision.

We might have some almost built-in tendencies towards color temperature and light levels – perhaps somehow tied to the cycles of the sun and our circadian cycles.  We might have a tendency to associate warm colors with fire light at night, and we might associate higher color temperatures with the mid-day illumination levels from the sun.  Who really knows.  Kruithof gave it a try, and the curve is what he determined.

The two sources in the graph are the color temperature of Western/Northern Europe at mid-day (D65), and a 2700 Kelvin MR-16 tungsten-halogen source, for reference.

Thanks, ArchLighting and SoLux!

Previous articlePimp My Hubble – New Deep Space Images
Next articleHow High Output Lamps are Made


  1. My PC CRT monitor reads 66 LUX on  full white brightness just before it blooms.  Kruitof implies that for best color perception at this 66 LUX level of illumination, the color temperature should be 2700K. 
    And yet, computer monitors for color work are universally set to 5300K or 6500K by all color calibration software.  Kruitof would also imply that for calibrating your printer against your monitor, the print should also be illuminated with a 2700K source at 66 LUX for your eyes to function best for evaluating color. 
    A 2700K tungsten halogen bulb has a spectral response curve very close to an ideal black body radiator.  Bulbs of 5300K or 6500K have very bumpy (dichromatic filter quartz halogen) or spiky (florescent) spectral response and would logically be unsuitable for accurate color work.
    Is it any wonder that the bulk of color calibration equipment results in such poor color results?

  2. The Kruithof curve isn’t about the ability to perceive color, it’s just about color preference.  It shows how people prefer warmer color temperatures at lower illuminances. 

    Computer monitors don’t have color characteristics that are anywhere close to black body (full spectrum) emission either at 2700K or 6500K or any other color temperature.  This is  because they use phosphors to generate color, and those phosphors only emit at certain narrow frequency bands (not across all frequency bands like an incandescent lights do, or like the sun does).  Fundamentally that’s why color calibration of monitors isn’t very reliable.

Comments are closed.