Prism Spectrum for Color Calibration

Tom Axford wrote:

SmoothOperator wrote:

I'm wondering, why use these expensive color charts for color calibration, when you could get the real thing from a prism.

There are infinitely many more colours than those in the spectrum. Most real colours are mixtures (e.g. browns) and do not occur in the spectrum which consists only of "pure" colours.

The are other colors yes but not many more. The spectrum consists of a near infinite variety of subtly changing colors. If you start investigating atomic spectroscopy you'll soon see just how many colours there really are. Humans can just about distinguish colors with a wavelength difference of 2nm giving around 200 colors in the spectrum. The same appearance of color can be achieved by mixing the correct proportions of primary colors.

The trouble with using a prism for your color calibration is that you don't have it calibrated at all you just get a band of varying color, and with a camera will often only see  relatively few distinct changes. Color charts have carefully chosen shades which are also separated to make them easier to tell apart.

FWIW I have used diffraction gratings to investigate the spectral distribution of lighting with my full spectrum camera (where the actual wavelengths/colors the camera responds to  stretches from 350nm to 1100nm rather than just the typical visual range around 400nm to 700nm. To get any useful data I would have to be very repeatable in positioning & make use of a range of optical filters. Fortunately the UV/Vis/NIR spectrometer at work allows me to measure absorbtion/transmission far more accurately them my photography can use (to better than 1nm wavelength & <0.1% transmission) but it only works using it's built in light sources. (The atomic spectrometer has at least a thousand fold better wavelength control but isn't practical for scanning wide ranges and can't hold useful items like lenses instead needing it's samples vapourised)