Although light sources having the same (correlated) colour temperature will also have the same colour appearance, this does not necessarily mean that coloured surfaces will look the same under them. As explained in the C/E system area, surface colour is due to selective reflection. In other words, those spectral wavelengths contained in the incident light that are reflected, will determine the colour impression we obtain from the surface.
This is all quite straightforward as long as the light source is a thermal radiator, and thus displays a continuous spectrum with all wavelengths represented. But a selective radiator, such as a discharge lamp, emits light in a selected number of spectral lines or bands only, the other wavelengths being absent.
The line spectrum of a mercury lamp (left) while the continuous spectrum of sunlight (right) both give an impression of ”white’ light.
How does this affect colour appearance and colour rendering?
That the colour appearance obtained from such a light source can nevertheless be white is explained by the theory of additive colour mixing. Any spectral colour together with its complementary colour will produce white light. As the complementary colour itself is generally also present in the spectrum – or can be obtained by mixing two other spectral colours – it is possible to obtain white light by the combination of only two or three single wavelengths. And although the white light thus obtained may be of a colour appearance comparable with that of a thermal radiator and therefore can be assigned a correlated colour temperature – surface colours illuminated by it will often be difficult to distinguish, as most of the colour shades they are composed of, are absent in the light falling upon them.
This effect is not restricted to white light (compare for example), the monochromatic yellow light from a low-pressure sodium lamp with that from an incandescent lamp tilted with a yellow filter. Although the colour appearance of the two light sources is the same, colours can be distinguished fairly well under the light of the incandescent lamp, whereas under the sodium light this is absolutely impossible.
Yellow light from an incandescent source, fitted with a yellow filter (left) ,allows fairly good colour perception, but under the monochromatic yellow light from a low-pressure sodium lamp (right), it is impossible to distinguish between colours.
The number, arrangement and relative intensity of the spectrallines or bands present in the visible part of the spectrum of a selective radiator, together determine how far a random selection of surface colours can be faithfully reproduced under this light This is called the colour rendering capability of the light source.
Standard colours for assessing the colour rendering index Ra
In 1965 the CIE developed a method for the quantitative assessment of the colour rendering capability on the basis of eight test colours. First the correlated colour temperature of the light source under test is assessed. Then, for each test colour, the colour appearance under the source is calculated as a percentage of that of a black-body radiator of the same colour temperature. The average result for the eight samples is called the colour rendering index (Ra). This is a number that may vary between no colour rendering *) for monochromatic light sources, such as low-pressure sodium lamps – and one hundred, for true black-body radiators.
The eight original CIE test colours used for assessing the colour rendering index (Ra) of a light source.
The colour rendering index (Ra for Rendition absolute) is given below for a selection of much-used lamp types:
For values below 25, the colour rendering index has no practical meaning. As a matter of fact because of the calculation method used – it can even take negative values!