Once an item has been defined, colour differences become the first parameter in the evaluation of quality, and also the main source of disputes. Metamerism is a special case of colour difference. By definition, the colours appear to be the same, but vary under certain lighting conditions. As always, metamerism is a source of frustration for colourists working in a traditional way. Today’s fashion trends have introduced the need to combine different materials, which emphasises this effect. The definition of reference lighting conditions is an initial palliative, which however does not necessarily solve the problem. As we shall see, the trichromatic nature of human vision represents a potential metameric situation. In recent decades, new cost-effective tools have become available that allow colourists to obtain the information necessary to resolve the problem, although a complete resolution is only possible by standardising the language and parameters of colour throughout the entire supply chain.
1. Definition of metamerism
A metameric pair can be defined as two colours with different spectral compositions that however generate the same colour stimuli under certain conditions such as lighting, size and angle of viewing or the chromatic sensitivity of observers. In fact, we talk about a metameric pair because this effect is evident when comparing at least two colour samples.
The colour depends on three main factors:
> An observer with his discrimination capacity and sensitivity
> A light source with its characteristic spectral composition.
> An object with its light absorption/diffusion/reflection.
2. The observer and the mechanism of vision
Physiologically, the retina is responsible for light perception (Fig. 1). This is a thin membrane that covers the ocular globe. Cones and rods (Fig. 2) are photodetectors within the retina. There is only one type of rod with maximal sensitivity at around 510 nm and three types of cones called ρ, γ, and β (corresponding to r, g and b, namely, red, green and blue in German), with maximal sensitivity at wavelengths 580 nm, 540 nm and 440 nm respectively. In figure 3, we can see the spectral sensitivity of the cones. We can distinguish two areas on the retina: one central area called the macula, containing the central fovea, rich in cones, and the middle and extreme area with prevailing rods. At the centre of the fovea, in the point where the optic nerve travels to the eye, there are no cones or rods. This is known as the blind spot, a point where there is no perception. The fovea is responsible for day or photopic vision, while the mean and peripheral fovea is responsible for mesopic and scotopic vision (intermediate lighting levels and night vision). Given that cones and rods may be degraded in the presence of high levels of lighting, the eye has a pupil, which is similar to a diaphragm in old cameras. The pupil regulates the right amount of light, and according to its dilation, allows different areas of the retina to be illuminated. Each cone and rod receives light, converting it into an electrical impulse that is transmitted through the optic nerve to the brain.
no images were found
The proportions of the various types of cones and rods will define a chromatic sensitivity which is unique for each individual (it may even be different for each eye). The defect or lack of one or different types of cones or rods will define the various defects in colour perception. To study spectral sensitivity of a certain number of observers, the standard colorimetric observer was established in 1931, which represents the response of an average observer. This is one of the data matrices adopted to emulate human vision in a spectrophotometric system.
3. Light sources/correlated colour temperature
The light source will bring a certain spectral component, which corresponds to a colour of light defined in terms of correlated colour temperature, based on Planck’s law of total radiation (also known as blackbody). Planck’s experience is based on the temperature increase of a hollow sphere with a small observation hole. According to the energy state, the sphere or radiator will emit a radiation with a characteristic colour, which varies from an emission in the infrared (greater than absolute zero) to visible red (800 K ca.) passing through white and blue. This effect can be observed when an iron bar is heated. It first turns red and then progressively shifts towards yellow as it is heated further, and becomes brighter. Each light source has its own distinctive spectral composition, as we can see in figures 4 and 5. The CIE (Commission Internationale dell’Ecclairage) has standardised different light sources, such as illuminants A, B, C, D50, D55, D65 and D75 etc. The importance of the light source is mainly related to the colour, but also to its yield. As can be seen in figure Fig. 4, when compared to a “D” illuminant, the “A” illuminant has a yellowish/orange dominance without components in the blue/green area. Clearly, this property will be seen as an increase in the yellow/orange dominance of the samples it lights, and a relative lack of blue/green shades. Figures 4 and 5 show the colour variability of the radiant source according to the energy content. Similarly, the blue and green shades lose strength at dusk, while the reds and yellows increase their relative intensity. Let’s evaluate what happens to the colour temperature during the day: early in the morning, at dawn, the colour temperature is around 900-1000 K (very red) increasing to 6,500 K at noon on a cloudy day (total diffusion of light) or above 11,000 K during a clear sky day, and returning to the same 900-1000 K with the sunset.
no images were found
Figures 6 and 7 shows the variations in the shades of the light sources with reference to the colour temperature.
Another factor to consider is the surrounding. So as not to influence the light composition, the surrounding should be medium grey, with 50 % reflectance and matt. In actual fact, the places where we evaluate the colours usually have walls and roofs in different colours, which contribute to a non-standard reflected component, bringing our most usual source of reference to the most extreme conditions of variability.
no images were found
Conditions necessary for the existence of metameric colours
Due to the trichromatic nature of human vision, the sensitivity of the observer as well as the lighting level will define potential metameric situations. This is because colours with different spectral compositions may produce similar colour perception stimuli or colour sensations. The example in figure 8 shows the spectral composition of two metameric colours.
no images were found
Samples 2 and 3 in figure 9 respectively correspond to the samples and tests in figure 8. Observed under a D65 illuminant, they appear to be the same, while in figure 10 under an A illuminant (Tungsten incandescent) they appear to be very different. Since there is an overlapping between the spectral curves of both colours, we are sure of seeing an extreme metameric pair.
no images were found
no images were found
4. Types of metamerism
We have seen that the eye does not respond to each visible wavelength of light, but to the integration of three trichromatic stimuli. In this way, infinite spectral compositions will be able to generate the same colour sensation. This effect will define the different metameric situations:
a) Observer metamerism: two observers will always have a different perception of colour stimuli, depending on the composition and distribution of the eyes’ cones.
b) Visual field metamerism: since the cones are concentrated in the central fovea, a variation of the visual field or viewing angle may generate a different colour stimulus.
c) Light source metamerism: a variation of the light source will evidence a different perception in the colour stimulus.
The problem of the observer’s metameric effect may be partially solved by working traditionally with colourists who have a very acute visual perception. Unfortunately, statistics show a high percentage of observers with visual defects, especially men (about 8%), while this percentage drops to 0.4% for women. This is because colour blindness has a genetic origin that is dominant on males and recessive on females.
Many tests are available to evaluate the discrimination capability of colourists. One of the best known is the Ishihara test (Figures 11 and 12). It is a confusion test where, depending on the perceived number, it is possible to deduce if an observer has normal vision or a visual defect. Figure 11 shows one of Ishihara’s confusion charts, and figure 12 shows a simulation of the perception of a “colour-blind” person.
Another method is the “Munsell – Farnsworth” chromatic discrimination test (fig 13). It involves a series of colours with minute differences in tone between them, to be organised in a progression of tones. Inserting the positions of the colours in a dedicated software, we can evaluate the discrimination capability of observers and the axes of confusion, generating a rating that will indicate the discrimination capability.
To evaluate the metamerism for variation of the visual field, simply adopt a method of visual colour evaluation under a normal light source (0°), with the observer at a viewing angle of 45°, then vary the angles.
In the case of light source metamerism, we can initially limit the problem by evaluating samples with a suitable observer in an appropriate environment and avoiding the use of solar light. For this purpose, we use cabins with standard lighting, which must meet the CIE recommendations: dull and grey neutral background and a standard D65 illuminant (6500 K), and then compare the matched colour with respect to the standard under a second illuminant (for example, TL 84) keeping as a reference the grey scale for evaluating the colour differences. If the colour difference is not greater than the customer’s limits (or our limits) it is fairly certain that the matched colours will not be very different if compared under other conditions or by other observers. Unfortunately, human physiological limitations to trichromy prevent the perception to single spectral components of the colours being matched. This factor may cause endless incorrect tests especially with greys, browns and beiges, where colour perception between the colours is shared by all three types of cones.
5. The spectrophotometric solution – calculation of the metamerism index
By means of spectrophotometry combined with computerised colour matching systems, it is possible to choose the right dyestuff or pigment combinations that will give the smallest colour difference. This is possible because these systems can evaluate the spectral trend at each wavelength and not as human trichromatic integration. In addition, through algorithms, these systems calculate the behaviour of two measured samples under different lighting. The values can be expressed in terms of colour difference, and as a metameric index, quoting the lighting used.
These systems allow matching the spectral curve of the reference colour by minimising the likelihood of metameric effects.
In recent years, different formulas have been developed to calculate the approximate of tolerances. The algorithm most suitable for this purpose is the CMC 2:1, which allows defining the tolerances of an ellipsoidal approximate that approach the differences perceived by the standard observer.
6. Standardisation of colour along the supply chain
To avoid interpretation differences and misunderstandings, it is fundamental to establish with customers and suppliers the conditions for colour evaluation. It would be ideal to apply international standards, normally this choice must be related to the environment in which the final articles will be exposed or used. In the case of car upholstery, the combination of textiles, plastics and metals should not alter in daylight, at night and under artificial lights. With regard to the finished articles, it is important to consider the lighting inside showrooms and shopping centres and stores.
Designers with the support of specialists must define the colours, with their ranges of tolerances and feasibility of realisation. The clearer and more unalterable the references, the greater the certainty of a successful matching. For example, we can define a Pantone reference with a small photo-colorimeter, measuring the reference colours, or defining them via a colour guide. It is essential to establish a control protocol of the colour, including the standard D65 illuminant and another usual reference (TL 84, D 11, A) to evaluate the metameric tolerances.
Light cabins of the latest generation have the possibility of storing lighting profiles, such as the type of illuminant and its intensity. Just the fact of defining the reference colours in this way, allows us to transmit the desired colours and lighting profiles via email, telephone or fax, without having to send physical references, which can deteriorate or perish with handling and time. The spectrophotometer in the hands of colourists should serve as an extension of his eyes to find the best solutions with a low metameric index.
At Ars Tinctoria, we have dedicated part of our laboratory in Santa Croce sull’Arno to the study of colour, with a support program aimed at defining and standardising colour and light, evaluating the perception of colourists, and organising tailor-made courses and workshops.