M1: “The all time desired mode”
The graphic arts industry uses standardized viewing conditions in order to minimize issues when communicating colour. The relevant standard is ISO 3664, which specifies CIE illuminant D50. Since 2009 the UV-content of D50 has to be met within closer tolerances than before. In order to make sure that optical brighteners “glow” to a similar extent when illuminated during a colour measurement as they do in a D50 viewing environment, ISO 13655 introduces the measurement mode M1. Compliance to M1 can be achieved in two ways.
Method 1: Illuminant Match
M1 can be achieved by using a light source that fulfils the requirements of ISO 3664:2009. This simply means, that if you build a normlight into the spectrometer, it complies to M1 (but remember the prerequisite of geometry). This sounds simple but can not be achieved in practice.
The obvious choice of using the same light source as used in most viewing cabinets can’t be realized as these are mostly fluorescent lamps which cannot be built into a spectrophotometer. In addition they do not perfectly match CIE illuminant D50 (approximation within defined tolerances).
Another possibility for achieving D50 is to use a combination of different LEDs, which produce a D50 spectrum. In practice a problem occurs when trying to mimic the UV content of D50 as current LEDs are not capable of perfectly reproducing the UV content of D50.
The last presented technical solution for achieving D50 as physical illumination is to use filtered light sources to mimic the spectral power distribution of D50. The advantage is that with this technique a close match to D50 can be achieved. This should also provide correct measurements for samples that show fluorescence active in the visible wavelength area (few inks and toners do show this behaviour to some extent). The disadvantage is that the light source might not be stable in terms of its UV content and therefore reliability over time has to be questioned.
Furthermore we have to ask ourselves if perfect D50 is really the best solution to be used in a measurement device. Normally we do not have perfect D50 as a viewing condition but only a simulation within tolerance. So the theoretical benefits can hardly be transferred to practical usage. How this problem is overcome will be explained later in this document.
Method 2: UV Calculation
The second possibility to achieve conformance to measurement mode M1 is related to the nature of optical brighteners. Optical brighteners absorb UV energy and emit blue visible light. To measure the effect of an optical brightener it is perfectly sufficient to assure a correlation between brightener excitation during the measurement and in the desired viewing environment. This is described by means of the ratio between UV-content and visible content in ISO 13655:2009.
In other words: Make sure that during the measurement the brightener glows as bluish as in your desired viewing environment.
This can be achieved in different ways. In the available literature some methods are described. Two will be discussed in the following.
As discussed, an optical brightener absorbs UV energy and emits this energy as blue light. Should we like to measure the amount of emission for a certain reference illuminant, we need to assure that the light source in the measurement device has enough energy in the wavelength area where the optical brightener is active.
If you were able to conduct two measurements, one using only UV energy, to give pure fluorescence, and the other without UV energy to provide pure reflectance, it was possible to calculate the resulting total radiance factor (often called reflectance factor although it is the combination of reflection and fluorescence).
The problem is that the method relies on the existence of a UV-only light source. The UV-LEDs that are available today have a varying spectral power distribution and also emit visible light. Thus not only fluorescence but also reflectance (caused by the visible light emitted by the UV-LED) are measured and introduce errors into the underlying model. Real-life instruments using this method would suffer from a varying measurement error.