L*a*b* – What does this mean in terms of colour?

Modern instruments have made colour measurement easy and highly accurate. However interpreting the data can be fairly daunting especially when the definition provided by sites like Wikipedia cause more confusion than clarity.

Users and Managers probably prefer simpler explanation as they report to or have hired nerds to already understand the detail, but usually don’t have time to sit down and have the talk.

The aim of this blog is to provide more clarity on the results that we receive when measuring colour in the most commonly used C.I.E colour space, L*a*b*.

How to interpret L*a*b*?

The image below shows the results that were received when measuring a Green Tile on the CM-5 Spectrophotometer in Reflectance mode.

• L*a*b* is a specific colour space developed in 1976 by the CIE to mathematically represent colour. There are a number of different colour spaces available, but for the purpose of this blog, we focus on L*a*b* which is the most commonly utilised colour space.
• L* = is a number from 0 (Black) to 100 (White). Essentially the L* value represents how light or dark an object is and has representation of colour. The closer the result is to 0 the darker the sample is, the closer L* is to 100 the lighter the sample is.
• a* represents the relationship between RED and GREEN and comes in two parts: a*+ (positive) represents RED while a*- (negative) represents GREEN. There are technically no maximums as this number can exceed 100 in some cases. The higher the value (+ or -) the more intense that colour is represented in a sample. We can say that this value denoted intensity of that colour.
  From the results above we can see that the a* value = -18,25. As this number is a negative value, we would conclude that this colour has a certain amount of GREEN within the sample.
• b* represents the relationship between YELLOW and BLUE and also comes in two parts: b*+ (positive) is YELLOW while b*- (negative) is BLUE. Like with a*, the value (+ or -) denotes the intensity of that colour within a sample.
  We see from the results above that the b* value = +24,14 which would indicate that this object contains some YELLOW.

Without looking at the colour, we can deduce that its a green hue with almost equal amounts of green and yellow. The L value (70,96) suggests the depth of shade is between medium and pale, but this perception depends on the application. As it is above 50, we can say that it is more of a lighter than a darker colour.

Can this data be presented graphically?

As this data creates a 3-axis plot, we can depict the colour information and present it onto a 3 dimensional plot. You can download a copy of this colour space directly on the Konica Minolta Sensing website.

 

Now this is where most users become confused because they believe that the results they receive are the property or actual colour of the object where in fact, it is only the perception of the observation of that object.

These values are only correct under the SET OF CONDITIONS used at the time of measurement and can therefore only be compared to other measurements taken under the SAME EXACT conditions. A colour space is designed to encompass ALL the conditions to allow an “Apples to apples” comparison.

What are the conditions that might affect colorimetric results?
There are quite a number of criteria that need to be identical to create the same context, as colour is not a property of the object measured, but only of the object under the current context measured.

The conditions are:

1. Does the instrument, software and firmware conform to C.I.E norms and standards?
2. Measurement Type – Reflectance, Transmittance or Emitted 
3. Instrument Geometry (d/8 or 45/0) – This is a fixed condition and cannot be changed by users. Instruments have a certain type of geometry that can never be changed. It is important to compare results from instruments with the same geometry.
4. What illuminant is used for the measurement? Illuminant settings such as D65 will affect results.
5. What is the observer angle that is used? Either 2° or 10°
6. Measurement aperture size? This can vary from device to device but typically there are 3 aperture settings, 30mm, 8mm and 3mm.
7. What colour space has the instrument been set to? If you are trying to compare L*a*b* with Lab, you will find a difference in the results.
8. What are the specular component settings? SCI or SCE
9. Wavelength range (360 – 740nm) and Wavelength pitch (10nm) settings. Again these are inherent in a device and cannot be changed.
10. Does the sample contain any Optical Brightening Agents (OBAs)?
11. Levels of accuracy, repeatability and inter-instrument agreement. Comparing a highly accurate device against one that is less accurate can cause numbers to be different.
12. Sample presentation and preparation – The samples should always be prepared and presented to the instrument in the same way.
13. What are the external environmental conditions such as temperature and humidity?
14. Is the device being calibrated daily and has the device been certificated for performance.

Some of the conditions above are fixed while others can be changed on the instrument itself. It is important to understand and note these settings for effective colour control practices.

How important are all of these criteria?
If you want to compare a master (Standard) colour to a sample (Production Batch) colour – Important
If you want to compare a master (Standard) colour to a sample (Production Batch) colour across many different sites – Critical

In most cases where a customer says that they suddenly have ‘different’ numbers, it usually comes down to a change in settings and conditions.

How can we compare Targets (Masters or Standards) to a Sample?

 

This particular screen has stored Target Data in L*a*b and in the identical setup is compared to a recently measured SAMPLE.

A new symbol Delta (Δ) is observed in the reading. The SAMPLE data has been subtracted from the TARGET data and the DIFFERENCE (Delta or Δ) returned.

Reading differences in the 2nd decimal are MEANINGLESS and can be ignored as no two readings are ever correct to such small values.

In perspective, depending on the object and the importance of colour to the object, a trained eye can see Δ 1.0 and a consumer can even be blind to Δ  3.0 if colour is less important than say value.
Automotive and Clothing may work to very close tolerances, while foodstuffs may be sometimes wider, although international food company’s closely monitor colour as part of their Branding.

Conclusion:

Here are the key points to remember:

1. Colour is NOT the property of the object, but is rather the perception of the observation of that object under defined conditions.
2. If you change a condition, the colour WILL change.
3. Colour data can be collected using an instrument which reports this on a C.I.E colour space such as L*a*b*
4. L*a*b* and ΔL*a*b* can only be compared between two samples under identical conditions.
5. Conditions and settings are crucial to any colour measurement project. Be aware of your instrument settings to avoid errors and miscommunications.