Technidyne Header Image

Sunday, June 30, 2013

L*a*b* Production Tolerances for Colored Papers


Color Tolerances
With the use of color systems, the magnitude and direction of color difference between a sample and standard can easily be determined and understood. The Delta E* (∆E*) values are overall color difference values, which take into account lightness/darkness differences as well as chromatic differences. The intended object of these systems is for a color difference (∆E*) of 1.0 unit to be exactly the same visual color difference anywhere in color space. In practice, this objective is seldom realized; therefore, the establishment of tolerances based solely on ∆E* is not recommended. Individual tolerances on ∆L*, ∆a* and ∆b* should be employed for rigorous color control.  Likewise, the same tolerance on each of these parameters may be a simple approach, but is does not take into account the non-linearity of the color space.  Ideally, a variety of samples would be produced that vary from the customers perfect shade (standard) and the customer would be asked to accept or reject the different shades. This process would help establish tolerances based on the visual acceptability of the customer.   An excellent visual presentation of the L*, a*, b* color space is offered in the publication, “Prismatic II: A Visual Display of Measured Color Difference”. Two of those images follow to help give a better visual understanding of the variation in tolerances that may exist for more or less saturated colors.



Note: The following illustration uses Hunter L,a,b and ∆E, however, the concepts are still the same for CIE L*,a*, b* and ∆E*.



Figure 1 - Bone White
Bone White: The top of Figure 1 shows a standard in the center and 6 samples surrounding it. Each of the 6 samples differs from the standard by ∆E = 1 unit. However, the differences are only in one direction of color space. For example, the sample above the standard differs from the standard by ∆L = +1.0 unit (lighter than the standard). The sample below the standard differs from the standard by ∆L = -1.0 unit (darker than the standard).  Likewise, the other samples differ from the standard by ∆a = +1.0 unit (redder than the standard), ∆a = -1.0 unit (greener than the standard), ∆b = +1.0 unit (yellower than the standard), and ∆b = -1.0 unit (bluer than the standard).



The bottom of Figure 1 shows a standard in the center and 6 samples surrounding it. Each of the 6 samples differs from the standard by ∆E = 5 units. However, the differences are only in one direction of color space. For example, the sample above the standard differs from the standard by ∆L = +5.0 units (lighter than the standard). The sample below the standard differs from the standard by ∆L = -5.0 units (darker than the standard).  Likewise, the other samples differ from the standard by ∆a = +5.0 units (redder than the standard), ∆a = -5.0 units (greener than the standard), ∆b = +5.0 units (yellower than the standard), and ∆b = -5.0 units (bluer than the standard).



Obviously, the difference of 5.0 units is much more noticeable than the difference of 1.0 unit.



Figure 2 - Banner Red
Banner Red: Like the Bone White samples, the top of Figure 2 shows a standard in the center and 6 samples surrounding it; each of the 6 samples differs from the standard by ∆E = 1 unit.  The illustration at the right-bottom of the page shows a standard in the center and 6 samples surrounding it; each of the 6 samples differs from the standard by ∆E = 5 units.




Color Tolerances - Summary
The thing to take away from this is that the obvious difference that exists at 5 units of color difference on the white sample is not as obvious with 5 units of color difference on the red samples.  Also, the white samples with a 1.0 unit difference from the standard may or may not be acceptable, whereas, the red samples with 1.0 units of difference from the standard all appear to be acceptable.  This helps exhibit the non-linearity of the color space compared to visual observation. Typically, more saturated shades (further from a* = 0 and b* = 0) can have wider tolerances and still be visually acceptable.  Based on feedback from a variety of paper producers in North American tolerances can vary.

Near Neutral, Pastel Colors

Saturated Colors
L* ± 0.50

L* ± 0.50
a* ± 0.50

a* ± 0.80
b* ± 1.20

b* ± 1.80

Note: These are aggregate tolerances based on a variety of paper manufacturers.  These tolerances may or may not be acceptable depending on your application. The customer’s visual observation is always the best way to set these tolerances.


This is a common issue for companies that make a wide variety of colors. If you have additional questions email me Todd Popson.