Dina Vees is an Assistant Professor of Graphic Communication at California Polytechnic State University. She has taught a variety of classes from front end web development to prepress and is also the advisor for University Graphic Systems, a student run print service provider located within in the department.

She taught Printing and Publishing at Waukesha County Technical College for 19 years prior to joining Cal Poly. Her industry experience is in prepress. She has always had a desire to embrace technological advances in our ever-changing profession. This drew her curiosity to research new low-cost devices used for measuring color. Dina is also the treasurer for Graphic Communication Educators Association (GCEA).

A Comparison of App-based Color Measurement Devices to Spectrophotometers

Dina Vees, MFA and Malcolm Keif, Ph.D., California Polytechnic State University

Maintaining color consistency across various print methods and substrates is a daily challenge for the printing professional.Accurately reproducing brand colors demanded by customers is of the utmost importance. It is common practice to manage color accuracy by measuring color attributes throughout the print workflow using colorimeters or spectrophotometers.

This empirical research study looks at three new low-cost app-based measurement devices to determine their intermodel agreement with two professional spectrophotometers. An experimental study was conducted to look at the five measurement devices to determine standard deviation, precision or repeatability of measurements,consistency of L*a*b* readings, and Delta-E 2000 values to understand the performance of low-cost devices.

The instruments studied in this research were the Nix Mini, Nix Pro 2, Variable Spectro 1™, X-Rite 530 and the Techkon SpectroDens. The X-Rite and Techkon are spectrophotometers with built-in processing and are considered professional instruments,while the Nix and Variable devices are app-based and often considered consumer devices, which sell for considerably less than the other professional models. We grouped these three and called them“low-cost” instruments for the purpose of this study.

Seven Pantone® colors were selected and twenty repeated measurements were taken for each device. All devices were calibrated and measurements were taken with little to no movement between twenty readings. In other words, all devices measured the approximate same spot, though aperture sizes varied among instruments. CIELAB measurements were captured using D50 illuminant, 2° Observer and 45°/0° geometry for all instruments. The SpectroDens, which is ISO 13655:2009 compatible,was set to measurement mode M0to try to maintain consistency with the other devices. A white backing was used during all measurements.

Three hypotheses formed the basis of the research:

  • H1Low-cost instruments have greater measurement variation across repeated measurements than professional spectrophotometers.
  • H2Low-cost instruments produce statistically different L*a*b* readings than professional spectrophotometers.
  • H3Low-cost instruments have greater calculated Delta-E 2000 values relative to professional spectrophotometers.

While all measurements were taken from a new Pantone® guide, we used Pantone’s digital L*a*b* values from Pantone’s website as the reference for calculating Delta-E 2000(DE00). Since the production of Pantone® guides are subject to printing variation too, we did not consider a particular DE00to be “correct”,but rather looked at trends and other evidence of variation in our study.

The low-cost instruments do not measure to the same level of significant figures as the professional devices,resulting in precision differences. This made direct comparisons somewhat challenging during analysis,particularly for H1 Low-cost instruments produce statistically different L*a*b* readings than professional spectrophotometers. Comparing standard deviations of the various measurements proved futile since the Nix Mini reported all readings to whole numbers and therefore had very little deviation of readings. We were not able to confirm H1 and instead rejected that hypothesis. Analysis of variance (ANOVA) gave us a better understanding of the performance of these devices and allowed us to analyze the other two hypotheses effectively despite the difference in precision among the devices.

Hypothesis H2, Low-cost instruments produce statistically different L*a*b* readings than professional spectrophotometers was confirmed using analysis of variance and post-hoc analysis using Tukey’s HSD at a 95% confidence level.

We also were able to confirm H3 that Low-cost instruments have greater calculated DE00 values relative to professional spectrophotometers by using Tukey’s HSD at a 95% confidence level and comparing the mean range of Da*00, Db*00, DL, DC, DH readings across all 7 colors.

The three low-cost, app-based devices are small, inexpensive and easy to use,off-loading all color computations to a mobile app, but do not provide the precision or consistency necessary to compare to professional spectrophotometers used in industry to meet critical color reproduction standards.

Range of Delta value calculations incorporating six color attributes: Da*00, Db*00, DL, DC, DH.

 

Cluster plots of various color difference dimensions by instrument. Note the Y-axis changes for each plot.