The Measurement and Assessment of Colour Differences for Industrial Use V—The Accuracy of Visual Assessments in Relation to Industrial Colour Passing

The quantitative data available for expressing the errors associated with visual assessments, measurements and colour-difference equations have been used to calculate, for given sets of circumstances, the pass and rejection rates at various stages from the dyer to the retailer to the final customer. The main errors arise in the visual assessments and the colour-difference equations and, for the average of the available equations, these two errors are roughly comparable in magnitude. They do, however, vary with size of colour difference in different ways such that for small differences (0- 1AN40 units) the equations agree better with the true colour differences than a single visual assessment, whereas the reverse is true for large differences (above 2 AN40 units). Detailed results depend on the distribution of samples over the range of colour differences considered, together with the sizes of colour-tolerance limits set. Two initial distributions for the dyer have been assumed, one with same number of samples at all colour differences and the other simulating the distribution used in the Davidson and Friede study. For each, the results of use of a ‘low tolerance’ (0.5 AN40 unit) and a ‘high tolerance’ (1.5 AN40 units) by the dyer have been calculated. At low tolerance, the average equation provides much greater satisfaction than one assessor and improved equations will not give much more satisfaction. At high tolerance, the average equation could give greater or less satisfaction depending on the precision of the particular assessor and the exact distribution of samples to be considered, but there are considerable benefits to be gained by improvement in equation performance. Some of the likely interactions between the dyer and his customers are illustrated, showing that difficulties could arise unless there is some degree of co-operation. This is particularly true if unilateral action for equation use is taken by the retailer working to his customary tolerance limit. It would appear that the introduction of the use of an equation should first be made by the dyer, but if made by the retailer then consultation with the dyer is essential. The use of an equation by a retailer provides little benefit except in reducing dyer-retailer disputes.     

Publications Sponsored by the Colour Measurement Committee—VIII

Methods of describing the extent of agreement between the ΔE values calculated from a colour-difference equation and the corresponding visual estimates of colour difference are discussed. Lack of agreement will be due to a combination of errors in the visual assessments, in the measurements, and in the colour-difference equation itself. If there were no error in the instrumental measurements, the equation error for a particular colour difference would be the difference between the calculated ΔE value and the mean visual assessment of a very large number of observers (ΔVtrue). Experimental data in the form of acceptances (%) can be converted to ΔV values directly proportional to the observed colour differences. The overall equation error for n colour differences can be calculated from Eqn 1. The Davidson and Friede data are considered to be the most satisfactory of those presently available for testing the suitability of equations for industrial colour-tolerance work and have been used to assess the accuracy of several well-known equations. After allowing for errors in the visual assessment and in the instrumental measurements, σ(log ΔE) for the 1964 CIE equation was 0.22. Similar values (0.16-0.23) were found for other equations. A lower a-value was found for a very simple empirical equation essentially based on the x, y chromaticity diagram rather than on any transformation of it. The usual transformations tend to be based on data covering the whole chromaticity gamut, whereas real surface colours cover only a fraction of the possible area. This fact, together with the knowledge that most equations are based on data corresponding to small fields of view or large colour differences, could account for the relative failure of the standard equations.     

Performance of lightness difference formulae

There are large variations between different previously published lightness difference experimental data sets. Two hundred and eight pairs of matt and glossy paint samples exhibiting mainly lightness differences were accumulated. Each pair was assessed about twenty times by a panel of fourteen observers using the grey scale method. The results were used to derive a new lightness difference formula (CII), and to a large extent, a possible new CIE lightness difference formula (CMC99). Both formulae were found to be more accurate than the typical deviation of an individual assessment from the mean of a panel of 20 observers, and outperformed the existing formulae using the present data set. The new CMC99 lightness difference formula is integrated into the new CIE colour difference equation CIEDE2000. The results also showed that special attention should be paid to measuring very dark samples. This is caused by poor instrument repeatability and inter-instrument agreement in this colour region.     

Factors Affecting Instrumental Measurement and Visual Assessment of Colour Differences

The factors affecting the instrumental measurement of colour differences are reviewed. For most types of sample the performance of modern instruments is more than adequate provided that sufficient care is taken. Problems may arise with metameric pairs and fluorescent samples. The main source of error lies in the colour–difference equations/ this is quite separate from errors due to the instrument. Problems associated with visual assessments are discussed. The use of a panel of three observers will produce more reliable results and may be preferable under certain circumstances.     

A general metric for the magnitude of observer metamerism

Observer metamerism is defined as a property of a pair of spectrally different stimuli having the same colour sensation for an individual (reference) observer. Frequently, samples in this pair no longer match if the observer is changed. In this article, a linear approximation formula is developed that predicts a metameric effect caused by small changes in the observer's colour‐matching functions. This approximation formula enables a general metric of observer metamerism, the observer metamerism potential, to be defined that is independent of any particular deviated observer but still provides a close link to ‘observer‐metameric’ colour difference. Numerical experiments were conducted to investigate the correlation between the observer metamerism potential and the maximum of 53 metameric colour differences caused by the change from the colour‐matching functions of CIE standard 10° observer to the colour‐matching functions of 49 Stiles and Burch's real 10° test observers. The proposed general metric, together with a previous metric proposed by the present authors, the illuminant metamerism potential, could be taken as a quantitative measure of the performance of spectral approximation methods.     

A cross‐cultural comparison of colour emotion for two‐colour combinations

Psychophysical experiments were conducted in the UK, Taiwan, France, Germany, Spain, Sweden, Argentina, and Iran to assess colour emotion for two‐colour combinations using semantic scales warm/cool, heavy/light, active/passive, and like/dislike. A total of 223 observers participated, each presented with 190 colour pairs as the stimuli, shown individually on a cathode ray tube display. The results show consistent responses across cultures only for warm/cool, heavy/light, and active/passive. The like/dislike scale, however, showed some differences between the observer groups, in particular between the Argentinian responses and those obtained from the other observers. Factor analysis reveals that the Argentinian observers preferred passive colour pairs to active ones more than the other observers. In addition to the cultural difference in like/dislike, the experimental results show some effects of gender, professional background (design vs. nondesign), and age. Female observers were found to prefer colour pairs with high‐lightness or low‐chroma values more than their male counterparts. Observers with a design background liked low‐chroma colour pairs or those containing colours of similar hue more than nondesign observers. Older observers liked colour pairs with high‐lightness or high‐chroma values more than young observers did. Based on the findings, a two‐level theory of colour emotion is proposed, in which warm/cool, heavy/light, and active/passive are identified as the reactive‐level responses and like/dislike the reflective‐level response. © 2010 Wiley Periodicals, Inc. Col Res Appl, 2012     

Regression analysis to determine the optimum colour tolerance level for instrumental shade sorting

An investigation of the correlation between visual colour assessment and instrumental colour acceptance determination using regression analysis has been carried out. Three colour-difference equations, CIELAB, CMC(2:1) and CIE94(2:1:1), were studied in order to determine which is the best for generating a uniform colour space/microspace for allocating the colour population in shade sorting. Determination of optimum colour tolerance for further shade sorting was also undertaken. Some 1320 pairs of dyed samples distributing around 20 shade standards were measured instrumentally and also evaluated visually by a panel of 32 observers. Percentage rejection was plotted against colour difference and different mathematical regression relationships were then imposed. As a result, both CMC and CIE94 showed better correlation between the two colour assessment methods than the CIELAB colour-difference equation. Consequently, optimum colour tolerance limits were determined for subsequent development of shade sorting, with the findings being equally applicable to colour acceptance (shade passing).     

Grading textile fastness. Part 3: Development of a new fastness formula for assessing change in colour†,‡

In Part 1 of this series of papers, a new method for assessing textile fastness using an imaging system based upon a digital camera was introduced. The results from the system agreed well with those obtained using conventional spectrophotometers. Although the ISO fastness formula for assessing change in colour gave reasonable correlation with the available visual data, the data sets included very few test specimens. In Part 2 a new formula was proposed. This markedly outperformed the current ISO staining formula, ISO 105‐A04, and its predictive errors were encouragingly few. Now, in Part 3, two new data sets have been examined, together with an earlier data set, and observer and instrumental variances evaluated. A new formula for colour change has been developed, and again its predictive errors found to be fewer than those of a panel of observers.     

Deriving instrumental tolerances from pass-fail and colorimetric data

Equations such as CIE94 and CMC are now in common use to set instrumental tolerances for industrial color control. A visual experiment was performed to generate a data set to be used in evaluating typical industrial practices. Twenty-two observers performed a pass-fail color tolerance experiment for a single high-chroma yellow color. Thirty-two glossy samples varying in all three CIE-LAB dimensions were compared with a single standard. A near-neutral anchor pair was used to define the quality of match criterion. The pooled pass data were used to fit a 95% confidence ellipsoid. The chromaticness dimension was well estimated by either CMC or CIE94. The lightness dimension was poorly estimated by either equation. Evaluating the sampling distribution of the 32 test samples via a covariance matrix revealed a poor sampling, particularly in the ΔL*Δb* plane. This sampling may have biased the visual experiment. The visual data were used to optimize various color-difference equations based on CIE94 and CMC, where the l:c and total color difference were adjustable parameters. Several methods of optimization are described including minimizing the number of instrumental wrong decisions and logistic multiple-linear regression. Some methods require only pass response data, while others require both pass and fail data. Because industrial tolerances are usually based on a single observer, ellipsoids were fitted for three observers to demonstrate the large variability between observers in judging color differences. It was concluded that when tolerances need to be set based on a single observer's visual responses of samples not well distributed about the standard, typical industrial occurences, one should only adjust the tolerance magnitude based on a statistically valid equation such as CIE94. One should not change l:c or derive a new ellipsoid. © 1996 John Wiley & Sons, Inc.     

Ordinal scale based description of colour rendering

Perceived colour differences of 17 test colour samples (uniform standalone patches) were evaluated visually between a test and a reference light source on three visual scales. Two graphical rating scales (a greyscale‐anchored colour difference scale and a similarity judgement scale) and a five‐step ordinal rating scale (excellent, good, acceptable, not acceptable or very bad colour rendering) were used. The experimental setup included tungsten halogen, gas discharge, fluorescent, and white LED light sources at two correlated colour temperatures, 2700 and 4500 K. There was an inverse relationship between similarity judgement and visual colour difference results. Each category of the five‐step ordinal rating scale had a characteristic mean visual colour difference value. Visual colour differences correlated best with the recently developed CIECAM02‐UCS colour difference metric. Latter metric was used to predict the observers' ratings of visual colour differences on the above five‐step ordinal rating scale. From the predicted ratings of 17 test‐colour samples under the test light source, a new ordinal rating scale based colour rendering index (RCRI) was defined and compared with previous colour rendering indices. RCRI correlated well with both the mean visual colour differences and the mean similarity judgements. Despite the significant interobserver differences of the visual assessment of colour differences, the RCRI method showed an overall performance of 73% in terms of good predictions of the rating categories. Validation experiments with complex still life (tabletop) stimuli are currently underway. © 2010 Wiley Periodicals, Inc. Col Res Appl, 2010     

The impact of viewing conditions on observer variability for cross‐media colour reproduction

This study investigates observer metamerism in cross‐media colour reproduction (CMCR) between monitor and physical colours. An LCD display was placed inside a viewing cabinet. The cabinet had a mid‐grey interior. Observers were asked to match a colour displayed on the monitor to a printed colour patch, which was also inside the viewing cabinet in two configurations, one consisting of two samples separated by a hairline gap (Configuration 1), and the other consisting of two samples separated by a large gap (Configuration 2). Eleven observers were asked to first match the background colour and then 10 test colours for each configuration, and this was repeated five times. The observers’ variability results showed little difference between inter‐ and intra‐variability, and between the two configurations. Comparing the observers’ results with the CIE 1964 standard colorimetric observer, Configuration 1 and Configuration 2 each showed similar agreement. In both configurations, the results of observer variability were smaller than those obtained by Oicherman et al. in 2008. In addition, each configuration's results showed better agreement with the CIE standard colorimetric observer than Oicherman et al.'s results. This implies that both configurations, one with two samples having a hairline gap and with two samples separated by a large gap in a viewing cabinet, could be recommended for future CMCR systems. However, if choosing between the two, then Configuration 2 is recommended rather than Configuration 1.     

Small and Moderate Colour Differences

In nineteen regions of colour space groups of about 30 samples, differing by small to moderate amounts from one designated as ‘the standard’ were generated by offset printing. Twenty observers estimated the perceived size of each of these differences twice, using cateogry scaling - the judgements were not based on acceptability. The samples were placed, in turn, adjacent to the standard on a middle gray surround, subtended 4d? on the retina, were illuminated with semidiffuse fluorescent light, simulating daylight at about 45d? incidence, and viewed along the normal. The visual data were converted from an ordinal to an interval scale. The samples were measured by abridged spectro-photometry, and CIE coordinates, obtained by integration using the spectral-power distribution of the illumination source and the 1931 2d? standard observer. Colour differences were calculated by eleven formulae and correlated with the visual interval-scale results. The usual low correlation coefficients were found. The data are analysed colour by colour and significant differences are noted. Collectively, the Adams Chromatic Value (ANL AB 40), Saunderson-Milner, and CIE 1976 L*a*b* formulae gave better performances.     

Accuracy of metameric indices in relation to visual assessments

Metamerism is one of the most fundamental perceptual phenomena of the visual system and has been extensively studied. It can be visualised in terms of two pieces of fabric, a standard and a batch, that match under one type of light, e.g. daylight, but fail to match under another, e.g. tungsten light. This project looks at the agreement between observers on the colour difference of metameric pairs and how the metameric indices for these pairs relates to the visual assessments. Twelve sample metameric pairs were prepared using computer dye recipe prediction and were assessed visually by two methods, one with adaption of the eye to the light source under which the sample was being matched and the second without adaption. There was a low degree of agreement amongst the observers. Poor correlation was observed between the metameric indices, based on both the colour difference and the difference in reflectance curves. and the observer assessments.     

Categorical observers for metamerism

A method is proposed to generate categorical colour observer functions (individual colour matching functions) for any field size based on the CIE 2006 system of physiological observer functions. The method combines proposed categorical observer techniques of Sarkar et al with a physiologically-based individual observer model of Asano et al and a clustering technique to produce the optimal set of categorical observers. The number of required categorical observers varies depending on an application with as many as 50 required to predict individual observers' matches when a laser projector is viewed. However, 10 categorical observers are sufficient to represent colour-normal populations for personalized colour imaging. The proposed and recommended categorical observers represent a robust and inclusive technique to examine and quantify observer metamerism in any application of colorimetry.     

Evaluation of daylight simulators. Part 2; Assessment of the quality of daylight simulators using actual metameric pairs†

This paper discusses two methods for evaluating the quality of daylight simulators, namely the band‐value method specified in British standard BS 950 and the CIE metamerism index method. Six daylight simulators of various types and manufacturers were used for the assessment and a psychophysical experiment was conducted to evaluate the two methods. A range of 70 metameric pairs was assessed by a panel of observers under each of the six simulators and the reliability of their visual results was examined in terms of observer accuracy and repeatability. The visual results were also compared for each pair of simulators to reveal how the results could be affected by using different simulators. The effectiveness of four colour difference formulae was tested using the visual results. Finally, four methods were developed using two statistical measures of performance factor, band‐value deviation and CIE metamerism index for evaluating the quality of daylight simulators.     

Methods for quantifying metamerism. Part 1ndashVisual assessment

This work forms part of the Society of Dyers and Colourists' CMC Working Group 1 project on metamerism and colour constancy. The aim of the project is to derive reliable indices for predicting metamerism and colour constancy. Some 76 metamers were prepared by dyeing wool with acid dyes. Two experiments were conducted to quantify the degree of metamerism using a grey-scale method. In the first experiment each metamer's colour difference was assessed against a grey scale by a panel of observers under seven light sources. The second experiment was carried out under three sources: D₆₅, A and TL₈₄. Each observer assessed not only the total colour difference as in the first experiment but also the separate lightness, chroma, and hue components. Observer variations were investigated in terms of observer precision, and within-and between-observer errors. Cross-over wavelengths for each metamer were also examined. It was found that intersections tend to converge on three wavelengths for almost all metamers used here, in agreement with previous findings.     

Comparison of unique hue stimuli determined by two different methods using Munsell color chips

Unique hue stimuli were determined by male and female observers using two different visual experimental procedures involving Munsell color chips of varying hue but identical chroma and value. The hypothesis was that unique hues can be more reliably established by explicit selection from a series of ordered stimuli than implicitly by hue scaling a series of stimuli in terms of neighboring UHs and this was statistically confirmed. The implicit selections based on long term memory of UHs appears to have been more challenging to observers since variability was increased by nearly 50% compared to when UHs were explicitly selected. The ranges of unique hues selected in the two methods were, however, comparable and no statistically significant difference was found between the results of females and males. The intra‐observer variability in picking a stimulus to represent a unique hue, for all observers and averaged for all hues, was approximately 12% of the mean spread of unique hues, confirming that the large inter‐observer variability is driven by differences in color vision and perhaps cognitive processes. © 2010 Wiley Periodicals, Inc. Col Res Appl, 2010     

Effects of age and ambient illuminance on visual comfort for reading on a mobile device

A psychophysical experiment was carried out to investigate visual comfort for reading on an iPad under various illuminance levels, ranging from 50 lx to 1200 lx, in order to see whether and how the following variables can influence the visual comfort: observer's age, gender, the illuminance of ambient lighting, and the background colour in a document layout. A panel of 21 young and 22 older Taiwanese observers participated in the study. The paired comparison method was used for data collection and analysis. The experimental results show that for all lighting conditions, young observers tended to prefer reading documents that had a moderate CIELAB lightness difference between text and background, while older observers tended to prefer reading those with an extremely large lightness difference. The results also show that female observers tended to feel less comfortable than male observers when reading documents with an extremely large lightness difference. These findings were found not to be affected by the ambient illuminance. Regarding the influence of document background colour on visual comfort, the observers tended to feel more comfortable reading documents that had a gray background than reading those with a background colour of either white or black. It was also found that the visual comfort was slightly higher for positive polarity than for negative polarity, but the difference between the two settings was insignificant. © 2016 Wiley Periodicals, Inc. Col Res Appl, 42, 352–361, 2017