Site-wide links

Measurement FAQ

 

 

How do one estimate color of an object by just looking at their reflectance data? What is the relationship?   (answer)

I am confused by the range value of the Y-tristimulus value. For emissive things, like screens, it is supposed to be the absolute luminance in foot-lamberts (so the range could be 0 to infinity in theory).

If want to compare two colors (via deltaE - which deltaE is the 'standard' nowadays), one from a emissive source and one from a reflective source, can I? It seems like a reflectance number couldn't quite compare to a direct luminance measure, but then how can I compare these two things in a color space?   (answer)

Can you explain the science behind gloss levels on the paintwork of a car. I own an auto reconditioning business and I know using different techniques I try to achieve maximum gloss, even at a microscopic level. Let's say you have the spiderwebbing swirls in your clear coat. Scientifically , what is happening to the light when it hits the swirls as opposed to hitting perfect paint? Is it the law of Refraction or should it be explained another way.   (answer)

Why is light source data normalized to 100 at 560 nm? Is this the source´s relative spectral power distribution or is that something else? Doesn't the normalization affect the computed XYZ or L*a*b* values?   (answer)

I'm interested in skin colour and I normally reckon it should lie between 0° and 30° in a HSB colour space. In the Munsell space the Hue shifts for less saturated colors. Are there recognized skin values in the Munsell notation?   (answer)

Is there any way to predict what color an object will appear if I change the color of the light source?   (answer)

Imagine driving down a street toward a traffic light. How fast would you have to go so that the red light (700 nm) would appear green (500nm)?   (answer)

Is there a way to measure gloss with a spectrophotometer or is it necessary to have a glossmeter? Is there a conversion between CIELAB values and gloss?   (answer)

What are L* a* and b*? Are they constants? If not, how can I find those values?   (answer)

Where can I find calculator or software for LRV?   (answer)

Is there is a device that can measure liquid gravure ink spectral transmittance? I need a method to confirm ink recipe components in a liquid ink sample. I could use a faster option than the traditional draw-down proofing method.   (answer)

How many lights are metameric to white light reflected from a paper?   (answer)

How did the CIE 1931 color matching experiments evaluate the whole gamut when we know that the RGB color gamut is a subset of that of the human eye? In other words, we know there are colors that cannot be matched by the RGB lights. How did they get around this?   (answer)

I measure timber colour using a Minolta Chromameter CR-400. I want to compare my values with those of other woods whose colors are listed in percent Luminance, Dominant wavelength (λ)and percent purity. Can I convert these values over to CIELAB?   (answer)

Is it possible to scientifically determine the original colors in an old black and white stereoview photograph if some of the colors are known?   (answer)

What wavelength(s) are used in the calculation of Gardner color?   (answer)

In his 1943 OSA paper "Specification of Small Chromaticity Differences" MacAdam describes constants g11, 2g12, g22 and gives values for them in graphical form. Is there a source for this data in conveniently tabulated form?   (answer)

I need a device that will give me color readings at an observer angle of around 45 degrees or so that will pick up almost 100% specular reflection.   (answer)

Why is it not possible to get an accurate ΔE colour measurement on a metallic powder coating?   (answer)

Do you know the origin of a measurement called the "flop index", which is defined as flop index=2.69(L15�-L110�)^1.11/((L45�)^0.86). My Xrite M68II reports this metric, however, X-Rite does not explain the source of this equation. Do you know where I might find the source of this equation, or another good way to mathematically explain the color difference at several angles?   (answer)

Are there any specific details in 45:0c reflectance measurement, such as the effect of distance between detector and sample or distance between light source and sample? Are there restrictions on the circumstances, like the color of the floor on which we put our sample?   (answer)

Is there any approximation to convert between Hunter Lab (illuminant C, 10 degree observer) and CIELAB (illuminant D65, 10 degree observer)?   (answer)

I received data from two sources. The first measured Hunter Lab on light source C under 10 degrees, the second measured CIELAB on light source D65 under 10 degrees. I don't have the spectra. Is there any approximation with which I can convert between them?   (answer)

What observer for XYZ values can be used in CAM02, 2 degree or 10 degree? Some articles use 2-degree observer, the others use 10-degree observer.   (answer)

Would a spectrophotometer be able to accurately read the color differences between chrome plated surfaces? (ie. mirror-like finish) What settings or type of instrument would best achieve this?   (answer)

Is it possible to quantitatively describe a color just by its reflectance or radiative spectrum?   (answer)

Is there a color model which would allow me to do the following? Given L*a*b*/XYZ/whatever coordinates of a color stimulus, given specified viewing conditions and color, what is the shape/size of the JND "ellipsoid" around the color, and what are the directions of the hue/chroma/lightness axes?   (answer)

Is 0.05ppm a suitable tolerance for measuring absorption on a spectrometer for beverages? When is it no longer a visible difference to the naked eye 1ppm, 0.1ppm, 0.01ppm, etc?   (answer)

In the CIELAB color space, what is the approximate difference for each of L*, a*, b*, and ∇E that most people would not percieve as a different color.   (answer)

Where can I find a chart showing color names, the colors themselves and CIELAB color coordinates? Ideally, this would also show the RGB color coordinates. I have found many RGB charts like this but none with the CIELAB coordinates.   (answer)

I need to know the Munsell color for 'school bus yellow.' The unsourced quote I have says it is 9.4YR, 7.5/9.2.   (answer)

Where can I purchase a poster size print of the CIE 1931 Chromaticty Diagram? I realize that the diagram is only a 2D projection of 3D matching data however I still find it useful to illustrate some of the other attributes of the data depicted (color mixing, color temperature, color gamuts)?   (answer)

Is black a natural color or is it just a very dark brown that people just see as black?   (answer)

We would like to know the complete Reflectance Measurement principle and formula for the dual beam color spectrophotometer including Light trap, PTFE standard and sample.   (answer)

How would I measure the yellowness of the sclera (whites) of the eye? Also, what instrument would be an appropriate measure haziness of the cornea?   (answer)

Is 0.05ppm a suitable tolerance for measuring absorption on a spectrometer for beverages? At what concentration is the difference not perceptible?   (answer)

"The Desktop Color Handbook" mentions using a "light source color rendering card" which has patches that look the same under light of the appropriate CCT, but different under other lights. Do you know who makes such cards, or how one could create one?   (answer)

Is there an accepted universal standard for the calibration of color Spectrophotometer? Are there white standards available with certified reflectance values?   (answer)

Are there maximum/minimum values for a* and b* like the min and max for L*?   (answer)

Do you know where I can purchase Hess Ives Color standards, both physical standards and papers publications?   (answer)

Is there such an instrument as a hand-held spectrometer for measuring the color temperature of light sources? If so, how much do they cost?   (answer)

Densitometers can be purchased with or without polarizing filters. Why are both types used? Which type provides the best correlation to perception?   (answer)

How do you convert ΔL ΔC ΔH values for colour differences in CIELCh to the equivalent ΔL Δa Δb colour difference values in CIELab, when you do not have access to the absolute values for the original measurements?   (answer)

Are there some basic averages for the LRV of human skin?   (answer)

I have used three online color space conversion calculators and I get three different sets of numbers when I convert from XYZ to RGB. What's happening? Who's right?   (answer)

Roughly how large is the overall remaining perceptual non-uniformity of color tolerances in CIELab space, when using the ΔE94 and ΔE2000 color distance formulas? (as compared to the 6:1 non-uniformity of simple ΔE*ab)?   (answer)

How do I calculate the dominant wavelength of a color sample with X,Y,Z or x,y? I do not want to draw it in the colorspace.   (answer)

How to convert CIELAB data to Yellowness and Whiteness index?   (answer)

What type of Pantone Guide shall we refer to & which instrument is suitable to measure the values of colour on aluminum foil with a flexo printing process? Since Aluminum has glossy as well as the matt finish then is there any need to change the settings of the instrument?   (answer)

We print on Aluminum foil with a flexo process. What type of Pantone Guide shall we refer to & which instrument is suitable to measure the values of colour? Since Aluminum has glossy as well as the matt finish then is there any need to change the settings of the instrument?   (answer)

How do I convert from spectral reflectance (from a hyperspectral imager) to CIELAB colorspace?   (answer)

Can a consumer grade flat bed scanner be used as a colorimeter?   (answer)

What are the colors of a complete spectrum? (Starting with the color with the longest wavelength.)   (answer)

What color measurement system would be best to define colors for a liquid, to be described in a US patent?   (answer)

What is the difference between the L*a*b* values and the L a b values when using a Minolta colorimeter?   (answer)

My spectrodensitometer states its measuring geometry is 0/45, and its standard observers are 2 and 10. Are these angles related? Which should I use?   (answer)

I am working on a plastic matched automotive black (GM 848) where the color is visually darker, but it measures LIGHTER than the standard. Can you please explain why?   (answer)

What is the mathematical equation to convert Lab or XYZ to Density?   (answer)

We know density is derived from the transmitted spectral power distribution. Can you suggest a reference that provides the responsivity functions for Status A and Status M densities, as well as the specific function for the calculation of density? Also, what is the proper illuminant to use?   (answer)

How can I meaure the color of clear solutions? We want an objective measurement (wavelength) of colors of various juices and wines.   (answer)

I am teaching color theory to our press operators and I am wondering if there is a full color spherical L*a*b model illustration that I can download, and if necessary purchase.   (answer)

Is there an existing paint industry standard system for expressing (and thus potentially matching) all the visual characteristics of a metallic or metallic flake paint, whether one-coat or two-coat (colored transparent over metallic or flake)?   (answer)

I am trying to calculate the colour of a coating. I have a graph of wavelenth versus reflectivity. Is there a simple way to convert this graph into RGB values or some similar standard?   (answer)

We have two items painted with Munsell N7 which are significantly different in shade. Is this normal? How can I measure if the differences are within specification?   (answer)

Why isn't E point (x=0.3333 - y=0.3333) used as a standard illuminant for color matching and ΔE?   (answer)

I am working on honeybee color vision. Using a honeybee specific color diagram, we know how to calculate dominant wavelength. How do I calculate the dominant wavelength of a color sample which resides in my color diagram in the "magenta" (or purple) zone.   (answer)

I work for an automotive finishing company and use an X-rite MA68II for measuring color. Which should I use, L*a*b*, or L*C*h�? Everyone seems to use L*a*b*, but I'm not sure why.   (answer)

I am developing paint color regulations to help buildings constructed on mountainsides blend into their surroundings to minimize their visual impact. What maximum LRV do you recommend for an environment typically consisting of darker browns, grays, and greens as the "base" colors in order to keep buildings from "standing out"? What would be the most appropriate instrument for measuring the LRVs of paint/stain for enforcement purposes?   (answer)

How does the color of a room affect people's moods?   (answer)

What are the equations for McAdams ellipses? Are they location dependent in the 1931 chromiticity diagram?   (answer)

It makes sense that I should want window glass with the highest Color Rendering Index available in the performance category that I need; but how do I judge the relative CRI when all options are between 87 and 95?   (answer)

Can CAM02 to predict Appearance Correlates of a color on different colored backgrounds?   (answer)

Is it possible to draw plankian line in CIE diagram? How?   (answer)

Why does the sheen of a finish affect the appearance of a color? Do the wavelengths of color get magnified, refracted, or intensified as the pass through the layers of a paint film?   (answer)

Can you explain the difference between value and luminance? Is Light Reflectance Value different than luminance?   (answer)

Where can I buy Color Standard for measurement calibration?   (answer)

To measure process color is it better to use a spectrometery or density, hue, and gray?   (answer)

Is there an index for the degree of yellowing? I want to characterize the degree of yellowing in trees as an indicator of poor health from aerial RGB photographs.   (answer)

Where can I buy the Munsell color chart that has the 14 samples used for CRI tests?   (answer)

Are there any instruments or methods which measure the intensity of gray color?   (answer)

I am trying to establish a color system criteria for silicon parts which are white. What do you suggest in setting up a criteria for the color white?   (answer)

What is meant by plankian locus? What is its importance?   (answer)

How do I calculate whiteness index (WI by ASTM standards) if I don't have a spectrophotometer but am in possession of reflectance data for materials that have <95% reflectance @ 600nm?   (answer)

I need a fairly fool-proof way for a moderately skilled technician to measure a large number of samples and present the data using the CIECAM02 coordinates. I am willing to buy a new instrument/software. Any thoughts?   (answer)

How can I get spectral reflectance information about Color-Aid samples?   (answer)

hat is the correlation between Hunter whiteness (L, a , b) and CIE whiteness (L*, a*, b*)? Can I relate them mathematically?   (answer)

I have a Minolta 2600d portable spectrophotometer. Can you tell me the best way to measure the fluorescence of teeth in vivo? Also, can I use the 2600d to measure the opalescence of both natural teeth and dental materials?   (answer)

What are the calculations used to convert spectrophotometric spectral data to density, CIE XYZ, and RGB variables/values?   (answer)

How can I convert CIELCH to CIE 1931 x,y coordinates?   (answer)

What is the utility of converting tristimulus values to L*a*b* values?   (answer)

Can my company's color references be accurately reproduced 10 years from now based on the tristimulus values we maintain ?   (answer)

What does the ratio b/a give or mean? Here a is negative and b positive.   (answer)

Can you help me transfer color temperature to 1931 CIE xy-coordinates?   (answer)

I'm measuring several 3000K white lamps. I'd like to know the color variation of these lamps. Is there any other way to measure color variation of a population of lamps, aside from standard deviation and MacAdam ellipse?   (answer)

What are good books to get detailed information on both the Munsell and CIE systems? I'm an artist and have a general education of both systems but would like to know indepth knowledge of each, including the math and reasoning behind the constructions of each system.   (answer)

How is the location of an illuminant on the Planckian Locus and R-G Axes related to colour vision tests?   (answer)

What is the "straight line rule" when referring to the CIE 1931 Chromaticity Diagram?   (answer)

If one measures same color on various spectrophotometers the spectral profile is the same,but there are differences of 2-3 % in reflectance. Can you explain why? Which is the most correct reading?   (answer)

Can I use BCRA standard tile for calibration of spectrometer? Where can I buy it?   (answer)

What are the advantages and disadvanatges of chromaticity diagrams?   (answer)

Can you please define "dominant wavelength" and "wavelength of maximum absorbance"? Which should be used in the calculation of strength?   (answer)

Would you recommend a source for standards tile for calibration of spectometer?   (answer)

What is the meaning of NET profiler of a color spectrophotometer?   (answer)

When I measure delta L on low-gloss (matte finish) parts, I notice more variation between measurments than with higher gloss parts. Is there a scientific reason for this? Does the gloss play a roll?   (answer)

I evaluate llighting systems by measuring u'v' with a spectroradiometer. How can I calculate deltaE color difference from this data?   (answer)

How is the Light Reflectivity Value (LRV) measured? Why do building codes require a certain percentage of LRV in the materials used?   (answer)

What sorts of jobs would a color scientist do?   (answer)

What is the relationship between the CIELAB a*-b* opponent axes and the perceptual unique hues?   (answer)

I'm looking for a Circular Spherical Vision Spectro Colorimeter for analyzing luminance and other parameters (axis etc...) of reflective materials in different colors for road construction. Can you help?   (answer)

I measure pigments from bird egg shells with a spectrophotometer CM-2600d, should I work with 100%UV or 0%UV? What about pigments from bills and legs of birds?   (answer)

What is chromatic textile? What properties do this textile have?   (answer)

How do I compute the correlated color temperature of an Lab/Luv color coordinate?   (answer)

I'm trying to understand whether there is an optical/physical basis for this "chameleon effect" (the ability of the materials to blend with surrounding tooth structures to create naturally appearing restorations) and whether there may be way to measure it.   (answer)

Can you help me calculate the color difference between my Macbeth Colorchecker and its image after capture and display on an LCD monitor?   (answer)

I would like to purchase a grey scale color variation kit for comparing the change in color wash for Denims and Textiles. Can you recommend a source for this kit?   (answer)

Where can I find a tabulation of all pairs of distict spectral monochromatic lights that are perceived as achromatic?   (answer)

How do I calculate the chromaticity coordinates of a single wavelength of light?   (answer)

What is the equation for plot a MacAdam ellipse? Also, what steps are the Macadam ellipses on the standard x,y chart and what do these "steps" mean?   (answer)

What is the best method to measure the opaque black color of a plasic material?   (answer)

What's the RGB composition of RAL 7005 & RAL 1021?   (answer)

How can I measure the LRV value of something. Or can I convert CIELAB data into an LRV value?   (answer)

What material has the most uniform reflectance spectra, including visible, IR, and UV?   (answer)

Can you help me convert between Pantone and RAL colors?   (answer)

What does the color blue mean?   (answer)

When reading a violet, what does the Δa and Δb values represent? Example is the Δa dirty/clean or red/blue?   (answer)

I am wanting to know how broad the primaries were for the 1931 color matching experiment. Were they monochromatic, as in a laser? If not, what was the spectral half width at half max for each of the colors?   (answer)

I would like to measure the color intensity of various parts of birds. What instruments will I use?   (answer)

Can you help me convert between candella and power wattage as measured by photodiode?   (answer)

What is the minimum ΔE that a human eye can detect? Which &Delta E"   (answer)

What's method that appropriate for measure color of gemstone?   (answer)

Can you help me find a cost-effective way to measure black-gray-white and all shades in between. My son is doing an science experiment to determine how well erasers work.   (answer)

Specifications on the web given by SAE, ECE eg, mention that the colour should fall within the given boundary:
y = 0.335 (yellow boundary)
y = 0.980 - x (purple boundary)
What does this mean?   (answer)

Is there anyone who I can pay to convert 12 colors with LCh values into Lab or Munsell values?   (answer)

How can I make a colored chromaticity diagram?   (answer)

What do the various CIE illuminant designation mean (D55, D65, etc)? How are they related to each other and to CCT?   (answer)

Can you help me transform Lch values to Lab?   (answer)

In literature it is mentioned that the difference between SCE and SCI color measurements gives an indication for the gloss of the surface. Which parameters are used to evaluate this: the difference between the SCI and SCE values for CIE-L, a, b or a combination?   (answer)

What would be the reason to use CIELAB, CIELCH, CMC, CIE94, and CIE2000 (all of the above together) to calculate color difference from sample to standard?   (answer)

What is the best way to test color fading in construction paper? My child is making an experiment on what color will fade the fastest when exposed to sunlight.   (answer)

Is there direct co-relation between CIE 76 and ANLAB 40 and ANLAB 50 color spaces?   (answer)

What is the primary calibration of a spectrophotometer?   (answer)

Can you derive CIELAB values directly by visually inspecting the reflectance curve?   (answer)

Can you explain the mechanism of UV filter in color spectrophotometer, to get universal Whiteness values on all spectros. Is it a wedge filter or an attenuator filter.   (answer)

I work in printing. Can you tell me what the effect of dot gain is on delta E? Also, can we get an accurate delta E measurement from a coated sheet to an uncoated sheet?   (answer)

How do I figure out what color a reflection of another color might be? For example, if I have a yellow car and something red is reflecting off of it, what color will show up as the reflection on the car?   (answer)

What effect renders UV lamp when selecting the colour with use light booth?   (answer)

When visually comparing colors, why is D65 light source better to use than Illuminant C?   (answer)

We are looking at colour and redness of scars on human skin. Which is better for the job, a hand held spectrophotometer or a hand held chromameter? Some literature validates the use of a chromameter CR300 in skin research, but spectrophotometers are apparently more accurate.   (answer)

I can get x,y,z color coordinates. So how would one relate to the materials response for a particular wavelength?   (answer)

How can I calculate hue angle from CIELAB? What is pi?   (answer)

What is the RAL number that should be used in the background of a light cabinet?   (answer)

We are a printing company. We print offset (litho), gravure and flexo, both on coated and uncoated paper. What spectrophotometer should we buy spectrometer for Quality Control and Color matching?   (answer)

I need to measure human skin color directly. What instrument should I use?   (answer)

How do yo measure ΔE*cmc, ΔE*94, ΔE*fmc2 from the L*, a*, b*, C*, h* ΔH*, ΔE*?   (answer)

When measuring the color of colored plastic plaques with a spectrophotometer, how thick should the plaques be and why?   (answer)

I am doing a problem on image permanence for large format inkjet. Should I use Delta E or Delta E 2000? Why and what is the difference between the two?   (answer)

We always measure the color difference by using d/8 spectrophotometer with included specular reflectance. But why do we always measure the white with excluded one ?   (answer)

I have an emission spectrum from a sheet of luminescent material. Can you help convert the emission spectrum to an RGB color to display?   (answer)

What is the best color model for use with colors that have a low saturation? CIELAB is not working for me because when a* and/or b* are close to zero I get poor values for hue.   (answer)

I have to purchase a system to compare shades of yellow in plastics with rough surfaces. Would you please recommend a company and/or instrument that would have good repeatability and be low maintenance   (answer)

Do the tiles in the Farnsworth Munsell 100 Hue test kit require calibration? If so, what can change that would require their recalibration? If they require calibration, what is the frequency?   (answer)

Does there exist a (known) correlation between a Minolta CM3500D and a BYK Gardner or does this depend on the product type?   (answer)

CIE chromaticity diagrams are usually exhibited in a shape of a hoof. Why are the single waves spread on the ellipse side instead of the straight?   (answer)

Can I determine the color rendering index of this source purely from the light-emission spectrum?   (answer)

Can you provide an example showing how to calculate ΔE using Δa*, Δb*, ΔL*, ΔC*, Δh for textile pigment?   (answer)

Can you help me compare RAL specification paint with ANSI paint for color match?   (answer)

Please can you explain what Macadam limits are (as mentioned in description of the 'real' set of Munsell values on your Munsell Renotation Data page)?   (answer)

I am a dental student. As part of my project I need to convert the RBG values obtained from Adobe Photoshop into CIELAB values in order to compare with values obtained with a spectrophotometer?   (answer)

Can you explain how to select the color center (or sample color) which can be used to evaluate color difference formula?   (answer)

As we know black colours, for example on textiles, usually have a pale tint. Is there any formula for measuring blackness?   (answer)

Would CIECAM02 or CIEDE2000 be better at predicting spot color differences?   (answer)

Can mixed paint be measured while it is still wet? In our factory we mix paint, we want to measure it's color before painting parts.   (answer)

I am trying to describe the colour of my sediment cores and have L*, a*, b*, and C* data from a spectrophotometer. I am looking for some directions on how to 'describe' the colour of my mud! Is there some tool or guide to convert numbers into a colour category?   (answer)

How does one measure color of ink in its raw liquid form? I need to measure the color in that form. What tools are required? Are there standardised methods to sample ink to measure it?   (answer)

Each of our spectrophotometers has its own calibration standard. If all of the spectros are not calibrated to the same standard, aren't we calibrating in variation between the spectros? We can see anywhere from a .2 to over a .9 difference in our ΔE readings when we read the same areas of a sample on the different spectros. What would happen if we lost one of the standards since they are suppose to be unique to the manufacture of one specific specto?   (answer)

I need a means to estimate the difference in general light reflection between walls in a windowless room painted matt magnolia (Munsell 10YR9/2), compared to matt brilliant white.   (answer)

What is the Reflectivity Index and what is the index for various colours, particularly silver?   (answer)

When I transform D65 illuminant values (in CIELAB) to RGB, I do not get R=G=B. Why?   (answer)

Various color systems (NTSC, SMTPE, CIE) use different white points. How are these defined?   (answer)

Can you provide information on Ultralume U5000? Also, what are ΔE GE, ΔE Audi,and ΔE PQ?)   (answer)

What are these devices: densitometer, spectrophotometer, and colorimeter?   (answer)

What is the equivalent RAL paint code for Munsell 3Y7.8/1.1?   (answer)

I want to write a paper on color measurement. Can you tell me the latest developments on this, and what we can do?   (answer)

What is LRV?   (answer)

I use a colorimeter that returns x,y,Y values. What units do they express?   (answer)

I have a densitometer that measures visual density. How do I convert the density values to CIE L*?   (answer)

I have a white printing on LDPE which is yellowing. How can I determinate the yellowing in a datacolor instrument?   (answer)

Is there a way to express Lch differences, in terms of units? (not referring to ΔL*, ΔC* or ΔH*) Hue for example is expressed in degree?s (eg 5°). What unit do I use to express the difference in lightness and chroma?   (answer)

Could you recommend a book that would serve as a good introduction to colour science for the lay reader? I am interested in learning the basics of colour science but am not a scientist or mathematician.   (answer)

Where I can find the exact value of Xn, Yn (the luminance factor for the illuminant or perfect white) and Zn?   (answer)

Where I should look for for the energy distribution of commercially available sources?   (answer)

What is the best method (color space, sci vs. sce, degree observer, Illuminant,etc.) for calculating Light Reflectance Value (LRV) on a spectrophotometer?   (answer)

Where I can find the ultraviolet reflectivity index of various colors? How about the heat absorption index for various colors?   (answer)

Is there a reference offering a method for the bi-directional calculation of densitometric quantities (particularly status M) and spectral transmittance to be used as weighting coefficients in the integration to solve for XYZ tristimulus values?   (answer)

What is the basis for 18% reflectance being the center of the lightness scale in many different systems? For example: Munsell, Ansel Adams zone system, and CIELAB. Coincidence? I doubt it.   (answer)

Would you explain the meaning of ΔH*?   (answer)

What is the scientific termonology for when you fold a piece of fabric and two areas appear to be entirely different colors? I've seen this demonstrated: after folding, both sections of fabric appeared to be different colors.   (answer)

I need to find a way to measure the saturation and brightness of printed colors and pen ink. I need to select hues of equal brightness and saturation in my studies. Can you provide information about machines that can make such readings?   (answer)

I have used a spectrophotometer to analyse colour samples. How do you interprete the curves, for example prussian red, since there are many peaks shown?   (answer)

My technical director asked if I could find an "easy to understand" explanation of "CIE 2000." He may mean the CIE DE2000 mentioned in this publication. Is there a publication or expert that could assist us?   (answer)

Can you provide me a graph of grayscale Values (0-255) verses density? I need it to evaluate a grayscale calibration wafer.   (answer)

How can I specify a color so that manufacturers the world over will use the same color paint?   (answer)

In relation to Kodak absorption filter graphs; What does "Transmittance" & "Diffuse Density" mean? and the difference between them?   (answer)

Can you provide spectral reflectance values for the Farnsworth-Munsell 100 Hue test?   (answer)

My spectrophotometer is getting old. I need a new instrument and software for evaluation and formulation of color paints and ceramics. What recommendations can you make?   (answer)

Would you explain what ΔE is and how is it measured?   (answer)

What is L-star?   (answer)

What is the established color difference tolerance for the plastics industry?   (answer)




How do one estimate color of an object by just looking at their reflectance data? What is the relationship? (926)
You cannot estimate the color accuracy from just the reflectance. The color depends on the combination of that reflectance distribution and the visual responses to light. That is why it is possible to have different reflectance spectra produce the same color. (Back to top)


I am confused by the range value of the Y-tristimulus value. For emissive things, like screens, it is supposed to be the absolute luminance in foot-lamberts (so the range could be 0 to infinity in theory).

If want to compare two colors (via deltaE - which deltaE is the 'standard' nowadays), one from a emissive source and one from a reflective source, can I? It seems like a reflectance number couldn't quite compare to a direct luminance measure, but then how can I compare these two things in a color space? (922)

The normal procedure is to select a luminance level that represents white (like the white point of a display) and then normalize the absolute tristimulus values to that point (making it 1 or 100% to match the reflective values). Then the two sets of values will be on the same scale and can be reasonably compared. (Back to top)


Can you explain the science behind gloss levels on the paintwork of a car. I own an auto reconditioning business and I know using different techniques I try to achieve maximum gloss, even at a microscopic level. Let's say you have the spiderwebbing swirls in your clear coat. Scientifically , what is happening to the light when it hits the swirls as opposed to hitting perfect paint? Is it the law of Refraction or should it be explained another way. (918)
It is probably a combination of things, but ultimately it comes down to scattering caused by diffraction at the edges of scratches and other imperfections. The diffraction redirects the light that would be reflected into a single angle on a perfectly glossy surface. You might also have some irregularity in the surface (such as orange peel) causing it to look less glossy since the gloss reflection will be in different directions for different locations on the surface. (Back to top)


Why is light source data normalized to 100 at 560 nm? Is this the source´s relative spectral power distribution or is that something else? Doesn't the normalization affect the computed XYZ or L*a*b* values? (917)
The normalization is a tradition. Any wavelength could be chosen, but 560nm was selected since it is the middle of the visual spectrum and approximately the wavelength of maximum visual sensitivity. Yes, these are the relative spectral power distributions and the normalization does not affect the XYZ or CIELAB computations because there is another normalization in the computation of XYZ (the k factor in the equations) that makes this one unimportant. (Back to top)


I'm interested in skin colour and I normally reckon it should lie between 0° and 30° in a HSB colour space. In the Munsell space the Hue shifts for less saturated colors. Are there recognized skin values in the Munsell notation? (913)
There isn't a specific standard, but one resource, The "ISCC-NBS Method of Designating Colors and a Dictionary of Color Names" indicates that colors that are named "flesh" and "pale flesh" can be found with Munsell hues ranging from 9RP to 8YR (that's roughly 36 degrees worth of hue). Of course, whether or not a color of those hues looks like skin would also depend on the lightness and chroma. There might be other synonyms for skin color in the ISCC dictionary if you want to search more (it is out of print, but usually can be found in online used book searches like bookfinder.com). (Back to top)


Is there any way to predict what color an object will appear if I change the color of the light source? (907)
Yes. The CIELAB values are computed from the measured spectral reflectance of the object. You could do these computations yourself for the illuminant of your choice (compute CIE XYZ and then CIELAB values from measured reflectance). A text on colorimetry (e.g. Berns' Principles of Color Technology, 3rd Ed.) can provide an overview of all the equations and their application. Another alternative is that your instrument manufacturer might be able to enter your custom illuminants into their software. (Back to top)


Imagine driving down a street toward a traffic light. How fast would you have to go so that the red light (700 nm) would appear green (500nm)? (893)
It would be approximately 2/5 times the speed of light. Faster than your car could go and stay intact. Sorry. (Back to top)


Is there a way to measure gloss with a spectrophotometer or is it necessary to have a glossmeter? Is there a conversion between CIELAB values and gloss? (892)
You cannot convert CIELAB values to gloss as they are designed to measure the diffuse color and not the gloss. You can get some measure of gloss with an integrating sphere spectrophotometer by comparing the Y tristimulus values for measurements in specular included and specular excluded geometries. A diffuse sample will have no difference while a very glossy sample will have a difference on the order of 4% or so. It is better to have a gloss meter. (Back to top)


What are L* a* and b*? Are they constants? If not, how can I find those values? (890)
L*, a*, and b* represent the lightness, redness-greenness, and yellowness-blueness dimensions in the CIELAB color space. The values are not a constant, but rather are different for every object and illumination condition. They are found by measuring the spectral reflectance of materials (or spectral radiance of self-luminous stimuli) and doing some colorimetric computations. I would recommend a good introductory text on color measurement like Berns&apost; "Principles of Color Technology, 3rd Ed." to get a good overview of the CIELAB space. (Back to top)


Where can I find calculator or software for LRV? (889)
I don't know of a specific LRV calculator or software, but it appears that LRV is the same as the CIE Y tristimulus value which is readily obtained from color measurement instrumentation. (Back to top)


Is there is a device that can measure liquid gravure ink spectral transmittance? I need a method to confirm ink recipe components in a liquid ink sample. I could use a faster option than the traditional draw-down proofing method. (885)
Many spectrophotometers are capable of measuring the transmittance of liquid samples placed in a glass cell. You can inquire with instrument manufacturers to get an instrument with appropriate capabilities (or find out if you already have one). The only question would be whether the inks are transparent enough to allow a measurement. If they are translucent (scatter light), then the measurements might prove difficult. (Back to top)


How many lights are metameric to white light reflected from a paper? (884)
There is no limit to the number of metameric lights that could be created. (Back to top)


How did the CIE 1931 color matching experiments evaluate the whole gamut when we know that the RGB color gamut is a subset of that of the human eye? In other words, we know there are colors that cannot be matched by the RGB lights. How did they get around this? (881)
It is really quite easy. RGB primaries are/were used and there are certainly many spectral colors (monochromatic lights) that cannot be matched by positive amounts of RGB. (The number depends on the selection of RGB.) The key here is the word positive. What is done is that the light to be matched is desaturated with one of the RGB primaries and then that desaturated light is matched using the other two. For example, 500nm is a saturated cyanish wavelength that is outside almost any RGB gamut. What is done is that it is mixed with a little bit of R and that mixture is matched by a combination of B and G as follows.

500nm + R matches G + B

Grassmann´s laws tell us that additive color matches follow simple algebra (that´s why the whole system works at all) and that allows us to subtract the amount of R from both sides of the match resulting in

500nm matches G + B - R

Thus, we can have negative amounts of light to match every possible color. That is why plots of the color matching functions in terms of RGB go negative at some wavelengths. (Note: For clarity I have left out the distinction between the amounts of RGB and the definition of them ... essentially, I left of the units.) (Back to top)


I measure timber colour using a Minolta Chromameter CR-400. I want to compare my values with those of other woods whose colors are listed in percent Luminance, Dominant wavelength (λ)and percent purity. Can I convert these values over to CIELAB? (879)
You are in luck. The computation is possible, but it is a little tricky. First, the % luminance is your CIE Y tristimulus value. Dominant wavelength and purity are directly related to the CIE xy chromaticity coordinates. Dominant wavelength is the wavelength at which a line from the white point (xy) through the sample (xy) intersects the spectrum locus on the chromaticity diagram. Purity is the fractional distance from the white point that the sample lies in the xy diagram. In other words, the dominant wavelength and white point define a line in the chromaticity diagram and the purity defines where on that line the sample falls. That would allow you to get xy chromaticities for your samples. Then together with Y (% luminance) you can convert to XYZ using the following.

X = Y(x/y)
Y = Y
Z = Y((1-x-y)/y)

With XYZ you can compute CIELAB coordinates. I would recommend you refer to a colorimetry text to get all the necessary details. Purity and dominant wavelength are discussed on p.61 of Berns Principles of Color Technology, 3rd Ed. You can also get all the details and a great reference book in Wyszecki & Stiles Color Science. (Back to top)


Is it possible to scientifically determine the original colors in an old black and white stereoview photograph if some of the colors are known? (878)
Unfortunately, no. You might be able to make some good guesses based on the objects in the photographs, but color is three dimensional and in a black and white photograph it is mapped into one dimension. So, for example, a light red and a dark blue might map to the same gray level in photograph and you have no way to tell which color the object was without some additional information from outside the photograph. And unfortunately, knowing about some objects does not tell you about others. (Back to top)


What wavelength(s) are used in the calculation of Gardner color? (876)
The Gardner Index is computed from CIE chromaticity coordinates (Yxy) and is therefore based on the entire visible spectrum of wavelengths from approximately 380nm to 760nm. (Back to top)


In his 1943 OSA paper "Specification of Small Chromaticity Differences" MacAdam describes constants g11, 2g12, g22 and gives values for them in graphical form. Is there a source for this data in conveniently tabulated form? (872)
The data required to compute the ellipses can be found in tabular form in Wyszecki & Stiles' "Color Science" book. They provide orientation angles and axes lengths rather than the gij coefficients, so it might not be exactly what you are looking for, but it is a commonly used source for this information. As you are using the ellipses, be sure to keep in mind that they represent data for only a single observer! (Back to top)


I need a device that will give me color readings at an observer angle of around 45 degrees or so that will pick up almost 100% specular reflection. (871)
You are correct that most color measurement instruments are designed to measure angles away from the specular reflection. There are gloss meters that measure the specular reflection, but they do not measure color. To do both you need a goniospectrophotometer that measures spectral reflectance at any combination of illumination and view angles. Murakami makes one such instrument and here is a link to some information from the US distributor. (Back to top)


Why is it not possible to get an accurate ΔE colour measurement on a metallic powder coating? (870)
It is possible, but the result is limited to the geometry of illumination and view present in the spectrophotometer. Metallic materials exhibit geometric dependencies in their color that are not easy to capture and represent with spectrophotometric measurements (and therefore ΔE*). Sometimes such materials are measured at several combinations of illumination and view angle to try to characterize these properties. The other potential limitation is that the metallic highlights have reflectance factors that significantly exceed those of normal reflecting materials and this results in CIELAB coordinates outside the normally well-understood range (e.g. L* can be much greater than 100). While such values can be computed and used in a ΔE* calculation, the meaning of such values remains very uncertain. (Back to top)


Do you know the origin of a measurement called the "flop index", which is defined as flop index=2.69(L15�-L110�)^1.11/((L45�)^0.86). My Xrite M68II reports this metric, however, X-Rite does not explain the source of this equation. Do you know where I might find the source of this equation, or another good way to mathematically explain the color difference at several angles? (865)
Unfortunately I cannot find a source for this, or any, flop index equation. It appears to be some form of empirical fit, so there must have been an experiment done somewhere. I would suggest you contact X-Rite to see if they can provide a source for the equation. This particular equation only depends on L* (measured at various angles), so it is only telling you about lightness differences. If you are interested in full color differences, you can compute normal CIELAB ΔE* color differences across the view angles. You might get some unusual values in comparison with color differences computed within a single geometry, but with care, the computed values could still be useful. (Back to top)


Are there any specific details in 45:0c reflectance measurement, such as the effect of distance between detector and sample or distance between light source and sample? Are there restrictions on the circumstances, like the color of the floor on which we put our sample? (860)
The specifications are in CIE Publication 15:2004. They include restrictions about illumination and view angle and solid angles, but nothing about distances. They also do not restrict the backing of the sample, although you might want to be careful of that if the sample is such that light might pass through. (Back to top)


Is there any approximation to convert between Hunter Lab (illuminant C, 10 degree observer) and CIELAB (illuminant D65, 10 degree observer)? (859)
There is no unique solution since there are many spectra that can produce a given set of colorimetric coordinates (metamerism). You could make a rough approximation by using the inverse HunterLab equations to convert back to XYZ tristimulus values and then compute the CIELAB values for them (perhaps just using Ill. C since it is close to D65 and this will be only an approximation). (Back to top)


I received data from two sources. The first measured Hunter Lab on light source C under 10 degrees, the second measured CIELAB on light source D65 under 10 degrees. I don't have the spectra. Is there any approximation with which I can convert between them? (858)
There is no unique solution since there are many spectra that can produce a given set of colorimetric coordinates (metamerism). You could make a rough approximation by using the inverse HunterLab equations to convert back to XYZ tristimulus values and then compute the CIELAB values for them (perhaps just using Ill. C since it is close to D65 and this will be only an approximation). (Back to top)


What observer for XYZ values can be used in CAM02, 2 degree or 10 degree? Some articles use 2-degree observer, the others use 10-degree observer. (857)
CIECAM02 was derived using the CIE 1931 2-deg. color matching functions, so they are the best choice. However, it will still work well with 10-deg. tristimulus values but the results will not be identical. (Back to top)


Would a spectrophotometer be able to accurately read the color differences between chrome plated surfaces? (ie. mirror-like finish) What settings or type of instrument would best achieve this? (853)
Yes, but it would take a special spectrophotometer designed to measure the specular (or regular) reflectance, which is the main component in a chrome surface. You could probably get a reasonable measurement from an total reflectance geometry (integrating sphere with specular component included), however their might be differences in the geometric distribution of the reflected light that would be masked in that measurement so you might also want to include an SCE measurement as a second confirmation. It is a challenging measurement, but with some care you might be successful. (Back to top)


Is it possible to quantitatively describe a color just by its reflectance or radiative spectrum? (852)
Perceived color depends on the reflectance of the object (or the radiance of a source), but it also depends on the illumination, the viewing geometry, and the response and adaptation of the human observer. The spectral properties alone are not enough to fully describe a color. (Back to top)


Is there a color model which would allow me to do the following? Given L*a*b*/XYZ/whatever coordinates of a color stimulus, given specified viewing conditions and color, what is the shape/size of the JND "ellipsoid" around the color, and what are the directions of the hue/chroma/lightness axes? (846)
The CIELAB color space would be your best bet as a starting point. The L*a*b* rectangular coordinates can be converted into L*C*h cylindrical coordinates to represent lightness, chroma, and hue angle respectively. Predicting the size and shape of JND ellipsoids is another matter that is a bit more complicated. You might want to look into the CIE2000 color difference equations to understand the relationships between color difference perception and appearance specification in CIELAB. (Back to top)


Is 0.05ppm a suitable tolerance for measuring absorption on a spectrometer for beverages? When is it no longer a visible difference to the naked eye 1ppm, 0.1ppm, 0.01ppm, etc? (837)
This would depend on what color the beverages are and what you are measuring. The only way to evaluate visual tolerances is with colorimetry, PPM on spectral absorptance functions tells you very little (perhaps nothing) about visual perception. (Back to top)


In the CIELAB color space, what is the approximate difference for each of L*, a*, b*, and ∇E that most people would not percieve as a different color. (831)
It depends significantly on the color in question and the viewing conditions and could range anywhere from less than 0.5 to as much as 3 or 4 units. The general rule of thumb is about 1 CIELAB unit. (Back to top)


Where can I find a chart showing color names, the colors themselves and CIELAB color coordinates? Ideally, this would also show the RGB color coordinates. I have found many RGB charts like this but none with the CIELAB coordinates. (829)
I don't know of any since it is a very difficult procedure to calibrate such a chart and RGB coordinates are generally undefined. I would suggest taking one of the RGB charts you have found and performing a conversion from sRGB values to CIELAB values (along with viewing them on a calibrated display). (Back to top)


I need to know the Munsell color for 'school bus yellow.' The unsourced quote I have says it is 9.4YR, 7.5/9.2. (824)
I'm sorry, I don't have the standard in terms of Munsell coordinates, but the Yxy values are 40.2, 0.4882, and 0.4205 respectively (From Color Standard SBMTC 008, 1997). Your Munsell coordinates seem about right, but I can tell from the Y value of 40.2 that the Munsell Value should be 7 instead of 7.5. The rest can be converted using the Munsell Conversion Freeware available from X-Rite. Unfortunately that won't run on my computer!

Update!
From The Physics and Chemistry of Color:The Fifteen Causes of Color. Second Edition by Kurt Nassau. John Wiley & Sons, Inc. 2001 p 395: "A full designation for an orange-yellow school bus in the sequence H V/C and using interpolation might then be 9.4YR 7.5/9.2." (Back to top)


Where can I purchase a poster size print of the CIE 1931 Chromaticty Diagram? I realize that the diagram is only a 2D projection of 3D matching data however I still find it useful to illustrate some of the other attributes of the data depicted (color mixing, color temperature, color gamuts)? (822)
Not only am I not aware, but I probably wouldn't tell you if I was!

Seriously, we really don't like them because they badly misrepresent the meaning of a chromaticity diagram. The diagram tells us nothing about color appearance since it is only based on matching data. It only tells us whether two colors match, but not what color they appear. (Back to top)


Is black a natural color or is it just a very dark brown that people just see as black? (820)
Technically, color is defined as a perception, so if something appears black to you, it is black. Beyond that, there are many naturally occurring materials that we would consider to be black (not just dark brown or other color). Obsidian is one thing that comes to mind (although it does occur in colors other than black as well). So, yes, black is a natural color. (Back to top)


We would like to know the complete Reflectance Measurement principle and formula for the dual beam color spectrophotometer including Light trap, PTFE standard and sample. (819)
I would suggest texts like Berns' "Principles of Color Technology, 3rd. Ed." for a fundamental introduction and Wyszecki and Stiles' "Color Science, 2nd Ed." for more details and references. (Back to top)


How would I measure the yellowness of the sclera (whites) of the eye? Also, what instrument would be an appropriate measure haziness of the cornea? (812)
The best instrument would probably be a spot spectroradiometer (also sometimes called a telespectroradiometer). You point it at the object of interest like a you would a camera and measure the indicated spot. This would allow you to choose the appropriate spot on the sclera (or cornea) to measure. With controlled lighting, or a reference white measured next to the eye, you could get quite reliable measurements. (Back to top)


Is 0.05ppm a suitable tolerance for measuring absorption on a spectrometer for beverages? At what concentration is the difference not perceptible? (807)
This would depend on what color the beverages are and what you are measuring. The only way to evaluate visual tolerances is with colorimetry, PPM on spectral absorptance functions tells you very little (perhaps nothing) about visual perception. (Back to top)


"The Desktop Color Handbook" mentions using a "light source color rendering card" which has patches that look the same under light of the appropriate CCT, but different under other lights. Do you know who makes such cards, or how one could create one? (803)
I have seen such cards, but I don't know of any current manufacturer. "color rendering" really isn't the correct name. These cards are based on a pair of metameric samples that will match each other under the appropriate lighting. They are produced by having the samples prepared with differing spectral reflectance functions, but designed to match under a certain light. For example, gray metameric pairs designed to match under a chosen light could be made on a CMYK printer by printing one sample with black ink only and the other with a combination of cyan, magenta, and yellow (but no black). (Back to top)


Is there an accepted universal standard for the calibration of color Spectrophotometer? Are there white standards available with certified reflectance values? (800)
The universal standard is the theoretical perfect reflecting diffuser (PRD). A PRD is a perfect reflector (reflects 100% of the incident light) and a perfect Lambertian diffuser. No such material actually exists, but pressed PTFE (sometimes called Halon) is a good approximation. Materials are calibrated by national standardizing laboratories to provide reflectance factor data relative to the PRD for a given illumination and viewing geometry. In the USA the national laboratory is NIST. Other examples are NPL in the UK, and NRCC in Canada. (Back to top)


Are there maximum/minimum values for a* and b* like the min and max for L*? (799)
No, there are not specific limits. The limits for given types of materials (or lights) would depend on both their material properties and the viewing conditions. You might look at the MacAdam Limits to see what some theoretical limits for reflecting objects look like. The a*b* limits depend on the hue and lightness of the object. In imaging applications, limits for 8-bit encoding are often set as the range from -127 to +128 in both a* and b*. This is adequate to cover the gamut of most systems. (Back to top)


Do you know where I can purchase Hess Ives Color standards, both physical standards and papers publications? (796)
I'm afraid I don't know of a source, but I would suggest contacting Tintometer since they make instruments capable of measuring Hess-Ives units. (Back to top)


Is there such an instrument as a hand-held spectrometer for measuring the color temperature of light sources? If so, how much do they cost? (794)
Yes, there are handheld colorimeters from companies like KonicaMinolta that are capable of such measurements. They range in price from several hundred to several thousand dollars. At the higher end, there are portable spectroradiometers from companies like PhotoResearch that are in the tens of thousands of dollars range, but also provide spectral radiance data in addition to color temperature. (Back to top)


Densitometers can be purchased with or without polarizing filters. Why are both types used? Which type provides the best correlation to perception? (793)
The polarizing filters help eliminate gloss in the measurement. This means less light is being measured and therefore density is higher. This measurement is used sometimes to allow the measurement of wet inks that are still glossier than the finished product, or to compare surfaces with different gloss characteristics. Observers tend to move the samples to minimize gloss when judging color. (Back to top)


How do you convert ΔL ΔC ΔH values for colour differences in CIELCh to the equivalent ΔL Δa Δb colour difference values in CIELab, when you do not have access to the absolute values for the original measurements? (792)
Unfortunately, you can not. (Back to top)


Are there some basic averages for the LRV of human skin? (791)
I don't know of any specific data and there would be quite a range, but an old photographers trick is to use the back of their hand to measure the light level if they don't have a standard gray card with them. This is equivalent to assuming that the skin has a reflectance of about 20% and that might well be a reasonable assumption on average. (Back to top)


I have used three online color space conversion calculators and I get three different sets of numbers when I convert from XYZ to RGB. What's happening? Who's right? (788)
It is quite likely that they are all accurate, however they all might be using slightly different definitions of RGB or assumptions on how the XYZ values were obtained. This is one problem with RGB; it is not specifically defined and is device dependent. It is probably best to just choose one converter and stick with it. (Back to top)


Roughly how large is the overall remaining perceptual non-uniformity of color tolerances in CIELab space, when using the ΔE94 and ΔE2000 color distance formulas? (as compared to the 6:1 non-uniformity of simple ΔE*ab)? (784)
It is a serious estimate, but if you are assuming 6:1 in CIELAB ΔE*ab, I would say that it would be about 3:1 or 2:1 in either ΔE94 or ΔE2000. Whether or not ΔE2000 is a significant improvement over ΔE94 is questionable. Both are significant improvements over ΔE*ab. (Back to top)


How do I calculate the dominant wavelength of a color sample with X,Y,Z or x,y? I do not want to draw it in the colorspace. (781)
I am not aware of any code to do this directly. You would have to do the computation in terms of xy. The task would be to define the line that connects the sample xy with the white point xy and then solve for the intercept with the spectrum locus (which is defined by the color matching functions and not analytically). It would have to be some form of numerical solution. There is also some information and suggestions from visionscience.com. (Back to top)


How to convert CIELAB data to Yellowness and Whiteness index? (779)
The equations are available in most textbooks on colorimetry, such as Berns'"Principles of Color Technology, 3rd Ed.". You will have to first convert the CIELAB coordinates back to CIE XYZ values and then compute the yellowness and whiteness indices you are interested in (there are several) from there. (Back to top)


What type of Pantone Guide shall we refer to & which instrument is suitable to measure the values of colour on aluminum foil with a flexo printing process? Since Aluminum has glossy as well as the matt finish then is there any need to change the settings of the instrument? (774)
I can't help you with the Pantone guide; I'd recommend contacting Pantone on that one.

On the measurement geometry, that is difficult with a metallic substrate. I suspect you are interested in the color of both the diffuse reflection and the specular reflection (especially if your ink is transparent). You really do need to make two measurements to evaluate both. I would suggest that you might use a traditional 45/0 geometry to evaluate the diffuse reflectance and a diffuse/0 (specular included) measurement to include the specular reflectance. It might also be insightful to look at the differences between the two measurements. As an alternative, you could use the specular included and specular excluded measurements on a single integrating-sphere instrument and compare those. (Back to top)


We print on Aluminum foil with a flexo process. What type of Pantone Guide shall we refer to & which instrument is suitable to measure the values of colour? Since Aluminum has glossy as well as the matt finish then is there any need to change the settings of the instrument? (773)
I can't help you with the Pantone guide; I'd recommend contacting Pantone on that one.

On the measurement geometry, that is difficult with a metallic substrate. I suspect you are interested in the color of both the diffuse reflection and the specular reflection (especially if your ink is transparent). You really do need to make two measurements to evaluate both. I would suggest that you might use a traditional 45/0 geometry to evaluate the diffuse reflectance and a diffuse/0 (specular included) measurement to include the specular reflectance. It might also be insightful to look at the differences between the two measurements. As an alternative, you could use the specular included and specular excluded measurements on a single integrating-sphere instrument and compare those. (Back to top)


How do I convert from spectral reflectance (from a hyperspectral imager) to CIELAB colorspace? (772)
You integrate the product of the reflectance, a selected illuminant, and the CIE color matching functions with proper normalization to obtain CIE XYZ tristimulus values. These values are then used in the computation of CIELAB coordinates. A colorimetry text such as Berns' "Principles of Color Technology, 3rd Ed." explains this procedure in detail. (Back to top)


Can a consumer grade flat bed scanner be used as a colorimeter? (771)
The general answer is no, but you can make a reasonable approximation if you know the characteristics of the material you are scanning. For example, if you are always scanning photographic prints that are made of the same three dyes, then it is possible to derive an accurate conversion from the scanner readings to XYZ values for that material. However, when you scan other materials, the conversion would not be accurate. (Back to top)


What are the colors of a complete spectrum? (Starting with the color with the longest wavelength.) (766)
The classic answer is ROYGBIV. That's shorthand for red, orange, yellow, green, blue, indigo, and violet. However, it really isn't that simple. When humans view the spectrum, they don't see just 7 discrete hues, but rather a continuous range of hues from red through violet. The number of distinct hues we can see would depend strongly on the specific viewing conditions, but it is safe to say that it could be well over 100! When naming colors, we tend to group them into categories of similar colors and that is why the ROYGBIV shorthand is so popular. (Back to top)


What color measurement system would be best to define colors for a liquid, to be described in a US patent? (765)
First, you would have to measure the spectral transmittance of a sample and define the measurement conditions such as the path length through the liquid that you are measuring. You could then do normal colorimetric computations with the measured spectral transmittance. You would have to select and appropriate illuminant and observer and then you could compute CIE XYZ and/or CIELAB coordinates to define the color. (Back to top)


What is the difference between the L*a*b* values and the L a b values when using a Minolta colorimeter? (763)
L*a*b* refers to CIELAB coordinates that are commonly used as the basis for color difference measurements. Lab refers to an older color space known as HunterLab. It was a predecessor of the CIELAB space and, unless someone specifically asks for HunterLab values, it shouldn't be used. The values are not equivalent. (Back to top)


My spectrodensitometer states its measuring geometry is 0/45, and its standard observers are 2 and 10. Are these angles related? Which should I use? (761)
0/45 refers to the geometry of illuminating the sample (0 degrees from the normal, or perpendicular) and detecting the reflected light (45 degrees from the normal). That is unrelated to the two and ten degree observers. Those refer to the field size when the original experiments were done to determine the standard observers for colorimetry. Think of them like have two different people look at the matches. Either observer would be appropriate for your application. It is best to be consistent, record which one you use, and use whatever is requested by your customers/suppliers or any standards you are following. More on all these topics can be found in books on colorimetry such as Berns' "Principles of Color Technology, 3rd Ed." (Back to top)


I am working on a plastic matched automotive black (GM 848) where the color is visually darker, but it measures LIGHTER than the standard. Can you please explain why? (759)
It's difficult to say without seeing the samples, but two possibilities come to mind. One is simply that there is some sort of problem with the measurements or the instrument. Assuming that is not the case (e.g., both samples measured at the same time, on the same instrument, in the same way), then I would suspect that there is a difference in the surface characteristics of the two samples. For example, the one that looks darker might have more of a glossy (or less matte) surface than the one that looks lighter. Depending on the instrument geometry you use, this difference might not show up in the measurements. If you take a glossy and matte surface that measure the same for total reflectance (0/d), the matte surface will usually look lighter. If you measure the two on a 0/45 instrument, the reading for the matte surface will be higher, more in line with the visual assessment. (Back to top)


What is the mathematical equation to convert Lab or XYZ to Density? (758)
I'm sorry, there isn't a direct conversion from CIELAB or XYZ to density. CIELAB and XYZ are based on visual color responses, while densitometers are designed to measure material properties (ink or dye amounts) of various systems. (Back to top)


We know density is derived from the transmitted spectral power distribution. Can you suggest a reference that provides the responsivity functions for Status A and Status M densities, as well as the specific function for the calculation of density? Also, what is the proper illuminant to use? (752)
The standard procedures for computing density are published by ISO (ISO 5 Photography - Density Measurements). It appears that part 3 (ISO 5-3) includes the spectral details. The procedures and data are also described in Hunt's book, "The Reproduction of Colour, 6th Ed." in chapter 14. (Back to top)


How can I meaure the color of clear solutions? We want an objective measurement (wavelength) of colors of various juices and wines. (749)
You would want to use a spectrophotometer or colorimeter capable of transmittance measurements using cells that can hold the liquids. There are a number of manufacturers of such instruments. Two that are well-known for these types of measurements are Hunterlab and Tintometer. (Back to top)


I am teaching color theory to our press operators and I am wondering if there is a full color spherical L*a*b model illustration that I can download, and if necessary purchase. (747)
If you do a google image search on "CIELAB", you will find a collection of images. I will leave it up to you to determine which have copyright restrictions. (Back to top)


Is there an existing paint industry standard system for expressing (and thus potentially matching) all the visual characteristics of a metallic or metallic flake paint, whether one-coat or two-coat (colored transparent over metallic or flake)? (744)
I am not sure if there are any industry standard procedures for these measurement and matching procedures. I would recommend checking with ASTM for the existence of such standards. However, there are commercial multi-angle instruments available for characterizing these materials. One example is the the X-Rite MA-68. You might contact X-rite to learn more about that instrument and any related standards and applications. (Back to top)


I am trying to calculate the colour of a coating. I have a graph of wavelenth versus reflectivity. Is there a simple way to convert this graph into RGB values or some similar standard? (743)
Unfortunately this is not quite so simple. RGB values are not standardized, so their meaning as a color specification is dubious at best. There are fairly straightforward colorimetric calculations that are performed to specify color matches (CIE XYZ) and color appearance (CIELAB) from spectral reflectance measurements. These might be helpful in your application. I would recommend a good introductory text on Colorimetry (such as Berns' "Principles of Color Technology, 3rd Ed.") to learn more about the details of these computations. (Back to top)


We have two items painted with Munsell N7 which are significantly different in shade. Is this normal? How can I measure if the differences are within specification? (735)
It is not unusual for materials that are supposed to have the same color to have perceptibly noticeable differences. Whether or not they are within specification depends on the tolerances set for a particular application. Such tolerances are normally established by setting acceptable lightness, chroma, and hue ranges in a color space such as CIELAB and using spectrophotometry to measure the samples. Alternatively, tolerances are sometimes setup visually. You could select the darkest and lightest samples that are acceptable to define the range that you would accept in your application. (Back to top)


Why isn't E point (x=0.3333 - y=0.3333) used as a standard illuminant for color matching and ΔE? (733)
It is used on occasion, but since it represents a purely theoretical illuminant (equal energy at each wavelength) it never occurs in a natural viewing environment. Therefore many prefer to use something like an average daylight distribution (such as Illuminant D65). (Back to top)


I am working on honeybee color vision. Using a honeybee specific color diagram, we know how to calculate dominant wavelength. How do I calculate the dominant wavelength of a color sample which resides in my color diagram in the "magenta" (or purple) zone. (732)
The way it is done in normal colorimetry is to draw a line from the color through the white point to the spectrum locus (rather than the other way) and then the wavelength of intersection is called a "complementary wavelength" or "complementary dominant wavelength". Thus you will end up with "green" complementary wavelengths for those "magenta" colors. (Back to top)


I work for an automotive finishing company and use an X-rite MA68II for measuring color. Which should I use, L*a*b*, or L*C*h�? Everyone seems to use L*a*b*, but I'm not sure why. (730)
The two represent different ways to represent the same CIELAB coordinates. L*a*b* are the rectangular coordinates and L*C*h are the cylindrical coordinates of the same space. So, in a sense, it doesn't matter which you use. L* is the same in both cases and correlates with our perception of lightness. Some people feel that L*C*h is more intuitive for describing colors and tolerances (C* representing chroma, and h representing hue) and choose to work in those coordinates. However, at times it is simpler to work in the rectangular coordinates where a* represents redness-greenness and b* represents yellowness-blueness. It really is a personal preference based on the application and user's experience, but normally L*C*h are a little bit more intuitive and effective (especially for tolerances). (Back to top)


I am developing paint color regulations to help buildings constructed on mountainsides blend into their surroundings to minimize their visual impact. What maximum LRV do you recommend for an environment typically consisting of darker browns, grays, and greens as the "base" colors in order to keep buildings from "standing out"? What would be the most appropriate instrument for measuring the LRVs of paint/stain for enforcement purposes? (721)
You would want the paint LRV to be similar to that of the background environment. In general this is taken to be about 20% for natural scenes. I would use that as a guideline and perhaps allow slightly higher values as well. Perceptually, a 20% reflectance appears to be midway between white and black, so this is a reasonable number. There are a variety of instruments that could perform this measurement, but a handheld spectrophotometer would be the most accurate, flexible, and easy to use choice. (Back to top)


How does the color of a room affect people's moods? (720)
Emotional responses to color can be very real, but they do vary in type and magnitude from person to person. Many times these responses are from learned associations. Faber Birren's book, "Color and Human Response" is a good source of well-grounded information on this topic. (Back to top)


What are the equations for McAdams ellipses? Are they location dependent in the 1931 chromiticity diagram? (719)
There are no equations for MacAdam's ellipses. They represent data on the variability of color matches for an observer when matching lights of various chromaticities and yes, they are location dependent in the 1931 xy diagram. I would suggest reading about this type of work in Wyszecki and Stiles' reference book "Color Science" and then exploring the original MacAdam papers on the topic if you need further details. (Back to top)


It makes sense that I should want window glass with the highest Color Rendering Index available in the performance category that I need; but how do I judge the relative CRI when all options are between 87 and 95? (718)
Color Rendering Index is computed for light sources, so I assume you are referring to daylight transmitted through the various glass samples. In your case, there isn't much to compare since all your CRI values appear quite high. If CRI is the most important variable in your application, then simply choose the highest. If there are other issues, such as cost, that would incline you to lean toward the lower CRI, then you won't be losing too much in this case. (Back to top)


Can CAM02 to predict Appearance Correlates of a color on different colored backgrounds? (716)
CIECAM02 is only capable of accounting for changes in the relative luminance of the background, not its color. To account for background color changes you would have to adjust the adaptation white point to be some weighted combination of the color of the light source and that of the background (essentially that is the same as incomplete adaptation to the background). (Back to top)


Is it possible to draw plankian line in CIE diagram? How? (714)
Yes, it is possible. The full procedure would be to use Planck's equation to compute spectral power distributions for a range of color temperatures and then compute the desired CIE coordinates from those spectral distributions. This is described in several places in Wyszecki and Stiles reference book, "Color Science". Conveniently, they also have a table of the xy chromaticities for a range of Planckian radiators in chapter 3 of that book. (Back to top)


Why does the sheen of a finish affect the appearance of a color? Do the wavelengths of color get magnified, refracted, or intensified as the pass through the layers of a paint film? (710)
The overall perception depends on both the diffuse reflectance of the material (what we often call color) and the surface properties (sheen or gloss, which is not colored). For a glossy material, all that surface reflection goes off in one direction and we can then see the nice saturated, or dark, color of the underlying material. As the material becomes more matte, more of that surface reflection is scattered in all directions and that results in the material looking lighter and less saturated no matter how it is viewed. Consider a black car and a white car both with equal levels of sheen on the final finish. The black car appears to be more shiny. Here, the sheen of the white is masked by all the diffuse reflection that makes the material look white. In the black material since there is really only surface reflection (the sheen), it is very apparent.

Generally for normal materials, there is no effect on the wavelength composition of the light. (Back to top)


Can you explain the difference between value and luminance? Is Light Reflectance Value different than luminance? (709)
LRV is related to the percent of light reflected and is proportional to luminance. Luminance would refer to the absolute amount of light, while LRV refers to the percentage that is reflected by the paint regardless of how much light is present. (Back to top)


Where can I buy Color Standard for measurement calibration? (705)
Measurement standards are general sold by national standards laboratories such as NIST in the USA, NRC in Canada, and NPL in the UK. There are also companies that sell more affordable secondary standards that are traceable to the national labs. If you do a Google search on terms like "reflectance standards", you can find a lot of information. (Back to top)


To measure process color is it better to use a spectrometery or density, hue, and gray? (704)
It is better to use spectrophotometry if you are interested in measuring the color appearance, but it might be better/simpler to measure density if you are only interested in the amount of ink present (and you don't measure different types of inks). Densitometers are designed to measure amounts of colorant (ink, dye, pigment, etc.) and not what the color looks like to humans. Spectrophotometers can be used to provide colorimetric parameters that correlate more closely with human perception. Thus, if you are measuring different sets of inks or different types of materials it is possible for density readings to match when the color doesn't. (Back to top)


Is there an index for the degree of yellowing? I want to characterize the degree of yellowing in trees as an indicator of poor health from aerial RGB photographs. (702)
There are yellowness scales that could be used. For example, there is a description on page 70 of Berns' "Principles of Color Technology, 3rd Ed." You could also look at color changes in a color space such as CIELAB (also described in the same book). The difficulty you will run into is calibration of your system to consistently convert from your RGB images to device-independent colorimetric coordinates. If the imaging system is not stable, or calibrated, you might have more variability in the RGB images than is present in the foliage of interest. Looking through Berns' text (or a similar text on colorimetry) will give you a sense of the issues involved. (Back to top)


Where can I buy the Munsell color chart that has the 14 samples used for CRI tests? (700)
Munsell products are sold by GretagMacbeth. I don't think there is a specific chart with all the color rendering samples, but it is possible to by individual Munsell samples. (Back to top)


Are there any instruments or methods which measure the intensity of gray color? (694)
Yes. These measurements are usually performed in the field of colorimetry using instruments called spectrophotometers. Spectrophotometers measure the percentage of light reflected by an object and then the results are used to compute colorimetric coordinates. In the case of grays one could use the Y tristimulus value to look at the relative luminance of the grays or the CIELAB L* parameter (which is a nonlinear transformation of Y) to have a number that correlates better with perceived lightness. All of this can be found in more detail in texts on colorimetry such as Berns' "Principles of Color Technology, 3rd Ed." (Back to top)


I am trying to establish a color system criteria for silicon parts which are white. What do you suggest in setting up a criteria for the color white? (687)
I'd suggest treating it like any other color. Choose a standard to aim for and characterize it colorimetrically (perhaps under multiple light sources) and then use your knowledge of the process and/or customers to set reasonable tolerances in CIELAB coordinates or using a color difference equation. There's nothing unique about white to make this procedure different than it would be for any other color. (Back to top)


What is meant by plankian locus? What is its importance? (686)
The Planckian locus is the curved line on a chromaticity diagram that represents the colors of Planckian, or blackbody, radiators. As a blackbody radiator is increased in color temperature, its color changes from red, through yellow, to blue and the Planckain locus maps out this change. These color temperatures are also used as descriptors of the colors of other types of light sources. I'd suggest you read about blackbody (Planckian) radiators, color temperature, and correlated color temperature in a text on colorimetry such as Berns' "Principles of Color Technology, 3rd Ed." (Back to top)


How do I calculate whiteness index (WI by ASTM standards) if I don't have a spectrophotometer but am in possession of reflectance data for materials that have <95% reflectance @ 600nm? (682)
I would recommend two texts to get more information on whiteness indices and colorimetric computations. These are Berns' "Principles of Color Technology, 3rd Ed." and Hunter and Harold's "Measurement of Appearance, 2nd Ed." The general form for computing a whiteness index is:
W = Y + 800(xn-x) + 1700(yn-y)
where Y is the sample relative luminance (Y tristimulus value), x and y are the sample chromaticitiy coordinates, and xn and yn are the white point chromatcities. The Yxy values can be computed from your spectral reflectance measurements using standard colorimetric procedures described in the texts. You might also want to refer to the original ASTM standard for more details. (Back to top)


I need a fairly fool-proof way for a moderately skilled technician to measure a large number of samples and present the data using the CIECAM02 coordinates. I am willing to buy a new instrument/software. Any thoughts? (678)
There are some examples of research code available on the internet. One option would be to adapt one of those to your needs and write the code in house. Since you are willing to buy a new instrument, it might be easier to convince the manufacturer to add the code to their system, if someone hasn't already. To my knowledge, vendors often add features to their software to make a sale. I'd suggest checking with several manufacturers to see if one hasn't already added CIECAM02, or would be willing to to get your business. (Back to top)


How can I get spectral reflectance information about Color-Aid samples? (664)
Unfortunately, I don't know of any source for these data. The best solution would be to buy a set of Color-Aid samples (very inexpensive) and then measure them with a spectrophotometer (unfortunately relatively expensive). Perhaps you can find someone that will allow you to use a spectrophotometer if you don't have one readily available. (Back to top)


hat is the correlation between Hunter whiteness (L, a , b) and CIE whiteness (L*, a*, b*)? Can I relate them mathematically? (651)
The HunterLab and CIELAB color spaces to which you are referring (they are more than whiteness metrics) are both computed from CIE XYZ tristimulus values for the sample and the illuminant. To go from one to the other, you would need to invert the first space to compute the sample XYZ values (assuming you know the illuminant values) and then use the normal forward equations for the second model. These could probably be concatenated into one set of equations, but I have never seen that and there would be little benefit. I would recommend Hunter and Harold's "Measurement of Appearance, 2nd Ed." as a good reference for details on these spaces. (Back to top)


I have a Minolta 2600d portable spectrophotometer. Can you tell me the best way to measure the fluorescence of teeth in vivo? Also, can I use the 2600d to measure the opalescence of both natural teeth and dental materials? (650)
The KonicaMinolta website doesn't give many details on that instruments and I am not personally familiar with it. However, the website does mention UV adjustment, so I assume it is possible to make measurements with and without UV energy in the illumination. Such measurements are one way to evaluate fluorescence. Any difference between the two indicates that the UV energy is being absorbed and re-emitted as visible light through fluorescence. As far as measuring the color of fluorescent materials, keep in mind that the measurement is only valid for the light source in the spectrophotometer. (The website doesn't say, but often these are D65 simulators.) I assume by opalescence, you mean changes in color with changes in illumination and view angle. There is little you can do with this particular instrument, but you could compare the specular component included and excluded measurements to see if there are any significant color changes across that change in geometry. (Back to top)


What are the calculations used to convert spectrophotometric spectral data to density, CIE XYZ, and RGB variables/values? (648)
All of these calculations share the same basic structure. You take the spectral data, weight it (multiply) by a set of responsivity functions, and then integrate to get the result. In the case of density, you do a logarithmic transformation after the integration and in the case of RGB, you have to do a nonlinear transformation (often) to account for display properties. The weighting functions are different in each case. The best way to learn about these calculations is through a good text such as Berns' "Principles of Color Technology, 3rd Ed." and Hunt's "The Reproduction of Colour, 6th Ed.". (Back to top)


How can I convert CIELCH to CIE 1931 x,y coordinates? (635)
The first step is to convert CIELAB L*C*h (there is no such thing as CIELCH) back to CIELAB L*a*b*. That is a simple coordinate transformation from the cylindrical coordinates (L*C*h) to the rectangular coordinates (L*a*b*). You will then need the CIE XYZ tristimulus values of the illuminant (which were used to compute CIELAB coordinates). With those, it is fairly straightforward to invert the CIELAB equations to compute the CIE XYZ values of your sample. From there you need to convert XYZ to xy in the normal way. All the necessary equations can be found in any good text on colorimetry such as Berns' "Principles of Color Technology, 3rd Ed." (Back to top)


What is the utility of converting tristimulus values to L*a*b* values? (634)
Converting to CIELAB (L*a*b*) allows you to better describe what the samples look like, their lightness, chroma, and hue, and to set tolerances for the production of those colors (i.e., measure color differences). It is another way to represent the same information, but it is a potentially more useful way. (Back to top)


Can my company's color references be accurately reproduced 10 years from now based on the tristimulus values we maintain ? (633)
Your references will be quite stable over 10 years. The CIE just celebrated the 75th anniversary of the 1931 system and it is still very widely used and will certainly be around for a good long time. No worries there. (Back to top)


What does the ratio b/a give or mean? Here a is negative and b positive. (630)
I assume you are asking about CIELAB a* and b*? The ratio itself has little meaning, but the inverse tangent of the ratio is called hue angle and is a good indication of the color's hue. When a* is negative and b* is positive, you are in a range of hues between yellow and bluish-green. (Back to top)


Can you help me transfer color temperature to 1931 CIE xy-coordinates? (627)
You would use Planck's equation to compute the spectral power distribution of a blackbody at 6500K and then use the CIE color matching functions to compute the chromaticity of that spectral power distribution. (Back to top)


I'm measuring several 3000K white lamps. I'd like to know the color variation of these lamps. Is there any other way to measure color variation of a population of lamps, aside from standard deviation and MacAdam ellipse? (623)
You can use the CIELAB color space and color difference equations for lamps, but you will have to make a selection of one to use as a reference white. Perhaps the ideal 3000K lamp could be your reference. You would then use its tristimulus values as the CIELAB white point and the other lamps as stimuli. You could then compute CIELAB color differences between each lamp and the reference, or between various lamps. If there are differences in absolute luminance this computation might get tricky, so I would encourage you to normalize all your lamps to a relative luminance of 100 (any number would do) before the CIELAB calculation. This will result in every lamp having an L* of 100 as well, but you can examine the color differences between the lamps by looking at the a* and b* differences that will remain. (Back to top)


What are good books to get detailed information on both the Munsell and CIE systems? I'm an artist and have a general education of both systems but would like to know indepth knowledge of each, including the math and reasoning behind the constructions of each system. (611)
A good starting point for the CIE system is Berns' "Principals of Color Technology, 3rd Ed." and it also includes some background on the Munsell system. Specifically on Munsell, there is "The New Munsell Student Set" by Long and Luke (which also includes an abridged set of samples) and you can still purchase "A Color Notation: An Illustrated System Defining All Colors and Their Relations" by Munsell and Farnum.

If you want to get deeper into the subject you might look at "Color Space and Its Division: Color Order from Antiquity to the Present" by Kuehni. (Back to top)


How is the location of an illuminant on the Planckian Locus and R-G Axes related to colour vision tests? (608)
The illuminant location defines the appearance of white in a given viewing condition. The appearance of samples in a color vision test is then determined relative to the adaptation point. Since these tests are designed to probe certain color discrimination properties of deficient observers, it can be important that they are used with the proper illumination. (Back to top)


What is the "straight line rule" when referring to the CIE 1931 Chromaticity Diagram? (607)
It refers to the fact that the chromaticity of additive mixtures of two colors fall on the straight line between the chromaticities of those colors. The exact location along that line is proportional to the relative luminance to each of the two colors in the mixture. (Back to top)


If one measures same color on various spectrophotometers the spectral profile is the same,but there are differences of 2-3 % in reflectance. Can you explain why? Which is the most correct reading? (605)
The most correct reading is that obtained by a national standardization laboratory on a reference instrument. Results from different instruments vary for many reasons. These include subtle differences in illumination and viewing geometry, sample area, sample presentation, calibration differences, and systematic errors in the instrument (e.g., wavelength, bandpass, photometric scale linearity, etc.). For fluorescent objects, the light source in the instrument will make a difference as well. (Back to top)


Can I use BCRA standard tile for calibration of spectrometer? Where can I buy it? (604)
Yes, they are available from the National Physical Laboratory in the UK. (Back to top)


What are the advantages and disadvanatges of chromaticity diagrams? (603)
The chromaticity diagram provides an abridged method to specify the hue and saturation of an unrelated stimulus or of metameric matches. That is really its only meaningful application. It has far more disadvantages such as having no representation of luminance, having no relationship to appearance, and having perceptually nonuniform spacing. I would recommend a review of the topic in Berns' "Principles of Color Technology, 3rd Ed" for more details. (Back to top)


Can you please define "dominant wavelength" and "wavelength of maximum absorbance"? Which should be used in the calculation of strength? (602)
Dominant wavelength is a colorimetric measurement from a chromaticity diagram. If you draw a straight line on a chromaticity diagram from the white point through the chromaticity of the sample to the spectrum locus, the wavelength at which this line intersects the spectrum locus is the dominant wavelength. The wavelength of maximum absorbance would be derived from simply finding the maximum on a spectral absorbance function. Dominant wavelength would have no correlation with strength and neither would the wavelength of maximum absorbance. However, the absorbance value at that wavelength might be useful in a calculation of strength. (Back to top)


Would you recommend a source for standards tile for calibration of spectometer? (600)
There are a series of National Measurement Laboratories around the world, some of which provide standards for spectrophotometry. For example NIST in the USA, NPL in the UK, and NRC in Canada provide these services. I am sure there are others as well. Also, some instrument manufacturers provide calibration services. (Back to top)


What is the meaning of NET profiler of a color spectrophotometer? (596)
Netprofiler is a commercial software application sold by GretagMacbeth.

Check the GretagMacbeth site for many more details. Briefly, this application communicates with GretagMacbeth via the internet. Information is exchanged which tries to improves the measurements and maintenance of your spectrophotometer. (Back to top)


When I measure delta L on low-gloss (matte finish) parts, I notice more variation between measurments than with higher gloss parts. Is there a scientific reason for this? Does the gloss play a roll? (595)
The measurements of the lower-gloss parts might be more sensitive to variations in measurement geometry. That would mean that differences in instrument geometry, sample presentation, sample flatness, etc. would be more likely to add variability to your measurements. (Back to top)


I evaluate llighting systems by measuring u'v' with a spectroradiometer. How can I calculate deltaE color difference from this data? (592)
It is a tricky calculation, but what I would recommend is normalizing the sources to the same luminance (Y=100) and then use Illuminant D50 (with Y=100) as the CIELAB reference point (XnYnZn). Then convert your u'v' measurements for other sources back to XYZ (scaled to Y=100) and use the test source as the stimulus (XYZ) in the CIELAB computation. You can then compute the delta-E between the reference D50 (L*=100, a*=0, b*=0) and your test source (which would be slightly different). This delta-E will be equivalent to the idea of illuminating side-by side white samples with the reference and test sources and assuming you adapt to the reference source. That is a completely valid and meaningful computation to do. (Back to top)


How is the Light Reflectivity Value (LRV) measured? Why do building codes require a certain percentage of LRV in the materials used? (591)
LRV is ill-defined, but the best metric appears to be the CIE Y tristimulus value (or the luminous reflectance). I am not familiar with the building codes, but I would suspect that it is an issue of energy conservation since lighter building materials would result in less energy required for achieving desired illumination levels and to cool the building interior. (Back to top)


What sorts of jobs would a color scientist do? (588)
Thanks for your inquiry. For one thing you made me realize we have to make a page on our website that gives examples of the types of jobs our graduates accept. We do have a page listing our sponsors and companies that students have gone to work for (there is a big overlap between the two).

Most of our graduates at both the M.S. and Ph.D. level end up going into industrial research and development positions. There are a variety of fields represented by the list I mentioned above. Recents grads have ended up working on things like printing/color reproduction of greeting cards at Hallmark, formulation of paints at Benjamin Moore, imaging systems at HP, cell phone cameras at Qualcomm, display systems at TI, electrophotographic printers at Xerox, imaging software/hardware at Apple, television quality at Philips and Samsung, etc. Many have gone on to help found technology companies or to very high level research positions at places like Kodak, HP, Microsoft, etc. (Back to top)


What is the relationship between the CIELAB a*-b* opponent axes and the perceptual unique hues? (587)
The unique hues do not directly line up with the a*-b* axes. One reason for this that we perceive different numbers of discriminable hues between each of the four unique hues (e.g. we see more purple hues between unique red and unique blue than greenish-yellow hues between unique yellow and unique green). In terms of CIELAB hue angle, the unique hues fall at approximately R:24 deg., Y:90 deg., G:162 deg., and B: 246 deg. Only unique yellow happens to line up with a CIELAB axis (positive b*). (Back to top)


I'm looking for a Circular Spherical Vision Spectro Colorimeter for analyzing luminance and other parameters (axis etc...) of reflective materials in different colors for road construction. Can you help? (569)
It sounds like you are looking for a 3D goniospectrophotometer. To my knowledge, there are none available commercially but rather research labs custom build instruments for their own use. There are 2D multi-angle instruments on the market that are used for applications like metallic automotive finishes. (2D means that the illumination and measurement angles are always in the same plane relative to the surface.) One example is from Datacolor, but there are others from manufacturers like GretagMacbeth and X-Rite. Perhaps one of those instruments would be helpful for your application. (Back to top)


I measure pigments from bird egg shells with a spectrophotometer CM-2600d, should I work with 100%UV or 0%UV? What about pigments from bills and legs of birds? (561)
The reason for the UV selection in spectrophotometers is for the evaluation of fluorescence (some materials absorb the UV and re-emit that energy at visible wavelengths). I am not sure if the materials you are interested in are fluorescent, but measuring with the UV included and excluded will give you that answer. If the results are the same, then there is no fluorescence present and it won't matter which you select for future measurements. If the results differ, you will have to decide which is more relevant to your work. I would think that including the UV would make most sense since these materials are normally viewed in daylight, which has plenty of UV. But, again, it only matters if the materials are fluorescent ... otherwise the results will be the same either way. (Back to top)


What is chromatic textile? What properties do this textile have? (558)
An achromatic color is a color devoid of hue (e.g., white, gray, black) and a chromatic color is a color with some appreciable amount of hue (e.g., different from gray). A chromatic textile would simply be a textile that has a hue. (Back to top)


How do I compute the correlated color temperature of an Lab/Luv color coordinate? (557)
You can't compute correlated color temperature from CIELAB or CIELUV coordinates. Correlated color temperature is a metric associated with light sources and is computed from tristimulus values (XYZ) or chromaticity coordinates (xy, or u'v'). CIELAB and CIELUV are computed by normalizing out the light source so those coordinates no longer have the information you need to compute correlated color temperature. (Back to top)


I'm trying to understand whether there is an optical/physical basis for this "chameleon effect" (the ability of the materials to blend with surrounding tooth structures to create naturally appearing restorations) and whether there may be way to measure it. (556)
This sounds like a fascinating color measurement question. I am not very familiar with the materials used or color measurement of teeth, but I can imagine a few things that might cause the "chameleon effect" you refer to. The first one would be scattering. It is possible that two materials have the same measured color and translucency, but one achieves it more through absorption and the other more through scattering. Whichever is more similar to the natural tooth material would blend in better under a variety of viewing conditions. Related to this could be index of refraction. The material with an index of refraction more closely matching the natural tooth would blend in better at the edges of the filling/repair. The third thing that comes to mind is metamerism. It is possible to produce the same color and translucency from different materials and end up with different spectral reflectance/transmittance properties. This could mean that a given material is a good match to the tooth under one viewing condition and a poor match under another. Difference materials could exhibit greater or lesser degrees of metamerism. (Back to top)


Can you help me calculate the color difference between my Macbeth Colorchecker and its image after capture and display on an LCD monitor? (553)
You need either an instrument to measure the XYZ values of the displayed chart (a colorimeter or spectroradiometer) or a characterization of the display that will allow you to convert the displayed RGB values to XYZ. You can then compute CIELAB values for both the original chart and the LCD display and from their calculate color differences using any of the recent CIE color difference equations. You might look in Berns' "Principles of Color Technology, 3rd Ed." for details on each of these steps. (Back to top)


I would like to purchase a grey scale color variation kit for comparing the change in color wash for Denims and Textiles. Can you recommend a source for this kit? (552)
I believe the kit you are referring to is produced and sold by the AATCC and can be directly ordered through: this site. They also have other related products, so you might want to look around their website beyond the single item mentioned. (Back to top)


Where can I find a tabulation of all pairs of distict spectral monochromatic lights that are perceived as achromatic? (550)
First of all, I assume you mean pairs of spectral lights that appear monochromatic when mixed together, in other words additive complementaries. I am not aware of any tabulated sets of complementary wavelengths. There are two reasons for this. One is that there is a large range of chromaticity coordinates that are perceived as achromatic due to chromatic adaptation in our visual system. Thus, the first step in this process is to decide which achromatic chromaticity is of interest. Once that decision is made, the complementary wavelengths are fairly easily computed from either the spectral chromaticities (a straight line through the achromatic point connects complementary wavelengths) or the color matching functions. (Back to top)


How do I calculate the chromaticity coordinates of a single wavelength of light? (546)
To compute the chromaticity coordinates for a monochromatic stimulus (a single wavelength, wl), simply read the values of the color matching functions (x-bar, y-bar, and z-bar) and do the following computation:

x = x-bar(wl)/(x-bar(wl)+y-bar(wl)+z-bar(wl))
y= y-bar(wl)/(x-bar(wl)+y-bar(wl)+z-bar(wl)) (Back to top)


What is the equation for plot a MacAdam ellipse? Also, what steps are the Macadam ellipses on the standard x,y chart and what do these "steps" mean? (537)
There are not equations to plot out the MacAdam ellipse data. Instead the data for each ellipse are tabulated in the original paper (D.L. MacAdam, Visual sensitivities to color differences in daylight, J. Opt. Soc. Am. 32, 247-274 (1942).). The steps refer to standard deviation of the color matches with 1 step meaning 1 standard deviation, etc. Please keep in mind that these data do not generalize for color discrimination of the population as a whole, but rather represent the experimental results for one observer in a particular viewing situation. (Back to top)


What is the best method to measure the opaque black color of a plasic material? (535)
The best method is to measure spectral reflectance with a high-quality benchtop spectrophotometer. The benchtop instruments will usually have larger measurement apertures, more powerful light sources, and more sensitive detection systems in comparison with handheld instruments. These features will help to make the measurement of very dark colors more accurate and precise. The measured spectral reflectance can then be used to compute CIELAB coordinates to describe the color of the material. (Back to top)


What's the RGB composition of RAL 7005 & RAL 1021? (531)
Unfortunately I cannot directly provide the RGB values you are looking for. For one thing RGB is is a device-dependent metric and the correct answer will vary from display to display. Secondly, the classic RAL designations are simply arbitrary color names with no published colorimetric coordinates (the system is proprietary and samples for each designation must be purchased). There is a newer RAL design series with designations that can be related to CIELAB coordinates thus enabling more flexible use of the numbers. There are a number of websites with example RAL colors, but I cannot confirm the accuracy of their representation or the methods they used to obtain RGB values for display. One such site, that includes the colors you mentioned, is martinindoor.com. That site might give you approximate RGB values that would be good enough for some applications. (Back to top)


How can I measure the LRV value of something. Or can I convert CIELAB data into an LRV value? (530)
LRV doesn't seem to be strictly defined (see this FAQ answer), but assuming that it is the CIE Y tristimulus value it can be easily calculated from the CIELAB L* value:
Y = Yn*((L*+16)/116)^3
Yn is the Y value for a perfect white under the same illumination and is normally 100. (Back to top)


What material has the most uniform reflectance spectra, including visible, IR, and UV? (529)
A material with the chemical name polytetrafluoroethylene (PTFE) that also goes by trade names of Halon and Spectralon is recognized as a very good and uniform (nonselective) reflector through the visible spectrum and into the ultraviolet and infrared wavelengths as well. It is sometimes mixed with carbon black to make a good, nonselective gray. As an example, Labsphere produces a selection of their Spectralon standards at various gray levels. They could provide more information on the reflectance properties outside the visible wavelengths. (Back to top)


Can you help me convert between Pantone and RAL colors? (527)
Both Pantone and RAL are proprietary systems. I have no way to convert between them and I am not aware of any. (Back to top)


What does the color blue mean? (517)
The color blue means whatever you want it to mean. A color by itself does not have meaning. However, within a culture and a context a color can develop a meaning. However, that meaning usually is only valid within the specific context (e.g. red usually means stop when it is a signal light, but not when it is the paint on a sports car). (Back to top)


When reading a violet, what does the Δa and Δb values represent? Example is the Δa dirty/clean or red/blue? (516)
With a violet color you are going to be in the positive a*, negative b* quadrant of the CIELAB space. While the precise answer will change with the precise color, the best general answer would be: +Δa* = more red, -Δa* = less red, +Δb* = less blue, and -Δb* = more blue. This is assuming the difference is computed as sample minus standard. (Back to top)


I am wanting to know how broad the primaries were for the 1931 color matching experiment. Were they monochromatic, as in a laser? If not, what was the spectral half width at half max for each of the colors? (513)
There wasn't a single experiment that resulted in the 1931 CIE color matching functions. There was one done by Wright using narrow-band primaries (essentially monochromatic light, but not lasers in the 1920s) and a second done by Guild using broad-band primaries (rough guess would be about 50nm full-width at half max). Wright performed the transformation of primaries necessary to mathematically represent his results in Guild's primaries and showed that the results were essentially identical. The CIE then took the averaged results and transformed them to two new sets of primaries, one a set of monochromatic RGB primaries to establish the seldom-used CIE RGB color matching functions and one a set of imaginary primaries that allowed for all-positive color matching functions resulting in the commonly-used (and equivalent) CIE XYZ color matching functions. (Back to top)


I would like to measure the color intensity of various parts of birds. What instruments will I use? (510)
This is a very challenging color measurement problem. Assuming the various areas are to be measured on live birds in their natural habitat, it is extremely difficult to make accurate measurements. The challenges include measurement duration (getting it done before the bird moves), variation in lighting, variation in illumination and view geometry, and issues about the size of the measurement spot. If all of these can be overcome, then a spot spectroradiometer (also called a telespectradiometer) would be the instrument of choice. However, it is probably unlikely that all these issues can be resolved to the degree necessary to meaningful measurements. A viable alternative would be to use a camera system. A high-quality digital SLR could be characterized to provide reasonably accurate colorimetric data for the hornbill features. The problems of illumination and geometry variations could then be overcome by capturing a large number of images and computing average colors for the features of interest. (Back to top)


Can you help me convert between candella and power wattage as measured by photodiode? (508)
You can't do this conversion unless your photodiode is filtered to have a photometric response or you are always measuring light sources with the same relative spectral power distributions. Just about any text on "Radiometry and Photometry" will provide the details necessary to understand these relationships. (Back to top)


What is the minimum ΔE that a human eye can detect? Which &Delta E" (495)
This is not at all a simple question with a single answer. As you note, the answer depends on the ΔE method or equation used, but it also depends on many other factors such as the viewing conditions for the samples (size, color, separation, luminance level, gloss, texture, background, etc.). As a very general rule of thumb, people take a simple CIELAB ΔE*ab of 1.0 to approximate a just noticeable difference. However, the CIELAB space is known to be non-uniform for color differences and it might take a significantly larger ΔE*ab for the difference to be noticed for a high chroma color. This non-uniformity is what prompted the development of weighted color difference formulas such as CIE94, CMC, and CIE2000. Since those weightings tend to reduce the average color differences, a rule of thumb for just noticeable color differences tends to be a bit smaller, closer to 0.5, when using those equations. In all cases, these are very general rules and shouldn't be taken too seriously. It is always best to derive a tolerance for the specific samples, colors, viewing conditions, application, etc. to get a reliable answer. (Back to top)


What's method that appropriate for measure color of gemstone? (490)
Measuring gemstones is not a trivial matter and some have constructed special instruments and sample holders for the purpose. It is reasonable to measure both transmittance and reflectance to get an idea of the gemstone color. The most perceptually meaningful measurement would be one carried out with more typical illumination and viewing conditions. This could be accomplished by arranging the desired illumination environment and then measuring the light from the gemstone at various angles using a telespectroradiometer. The Gemological Institute of America has developed measurement techniques and published on the topic in the past. Perhaps an inquiry to them or a search for their publications would be helpful. (Back to top)


Can you help me find a cost-effective way to measure black-gray-white and all shades in between. My son is doing an science experiment to determine how well erasers work. (485)
I can think of two techniques that might work. One would be to use a computer monitor and some graphics software to adjust a patch until it matches the sample and then record the RGB values (or other control ... some programs let you control a single value for light-dark) required for a match. The second would be to use a spinning disk. By making white and black disks that can be varied in proportion and spun to mix into gray, the gray level can be measured by recording the proportion of black and white. This web page gives an example for mixing red, green, and blue. The same can be done with just black and white. There are many more examples of spinning disks to mix colors on the web. (Back to top)


Specifications on the web given by SAE, ECE eg, mention that the colour should fall within the given boundary:
y = 0.335 (yellow boundary)
y = 0.980 - x (purple boundary)
What does this mean? (482)

Under typical viewing conditions red colors tend to plot in the lower-right side of the chromaticity diagram. So it is possible to figure this out given the color names you used (yellow, purple, red) and the numerical boundaries. The chromaticity coordinates for your red sample would need to be below the yellow boundary and to the right of the purple boundary. So a sample would pass if the following inequalities held true:

   y < 0.335 and x > 0.980 - y.

In your case y = 0.151 which is less than 0.335 so the sample is below the yellow boundary and x = 0.789 which is less than 0.829 (0.980 - 0.151) so your sample is to the left of the purple boundary (too purple). Therefore your sample fails to be within the defined red region. It would either need to have an increased y, and increased x, or some combination of the two to fall within the boundaries.

I'm not sure if the example you gave was just hypothetical, but those values fall outside the spectrum locus of the chromaticity diagram so cannot be a real object or light source. If they are intended to be real values, you might want to check for errors in your measurement or computation procedures. (Back to top)


Is there anyone who I can pay to convert 12 colors with LCh values into Lab or Munsell values? (481)
I could certainly put you in touch with someone, but the L*C*h to L*a*b* conversion is a very simple calculation given in this answer and the L*a*b* to Munsell conversion can be performed using GretagMacbeth's Munsell conversion freeware available from www.gretagmacbeth.com. (Back to top)


How can I make a colored chromaticity diagram? (476)
Since a chromaticity diagram does not represent the information necessary to define color appearance (but rather defines color matches ... and only a 2D projection of the 3D match data), we do not recommend making colored chromaticity diagrams. They are always incorrect and misleading. (Back to top)


What do the various CIE illuminant designation mean (D55, D65, etc)? How are they related to each other and to CCT? (472)
There is no simple correlation. On a clear day, the CCT of daylight will increase as the sun gets higher in the sky. However, other variables such as season, location, and weather conditions will also play a role. (Back to top)


Can you help me transform Lch values to Lab? (469)
CIELAB L*C*h and L*a*b* are two representations of the same color space. They are related by the transformation from cylindrical (L*C*h) to rectangular (L*a*b*) coordinates. The equations are:
L* = L*,
a* = C*cos(h),
b* = C*sin(h). (Back to top)


In literature it is mentioned that the difference between SCE and SCI color measurements gives an indication for the gloss of the surface. Which parameters are used to evaluate this: the difference between the SCI and SCE values for CIE-L, a, b or a combination? (464)
While a change in gloss would have an affect on all three CIELAB dimensions, the most significant (and consistent across colors) affect will be on the L* dimension. The glossier a sample, the larger the difference in L* between an SCI and SCE measurement will be. (Back to top)


What would be the reason to use CIELAB, CIELCH, CMC, CIE94, and CIE2000 (all of the above together) to calculate color difference from sample to standard? (463)
There really is no reason to use all of the color difference metrics together. The reason that many instruments/software systems provide all of those metrics is that different people are accustomed to using different metrics and the instrument manufacturers like to provide all of them. Generally in most applications you could settle on consistently using just one of the metrics. (Back to top)


What is the best way to test color fading in construction paper? My child is making an experiment on what color will fade the fastest when exposed to sunlight. (462)
There are instruments to measure color, spectrophotometers, that would provide a variety of data to measure the changes in color due to fading. You could also use a simpler instrument, like a photographic light meter, to measure the change in "lightness" of the paper. Even with the measurements there will be issues of interpretation. For example a 2% change in black paper might look far more objectionable than a 2% change in white paper. The PANTONE system is really a color naming system, so it doesn't provide a systematic way to measure changes and compare them across different colors. It might be best to do the evaluation visually and simply have some observers rank order the changes from smallest to largest after various periods of exposure to sunlight.

As far as preserving the colors with sunlight exposure, it is possible that some type of varnish or clear coating would help but it is hard to say and would require some experimentation. Even a piece of glass might help (e.g., the paper in direct sunlight might fade faster than that covered with glass since the glass will absorb some ultraviolet ... the same reason we don't get sunburned through windows). (Back to top)


Is there direct co-relation between CIE 76 and ANLAB 40 and ANLAB 50 color spaces? (459)
The CIELAB equations can be thought of as an evolution from the ANLAB color space. However, there isn't a direct mapping between them. One could always invert one space back into XYZ tristimulus values and then compute the other and it might be possible to derive a set of equations for that whole process if that is really necessary. (Back to top)


What is the primary calibration of a spectrophotometer? (458)
The primary reference standard for spectrophotometry is the perfect reflecting diffuse (PRD). A PRD is a theoretical material that is perfectly diffuse (Lambertion) and reflects perfectly (100% of all incident light). Properly calibrated spectrophotometers report values relative to the PRD. (Back to top)


Can you derive CIELAB values directly by visually inspecting the reflectance curve? (454)
With experience, one can learn to estimate CIELAB values for various illumination conditions by viewing spectral reflectance curves. However, this could only be very approximate. The actual values are computed from the spectral reflectance curves multiplied by a standard illuminant and weighted by the color matching functions. Finally a nonlinear transform is applied to obtain CIELAB values. While there is a direct computational relationship, it is not easily apparent from visual inspection of the reflectance curves and there is certainly no way to visually derive accurate CIELAB values from the curves. (Back to top)


Can you explain the mechanism of UV filter in color spectrophotometer, to get universal Whiteness values on all spectros. Is it a wedge filter or an attenuator filter. (453)
Generally the UV filter in a spectrophotometer is a cut-off filter that eliminates ultraviolet radiation incident on the sample below a certain wavelength (approximately 380nm). When the filter is in place, no UV radiation is incident on the sample and when the filter is removed, whatever UV is in the instrument's light source strikes the sample. For non-fluorescent materials, the resulting spectral reflectance measurements will be identical with or without the filter in place. For fluorescent materials, the measured reflectance factor (and therefore computed whiteness for white samples) will be higher when the UV energy is allowed through the system (no filter in place). Essentially, the UV filter allows a measure of the sample with and without fluorescent emission (approximately, since the filter might not remove all fluorescence, but just attenuate it). (Back to top)


I work in printing. Can you tell me what the effect of dot gain is on delta E? Also, can we get an accurate delta E measurement from a coated sheet to an uncoated sheet? (450)
Dot gain has an effect on the color of the print and this will effect the measured Delta-E, but there is no direct relationship between the two. The measurement of Delta-E is independent of whether or not there is dot gain. If the dot gain happens to bring the printed color closer to the proof, then the measured Delta-E will be smaller (and vice versa). You can accurately measure Delta-E across changes in paper surface. However the perception of the color differences will depend on the viewing geometry (due to the gloss differences) and that might effect the correlation between visual assessment and computed Delta-E. It is important that the measurement geometries (location of light source and detector) and the viewing geometries (location of light source and observer) are similar. (Back to top)


How do I figure out what color a reflection of another color might be? For example, if I have a yellow car and something red is reflecting off of it, what color will show up as the reflection on the car? (447)
The process of reflection is properly modeled with a mathematical multiplication in this case. If you know the spectral power distribution of the red light striking your yellow car, then the resulting light would be the product of that distribution with the spectral reflectance of the car. In color terms this would work like subtractive color mixing. Your yellow car is absorbing blue light and the red light is already missing green and blue. Thus when the red light strikes the yellow car, no more light will be removed and the reflected light will be red (approximately since it will depend on details of both the red light and yellow car). (Back to top)


What effect renders UV lamp when selecting the colour with use light booth? (446)
Generally the UV lamp in a light booth is used to check for the presence of fluorescence in a sample. If the sample is viewed under the UV lamp (only) and appears to glow in some color other than a deep violet, then it is fluorescent (absorbing the UV and emitting at different wavelengths). The UV lamp is also used with the daylight illumination in the booth to enhance the UV content, making it more like real daylight, to better render the appearance of fluorescent materials in daylight. It won't make any difference for non-fluorescent samples. (Back to top)


When visually comparing colors, why is D65 light source better to use than Illuminant C? (441)
A D65 simulator isn't necessarily better than an illuminant C simulator. For evaluating fluorescent materials, D65 might be better since it has UV content that more closely represents real daylight. However, you would need to be sure that the D65 simulator you are using truly has that UV content. The best choice for visual evaluation is always the light source that most closely approximates the standard illuminant data used in your colorimetric computations. So if your colorimetry (XYZ, L*a*b*, etc.) is computed using illuminant D65, then view samples under a D65 simulator. If you only have an illuminant C simulator to view samples, then it would probably be best to do your colorimetry with illuminant C. (Back to top)


We are looking at colour and redness of scars on human skin. Which is better for the job, a hand held spectrophotometer or a hand held chromameter? Some literature validates the use of a chromameter CR300 in skin research, but spectrophotometers are apparently more accurate. (440)
You are correct that spectrophotometers are generally more accurate and flexible in the colorimetric values they report. However, since you will be measuring a fairly limited range of colors, a colorimeter will likely provide you with sufficient data and accuracy. So, a spectrophotometer would be better for the job, but the colorimeter might be adequate if you are looking to spend a bit less on an instrument. (Back to top)


I can get x,y,z color coordinates. So how would one relate to the materials response for a particular wavelength? (438)
First of all, given the values in your question, I think you are referring to CIE tristimulus values, XYZ (upper case), and not CIE chromaticity coordinates, xyz (lower case). The values mean very little without more information. For example you need to know the illuminant, standard observer, and measurement geometry used in their computation to define the numbers. With that information (or similar information if it is a self-luminous stimulus) you could compute CIELAB coordinates that would give you some indication of the color appearance of the stimulus. However, you have asked a more complicated question. You cannot uniquely define the spectral properties (usually reflectance) of a material from its tristimulus values. This is really the fundamental definition of metamerism. There are many spectral curves that would map to the same sets of tristimulus values. You can't go backwards in the computation. There is an exception to this when you have a limited number of colorants with which to produce the stimulus. In such cases you might be able to find a one-to-one mapping between tristimulus values and spectral reflectance. (Back to top)


How can I calculate hue angle from CIELAB? What is pi? (437)
Hue angle is computed as the angle between the line that connects the a*b* origin with the stimulus a*b* values and the positive a* axis. It is computed as the inverse tangent of b*/a*. Since most programming languages will return the computed angle in radians, it then becomes necessary to convert to degrees. The conversion from radians to degrees is to multiply the angle in radians by 180/pi. pi is the numerical value of the ratio of the circumference of a circle to its diameter (approximately 3.14159).

Another way to look at this as a conversion from rectangular coordinates L*a*b* to cylindrical coordinates L*C*h (lightness, chroma, hue angle) in the CIELAB color space. (Back to top)


What is the RAL number that should be used in the background of a light cabinet? (435)
I don't have any way to convert from colorimetric quantities into RAL numbers. However, the most common color for light booths is a neutral gray with Munsell Value 7 (a light gray). In CIELAB that would translate into an L* of 70 and C* of 0 (hue is undefined for neutrals). You might be able to index into some RAL notations using the CIELAB values. (Back to top)


We are a printing company. We print offset (litho), gravure and flexo, both on coated and uncoated paper. What spectrophotometer should we buy spectrometer for Quality Control and Color matching? (432)
0/45 spectrophotometers are essentially the "standard" in the printing industry. There are probably two main reasons for this. First, the 0/45 geometry is best for comparing the colors of materials with various levels of gloss (common with different paper types) and secondly, the standards for densitometers in the printing industry also use the 0/45 geometry. In your case, it seems like the clear choice is to go with 0/45 instrument. (Back to top)


I need to measure human skin color directly. What instrument should I use? (425)
There are a number of small, hand-held spectrophotometers available on the market that would be very capable of making the measurements you need and reasonably priced. To find the best instrument for you and your budget, you would need to contact the manufacturers. Check online with KonicaMinolta, X-Rite, Hunterlab, BYK-Gardner, Datacolor, and others. (Back to top)


How do yo measure ΔE*cmc, ΔE*94, ΔE*fmc2 from the L*, a*, b*, C*, h* ΔH*, ΔE*? (423)
CMC and CIE ΔE94 color difference values are computed from the CIELAB values you mentioned. The FMC2 color difference equation pre-dates the CIELAB color space and is based on earlier CIE colorimetric coordinates. I would recommend the texts "Principles of Color Technology, 3rd Ed." by Berns and "The Measurement of Appearance, 2nd Ed." by Hunter and Harold as good places to learn more about the history of these equations and the equations used for computation. (Back to top)


When measuring the color of colored plastic plaques with a spectrophotometer, how thick should the plaques be and why? (420)
Ideally you would like the plaques to be thick enough that they are completely opaque (no light passes through them). This way it is only the plaque itself that influences the color measurement and not whatever happens to be behind it. If it is not possible to use opaque samples, then be sure to always be consistent in how the samples are backed (e.g. with a black object) when measurements are made. (Back to top)


I am doing a problem on image permanence for large format inkjet. Should I use Delta E or Delta E 2000? Why and what is the difference between the two? (416)
Using either the simple CIELAB Delta-E or CIEDE2000, would be appropriate for your work. The CIEDE2000 equations incorporate some corrections for nonuniformity in the perception of color differences across the CIELAB space. Thus, if it is important for you to compare the magnitude of a color differences for widely varying colors, CIEDE2000 might be a better choice. However, if you are simply looking at quantifying change in the samples and you will always be comparing like colors (e.g. red changed by so much for one technology and a different amount for another technology) then the simpler Delta-E equations will be more than adequate. You really can't go wrong with either. If I were to choose, I would use the simple Delta-E equations for this application since the added complexity of the CIEDE2000 equations are not likely to provide any additional insight. (Back to top)


We always measure the color difference by using d/8 spectrophotometer with included specular reflectance. But why do we always measure the white with excluded one ? (412)
I don't know. Normally an instrument will have calibration values for the standard for both specular included and excluded. A more normal procedure is to set up the instrument in the desired configuration and then calibrate with the tile. It might be possible that your instrument can be calibrated in one geometry and used for measurements in the other, but I would encourage you to use the same setup for both the calibration and measurements if that is possible. (Back to top)


I have an emission spectrum from a sheet of luminescent material. Can you help convert the emission spectrum to an RGB color to display? (411)
The easiest way is to first convert your spectral curve into CIE XYZ tristimulus values (see texts such as Berns' "Principles of Color Technology, 3rd Ed." for details) and then convert those tristimulus values to default RGB values such as sRGB (see w3.org). You could do better by having a specific calibration for your display to convert XYZ to accurate RGB values for your display, but the sRGB values would be a simple, and commonly used, approximation. (Back to top)


What is the best color model for use with colors that have a low saturation? CIELAB is not working for me because when a* and/or b* are close to zero I get poor values for hue. (409)
I assume that by poor hue angles you are suggesting that the CIELAB hue angles do not match the your perceived hue angle. I am not familiar with any deficiencies win the CIELAB equations that cause systematic problems with hue angles for near neutrals. There are two difficulties that you might be running into. One is that the actual neutral for your viewing conditions is slightly different from that computed by CIELAB. A small shift will produce very large discrepancies in hue angle (as much as 180 degrees). Another might be observer metamerism. This is the difference between your visual response and that of an average observer represented by the CIE standard colorimetric observer. That would also result in significant sensitivity for near neutrals. I would suggest sticking with CIELAB, but to use the a*b* dimensions to evaluate the relative redness-greenness and yellowness-blueness of your colors rather than using the C*h dimensions. If you see a systematic trend that your perceived neutral is always at a given set of a*-b* values, you could subtract those values from the other a*-b* pairs prior to computing chroma and hue angle to make an adjustment for your purposes. (Back to top)


I have to purchase a system to compare shades of yellow in plastics with rough surfaces. Would you please recommend a company and/or instrument that would have good repeatability and be low maintenance (407)
There are a number of manufacturers of color measurement instrumentation that can provide instruments capable of making the measurements you describe. We don't have the data, or the access to all current instruments, that would be required to make a recommendation of a particular manufacturer or model. (Back to top)


Do the tiles in the Farnsworth Munsell 100 Hue test kit require calibration? If so, what can change that would require their recalibration? If they require calibration, what is the frequency? (401)
To my knowledge, and in my experience with our F-M 100 Hue tests, they are very stable with respect to color change. Our very old tests appear to provide results consistent with out new tests. (That said, note that the manufacturer of these tests, GretagMacbeth, certifies the caps for a period of two years.) The color chips could fade over several years if left exposed to light, but that would be rare for F-M 100-Hue chips since they come in boxes, and should be stored there when not in use. More likely of concern than fading of the colors would be simple wear and tear damage from the handling of the chips. Sometimes they get dirty or damaged by finger prints. When such damage becomes noticeable, it is likely time to get a new test. The tests, and more information on their stability, are available commercially from GretagMacbeth. They produce and sell all Munsell products and are not associated with our laboratory (we're just named after the same person). (Back to top)


Does there exist a (known) correlation between a Minolta CM3500D and a BYK Gardner or does this depend on the product type? (400)
Theoretically, instruments of the same measurement geometry from different manufacturers should provide similar results. However there remain differences in the details of the optical configuration and calibration procedures that will likely result in differences in measured values. In general, the best inter-instrument agreement is obtained by using instruments of the same manufacturer and model. Obviously, this is not always practical. There are some techniques (described in Berns' "Principles of Color Technology, 3rd Ed.", Chapter 3) to evaluate the accuracy of instruments and then correct measurements from various devices to improve agreement. (Back to top)


CIE chromaticity diagrams are usually exhibited in a shape of a hoof. Why are the single waves spread on the ellipse side instead of the straight? (396)
The shape of the chromaticity diagram is an artifact of the shapes of the XYZ color matching functions and the projective transformation to xy. The curved part represents the gradually changing responsivities of the three cone types in the visual system as we sweep through the spectrum. It is curved due to the overlapping sensitivities. Near the long-wavelength end of the spectrum it becomes a straight line since only two visual responses are active for those wavelengths and not three. The purple-line (the straight line connecting the red and blue end of the spectrum) is simply derived from the additive mixture of the two ends of the spectrum, which is visible and has to fall on a straight line due to the laws of additive color mixture (Grassmann's Laws). (Back to top)


Can I determine the color rendering index of this source purely from the light-emission spectrum? (390)
Yes. You can compute the CIE color rendering index from the emission spectrum of your source. The technique is described in CIE Publications 13.3 and 15:2004. (See these CIE pages for more information. (Back to top)


Can you provide an example showing how to calculate ΔE using Δa*, Δb*, ΔL*, ΔC*, Δh for textile pigment? (385)
A color difference, ΔE*ab is computed as the Euclidean distance between the CIELAB coordinates. Mathematically, that is ΔE*ab = (ΔL*^2 + Δa*^2 + Δb*^2)^(1/2). I would recommend a basic colorimetry book such as Berns' "Principles of Color Technology, 3rd Ed." for more details. (Back to top)


Can you help me compare RAL specification paint with ANSI paint for color match? (376)
The most direct, and accurate, way to compare the two specifications is to obtain a sample of each paint and have it measured colorimetrically. You could then compute a color difference, such as a CIELAB ΔE* to quantify the magnitude of the difference. (Back to top)


Please can you explain what Macadam limits are (as mentioned in description of the 'real' set of Munsell values on your Munsell Renotation Data page)? (375)
The MacAdam limits are a definition of a theoretical maximal color gamut for object colors. In fact the MacAdam limits are far beyond any realizable object colors since they were computed, by David MacAdam, from theoretical reflectance curves that were equal to either zero or 100% at every wavelength (with a maximum of 2 transitions). Any color designations outside the MacAdam limits cannot possible be created with non-fluorescent object colors. Thus they provide a way to limit color spaces, or samplings of color spaces, to colors that might be realizable as objects. (Note that light sources can have colors that exceed these limits.) (Back to top)


I am a dental student. As part of my project I need to convert the RBG values obtained from Adobe Photoshop into CIELAB values in order to compare with values obtained with a spectrophotometer? (373)
Performing this conversion accurately requires knowledge of the specific RGB specification used. You can get a reasonable conversion in Adobe Photoshop itself (which will depend on the color settings of Photoshop to define the exact transform). Simply take an image in with the RGB values and change mode to LAB. (Under Image->Mode->Lab color.) That will give you CIELAB coordinates. You can also find both sets of coordinates in the Photoshop color picker. Just type in the RGB values and you will see equivalent LAB values for the current color settings. (Back to top)


Can you explain how to select the color center (or sample color) which can be used to evaluate color difference formula? (372)
Various researchers use different strategies to determine the color centers. The most important objective is to be sure you have a variety of colors that reasonably sample the range of color space you are interested in. There are also 5 color centers that were defined by the CIE for coordinated research. You can read more about them in A.R. Robertson, "CIE guidelines for coordinated research on colour-difference evaluation," Color Research and Application 3, 149-151 (1978). (Back to top)


As we know black colours, for example on textiles, usually have a pale tint. Is there any formula for measuring blackness? (368)
I do not know of any specific blackness indices. I would recommend simply using the CIELAB L* metric to quantify how dark a black sample is. You can also use the C* metric to quantify how close to neutral the sample is. (Back to top)


Would CIECAM02 or CIEDE2000 be better at predicting spot color differences? (363)
Most of the research on predicting color differences has been aimed at optimizing equations in the CIELAB color space (such as the DE94 and DE2000 equations). Similar work is being done within the CIECAM02 color space, but there is not yet a recommended technique for performing the computations. There is certainly reason to believe that one will be developed in the future. For now, and for the application you mention, the CIEDE2000 equations would be the recommended technique. For a simpler equation, that is likely to perform just about as well, you might want to consider the CIE DE94 equations. (Back to top)


Can mixed paint be measured while it is still wet? In our factory we mix paint, we want to measure it's color before painting parts. (361)
Certainly paint can be measured wet (preferably with a non-contact instrument). However, for most paint the color will change significantly as the paint dries. Thus you will either have to learn how the colors change and compensate for that when you make wet measurements, or simply wait for samples to dry before measuring them. The best way to evaluate the color of finished paint is to create a sample as close as possible to the intended application and then measure it after the paint has dried. (Back to top)


I am trying to describe the colour of my sediment cores and have L*, a*, b*, and C* data from a spectrophotometer. I am looking for some directions on how to 'describe' the colour of my mud! Is there some tool or guide to convert numbers into a colour category? (359)
The CIELAB coordinates correspond approximately with color appearance descriptions so it is possible to make some assessment of categories directly from the values. L* is lightness with 0 being black, 100 being white and 50 being a perceptually middle gray. The a* dimension is redness-greenness with positive values indicating redness and negative greenness. Likewise for b*, which is yellowness (positive) and blueness (negative). If the a* and b* values are both zero, you have a neutral gray. The C* metric is chroma, which is simply a measure of the distance from the origin in the a*-b* plane (C* = 0 for gray). The magnitude of the C*, a*, and b* values is similar to the L* dimension. With some experience, it is not too difficult to learn to correlate CIELAB values with appearance.

Another option is to convert the CIELAB values into Munsell coordinates, which are direct descriptors of lightness, chroma, and hue. Munsell descriptors are also commonly used in soil science., so your mud would fit right in. GretagMacbeth provides free software for these conversions. (Back to top)


How does one measure color of ink in its raw liquid form? I need to measure the color in that form. What tools are required? Are there standardised methods to sample ink to measure it? (353)
I guess I would ask what meaning the color of ink in raw liquid form has? Does anyone really care what it looks like then? It makes much more sense to measure the ink in a way similar to how it will be used and viewed. There are probably standard methods for preparing and measuring ink samples. Unfortunately I am not aware of details. I would suggest you look for ASTM standards on the topic or make an inquiry with an ink manufacturer to find out about standard techniques. (Back to top)


Each of our spectrophotometers has its own calibration standard. If all of the spectros are not calibrated to the same standard, aren't we calibrating in variation between the spectros? We can see anywhere from a .2 to over a .9 difference in our ΔE readings when we read the same areas of a sample on the different spectros. What would happen if we lost one of the standards since they are suppose to be unique to the manufacture of one specific specto? (351)
You are correct that you might be calibrating in variation between your instruments. However, they will each have different data for the calibration standard to compensate for some of that. If you were to decide to use one standard for all your instruments, then you would also need to make sure they were all using the same internal calibration data for that standard. Such a practice might well help you to obtain better inter-instrument agreement.

If you were to lose a standard you could get a replacement from the manufacturer and they could update the calibration data for the new standard.

Unfortunately, the color differences you have observed between various instruments are not unusual. (Back to top)


I need a means to estimate the difference in general light reflection between walls in a windowless room painted matt magnolia (Munsell 10YR9/2), compared to matt brilliant white. (344)
Your magnolia is very light with a Munsell Value of 9. I would guess that a good white would also have about the same Munsell Value, so the two would be about equivalent. (With a Chroma of 2, your magnolia color is pretty close to white anyway). Munsell Value 9 is approximately a luminous reflectance of about 76% percent if that might help in your calculations. (Back to top)


What is the Reflectivity Index and what is the index for various colours, particularly silver? (343)
"Reflectivity Index" is not a term commonly used in color science. It probably refers to the luminous reflectance factor, or CIE Y tristimulus value. The value itself depends on the measurement geometry and can range from 0-1 (or 0 -100%). A good silver surface, measured in a total reflectance geometry, might have a reflectance factor as high as 0.98 (98%) or so. (Back to top)


When I transform D65 illuminant values (in CIELAB) to RGB, I do not get R=G=B. Why? (339)
The CIELAB L*a*b* values for an illuminant are always 100, 0, 0. Whether or not a given white point (e.g., D65) transforms to R=G=B linear values depends on the definition of the RGB system. Some RGB systems are defined with D65 white points (meaning that, by definition, D65 maps to R=G=B) and others are not. White points do lie at the origin in CIELAB because CIELAB is normalized to the white point. This is essentially an assumption that chromatic adaptation is complete and the white point appears .... "white." (Back to top)


Various color systems (NTSC, SMTPE, CIE) use different white points. How are these defined? (338)
The CIE system does not have a fixed white point. White is defined by the user for a given situation and application. It is not fixed to any one chromaticity coordinate. The SMPTE and NTSC systems you mentioned are examples of specific applications in which a given white point has been selected. Both CIE illuminants D65 and C represent average daylight (D65 being the newer standard) and they were selected for TV systems since this is a good whitepoint to represent many typical scenes and for displays that appear neutral under a wide range of viewing conditions. (Back to top)


Can you provide information on Ultralume U5000? Also, what are ΔE GE, ΔE Audi,and ΔE PQ?) (336)
The only reference to Ultralume U5000 I can find on the internet is in brochures about the KonicaMinolta color control software. Perhaps someone there could let you know exactly what the light source was. Ultralume sources are usually narrow-band fluorescent sources and I would guess that U5000 might refer to a 5000K correlated color temperature.

Likewise, the color difference metrics you mention appear to be customer specific. Sometimes instrument manufacturers will add specific equations to their software for certain large customers and then just leave it in for everyone in case they want to use them. I would assume that GE refers to some General Electric specific equation, Audi to the automobile manufacturer, and PQ ... I have no idea. They are not CIE color difference equations. (Back to top)


What are these devices: densitometer, spectrophotometer, and colorimeter? (334)
A spectrophotometer is the most fundamental of the three instruments you mentioned. Spectrophotometers measure reflectance or transmittance as a function of wavelength. From those data, one can compute other quantities such as colorimetry (XYZ, CIELAB, etc.) or densitometry (Status A, Status M, etc.). A colorimeter directly measures colorimetric quantities through filtration of the sensors. Thus a colorimeter directly reports XYZ or CIELAB values, but spectral data are not available for additional computations. Likewise, a densitometer reports integrated values, called densities. There are various standards for densitometers, but they are not designed to measure color. By that I mean reported densities are normally not intended to correlate with human color perception, but rather to act as measurements of the quantities of dye or ink present. (Of course, with added knowledge about the system it is sometimes possible to convert densities into colorimetric quantities.) Again, densitometers do not provide spectral data that could be used for further computations. (Back to top)


What is the equivalent RAL paint code for Munsell 3Y7.8/1.1? (330)
I cannot give you an answer. First of all there are different types of RAL designations, some of which are not related to colorimetry. Some RAL designations are directly related to CIELAB coordinates. If you are looking for one of those designations, then you could convert from Munsell to CIELAB (GretagMacbeth provides software that does that) and then from CIELAB to RAL. Both these conversions are addressed in this FAQ. (Back to top)


I want to write a paper on color measurement. Can you tell me the latest developments on this, and what we can do? (329)
The first step in any research is to read, and read a lot. You need to learn about all the background of what has been done in the past and bring yourself up to date so that you can make a new contribution. Searching the internet can provide a lot of resources and insight. You might try scholar.google.com to search for papers on topics that interest you. In color measurement, "Color Research and Application" is the most well-known and respected journal. You can learn more about what's been published on their website. Lastly, in the USA, the organization that deals most with measurement issues is the Council for Optical Radiation Measurements, CORM. You can learn more about them at corm.org. As with learning anything, it is also helpful to work with a good teacher (someone already knowledgeable about the literature and the field). (Back to top)


What is LRV? (322)
LRV is Light Reflectance Value. It seems to be a term commonly used for paints and signage. It is defined as the amount of visible light a surface reflects, but not more strictly than that. I could not find a strict definition, but it appears to be the CIE Y tristimulus value for appropriate viewing conditions. (Back to top)


I use a colorimeter that returns x,y,Y values. What units do they express? (320)
The x,y,Y values you get from your instrument are chromaticity coordinates (x,y) and luminance (Y). The chromaticity coordinates describe the hue and saturation of the stimulus with no information about the luminance. They are unitless, normalized versions of the XYZ tristimulus values. They are computed as follows:
x = X/(X+Y+Z)
y = Y/(X+Y+Z).

There is no way to compute Y from x an y. The x,y values range from zero to one and are generally reported to 3 or 4 digits beyond the decimal point. You can read more about chromaticity coordinates in almost any introductory text on colorimetry. One good example is Berns' "Principles of Color Technology, 3rd Ed.". (Back to top)


I have a densitometer that measures visual density. How do I convert the density values to CIE L*? (319)
The full CIELAB L* equations are as follows:

L* = 116(Y/Yn)^1/3 - 16 for Y/Yn > 0.008856
L* = 903.29(Y/Yn) for Y/Yn <= 0.008856
Y/Yn is the relative luminance of the stimulus (relative to white) in a range from 0 to 1.0. You can convert density to relative luminance using:

Y/Yn = 10^-D

This gives you the equation from Poynton for Y/Yn > 0.008856, which happens to correspond fairly closely with a density of 2. So the Poynton equation is limited since it does not change to the alternative form for Y/Yn <= 0.008856. All you need to do is first convert from density to Y/Yn and then use the two-part L* equations given above and you can compute L* for any density values. (Back to top)


I have a white printing on LDPE which is yellowing. How can I determinate the yellowing in a datacolor instrument? (317)
I would suggest simply using your instrument to measure CIELAB values and examine the L* and b* values. As it yellows, your sample will get lower in L* (lightness) and higher in b* (yellowness-blueness). With some experience, you will be able to establish tolerances for changes in these dimensions that are unacceptable for your application. There are also whiteness indices that you might look into. The software with your instrument can probably calculate them. The only advantage of a whiteness index is that it might give you a reasonable single number to look at. (Back to top)


Is there a way to express Lch differences, in terms of units? (not referring to ΔL*, ΔC* or ΔH*) Hue for example is expressed in degree?s (eg 5°). What unit do I use to express the difference in lightness and chroma? (313)
The units of color difference in the CIELAB space are an interesting bit of color science trivia. Looking at the CIELAB equations, you can see that there are no explicit physical units for the color correlates. That is because all of the tristimulus values (which can have units) are divided by the tristimulus values for the white (which would have the same units). People tend to just refer to the magnitudes of differences in CIELAB as "CIELAB units". This is necessary to indicate the specific color space. There are no physical units to attach to those dimensions (which are correlates of perceptions). (Back to top)


Could you recommend a book that would serve as a good introduction to colour science for the lay reader? I am interested in learning the basics of colour science but am not a scientist or mathematician. (293)
One of my favorite introductory books, and a fun one to read, is "Colour: Why the World isn't Grey" by H. Rossotti. It is also very affordable. There are several other books that might also fit the bill. In order of increasing complexity/mathematical content I would suggest: "Color: Essence and Logic" by R. Keuhni (out of print, but used copies are available) "Color: An Introduction to Practice and Principles by R. Keuhni (an extended and revised version of "Essence and Logic") "Principles of Color Technology, 3rd Ed." By R. Berns (contains some more math, but is also very readable)

For a more general range of related topics, you might also want to look at "Seeing the Light: Optics in Nature, Photography, Color, Vision, and Holography" by D.R. Falk et al. (Back to top)


Where I can find the exact value of Xn, Yn (the luminance factor for the illuminant or perfect white) and Zn? (292)
Xn, Yn, and Zn are the tristimulus values of the illuminant used in the computation of the XYZ tristimulus values of the sample. They are computed the same way as the sample's tristimulus values with the exception that reflectance is removed from the equations (equivalent to setting it equal to 1.0 or 100%). Due to the normalization of the tristimulus value equations, Yn is always 100.0. Xn and Zn vary above and below 100 depending on the particular illuminant. For CIE standard illuminants, the XnYnZn values can often be found in tabulations of the illuminant spectra in standard texts such as Berns' "Principles of Color Technology, 3rd Ed." (Back to top)


Where I should look for for the energy distribution of commercially available sources? (290)
This is a great question. There are really only two sources. Many of the manufacturers will publish nominal spectral power distributions for their light sources. Sometimes you can even find these specs on the internet, but you should be able to get them by making a request directly to the manufacturer. The second option is to measure sources yourself. That's expensive and time consuming, but will actually provide more useful results for real-world settings. I believe the CIE did a measurements of a lot of real fluorescent light sources to create the newer F illuminants in publication 15:2004. (Back to top)


What is the best method (color space, sci vs. sce, degree observer, Illuminant,etc.) for calculating Light Reflectance Value (LRV) on a spectrophotometer? (289)
I can't find a specific definition of light reflectance value (LRV). Based on what I can gather, it is an overall measure of the average reflectance of a sample in the visible wavelengths. Most likely it is computed from the Y tristimulus value (0 for perfect black and 100 for perfect white). Unless there are specific specifications to answer your question, I would select the following:
  • Color Space = XYZ tristimulus space, Y is the light reflectance value
  • Specular Component Included, measure total reflectance
  • CIE 1931 2-degree Observer, this is the strict definition of luminance so is probably slightly more appropriate
  • CIE Illuminant D65 (you might want to pick others for specific applications, but this is the generic average daylight choice)
(Back to top)


Where I can find the ultraviolet reflectivity index of various colors? How about the heat absorption index for various colors? (280)
Unfortunately I can't help you. While there might be some correlation between UV and IR absorption and an object's color (absorption in the visible), it is not perfect and colors can have one property in the visible wavelengths and different in the UV and IR. In general, colors that reflect more light in the short-wavelength regions of the visible spectrum are more likely to reflect more UV as well. These would be blues and purples. Likewise, colors that reflect in the long-wavelengths (yellows, oranges, reds) are also likely to reflect a lot of IR (and therefore absorb more heat). The IR end is much more difficult to predict however. An object heating up is not just because it absorbs IR (regardless of wavelength) but also because it absorbs other wavelengths. For example, black objects that absorb all the visible wavelengths will heat up more than white objects that reflect all visible wavelengths (assuming their behavior at non-visible wavelengths. is similar. (Back to top)


Is there a reference offering a method for the bi-directional calculation of densitometric quantities (particularly status M) and spectral transmittance to be used as weighting coefficients in the integration to solve for XYZ tristimulus values? (278)
In general it is not possible to perform the reverse computation of spectral transmittance from densities. The density values are a weighted integration of spectral transmittance and it is impossible to recreate arbitrary spectral transmittance functions (perhaps 31-dimensional) from the 3-dimensional density data. If however, you are working with density measurements for a single system, particularly a 3-color system like photographic film, it is possible to derive an empirical relationship (based on sound theory) between the density measurements and the dye or pigment concentrations. You could then compute spectral transmittance (or reflectance) from the dye concentrations and knowledge of the dye properties. This has been previously done in the context of media-dependent scanner calibrations. A reference on that topic is R.S. Berns and M.J. Shyu, "Colorimetric characterization of a desktop drum scanner using a spectral model," Journal of Electronic Imaging 4, 360-372 (1995). (Back to top)


What is the basis for 18% reflectance being the center of the lightness scale in many different systems? For example: Munsell, Ansel Adams zone system, and CIELAB. Coincidence? I doubt it. (276)
You are correct that it is not a coincidence. The number you are referring to ranges between about 18% and 20% reflectance in various systems. There are two bases for this number. The first is that the average reflectance of natural scenes tends to be about 18-20%. This was measured by various researchers, perhaps the most famous of which was by Jones and Condit at Kodak in 1941. This value has been used to set proper exposures in photography. The second basis is human perception. The appearance of middle gray (perceptually half way between white and black) has a reflectance of about 20% (and hence the definition of Value 5 in Munsell or L*=50 in CIELAB). Is it a coincidence that the average reflectance of natural scenes and the human perception of middle gray have the same reflectance factor? I don't think so. The human visual systems functions by sensing contrast from the background average and evolved in a world with this average reflectance. No coincidence there. (Back to top)


Would you explain the meaning of ΔH*? (273)
Color differences in CIELAB, ΔE*ab, represent Euclidean distances between the coordinates of the two colors in question in the L*, a*, and b* dimensions. These differences can be expressed in terms of lightness, ΔL*, chroma, ΔC*, and hue, ΔH*, difference components in the same Euclidean units. ΔH* is computed as the residual difference in the overall color difference, ΔE*ab, after the lightness, ΔL*, and chroma, ΔC*, components can be removed. That's the long way to say that ΔH* is a measure of the perceived hue difference in Euclidean CIELAB units (same size as L*, a*, and b* units) as opposed to a difference in hue angle in degrees (Δh ... note it's a lower-case h for hue angle difference). (Back to top)


What is the scientific termonology for when you fold a piece of fabric and two areas appear to be entirely different colors? I've seen this demonstrated: after folding, both sections of fabric appeared to be different colors. (271)
I've asked a couple of people and we can't come up with a specific term for that situation. One possibility could be "goniochromatic", which means something changes color with viewing angle. Another possibility might be "geometric metamerism", but that would really be a misuse of the term "metamerism" (which would imply a pair of samples that match for one viewing geometry, but mismatch for another). (Back to top)


I need to find a way to measure the saturation and brightness of printed colors and pen ink. I need to select hues of equal brightness and saturation in my studies. Can you provide information about machines that can make such readings? (270)
Your best bet is to look at mamnufacturer's web pages to see what's available that could meet your needs. Konica-Minolta is one company that makes a line of rather inexpensive colorimeters and spectrophotometers that might meet your needs. The other companies (Datacolor, X-Rite, Hunterlab) might have appropriate instruments as well (the X-Rite Eye-One is an instrument that comes to mind).

What you will probably want to measure are CIELAB coordinates. This color space provides three parameters, L* which correlates with perceived lightness, h which correlates with perceived hue, and C* which correlates with perceived chroma. Chroma and saturation are similar perceptual dimensions. Color scientists worry about the differences, but for most applications the chroma dimension is the one that people are most interested in. (Back to top)


I have used a spectrophotometer to analyse colour samples. How do you interprete the curves, for example prussian red, since there are many peaks shown? (269)
Unfortunately I cannot answer your question without more details on what you are hoping to accomplish. Some applications, like identification of pigments or their concentrations, are performed directly with the spectral reflectance measurements. Others, like color evaluation and control are completed using colorimetric parameters such as CIE XYZ or CIELAB L*a*b* values computed from the spectral measurements (often by software that comes with the instrument). A great introductory reference to learn about these different ways to use spectrophotometric data is Berns' Principles of Color Technology, 3rd Ed. (Back to top)


My technical director asked if I could find an "easy to understand" explanation of "CIE 2000." He may mean the CIE DE2000 mentioned in this publication. Is there a publication or expert that could assist us? (268)
I think you are correct in assuming that he was referring to the CIE DE2000 color difference equation. I don't know of any other reference to "CIE 2000". There is another CIE publication that describes the equations in more detail. Unfortunately I don't know of any textbooks that yet have an explanation of the equations. (Berns' "Principles of Color Technology, 3rd Ed. describes much of the background work, but was published just before DE2000.) You may also want to look at a paper that one of my colleagues and I wrote on color difference measurement for images. It contains a lot of spatial imaging information that is not necessary, but starting on page 14 is a fairly brief discussion of DE2000. This paper can be found in Color Research and Application in 2003 (G.M. Johnson and M.D. Fairchild, "A top down description of S-CIELAB and CIEDE2000," Color Research and Application, 28 425-435 (2003)). (Back to top)


Can you provide me a graph of grayscale Values (0-255) verses density? I need it to evaluate a grayscale calibration wafer. (267)
Unfortunately the answer is "no." There are simply too many undefined variables. Luminance (or at least relative luminance) would be the common characteristic, but you need to have a monitor calibration and characterization in order to relate the digital counts (0-255) to the relative luminance values displayed. On the density side, if you have visual density, then that is logarithmically related to relative luminance, but the relationships for other density metrics might not be so direct.

So, to derive the answer, you need to first convert density to relative luminance using the definition of density you are interested in, and then relative luminance to digital counts using the display characterization. That can give you the relationship for your particular devices and measurements. (Back to top)


How can I specify a color so that manufacturers the world over will use the same color paint? (266)
Yes, there is a standard way to match your color. What you want is a colorimetric specification and tolerance. Typical specifications are in terms of CIELAB coordinates of L*, a*, and b* and tolerances are normally set around the desired color in terms of some color difference equation. When the paint manufacturers matched your samples in the past, it is likely they measured existing samples and then matched those CIELAB coordinates. Perhaps you can get those values from one of the paint manufacturers to set your standards. Alternatively you can choose a standard sample and send it to a paint supplier and they can measure it. Regardless, it is always good practice to have a physical sample as a standard to allow visual checks of the instrumental color matches. (Back to top)


In relation to Kodak absorption filter graphs; What does "Transmittance" & "Diffuse Density" mean? and the difference between them? (260)
Our collective wisdom is that the Wratten filters were measured using a GE Hardy spectrophotometer with the filters up against the sphere port. Thus the geometry was 8° incidence and total hemispherical collection. Assuming Helmholtz reversibility of optics, this is equivalent to diffuse illumination and nearly straight on viewing, a common photographic geometry, e.g., transparency viewer. Transmittance is the amount of light transmitted compared to the amount incident. Density is the "common logarithm of the reciprocal of transmittance" according to the Kodak Filter guide. So, 10% transmittance is equal to a density of 1. (Back to top)


Can you provide spectral reflectance values for the Farnsworth-Munsell 100 Hue test? (252)
We don't have them, but I recall someone asking this question recently and the best I could find on the internet was this reference www.ncbi.nlm.nih.gov to a paper that did eigenvector analysis of the reflectances. I haven't looked at the paper itself, but the abstract says they measured the reflectances and even the eigenvector analysis (if it is fully presented in the paper) might let you make reasonable reconstructions. (Back to top)


My spectrophotometer is getting old. I need a new instrument and software for evaluation and formulation of color paints and ceramics. What recommendations can you make? (248)
There are several companies currently manufacturing spectrophotometric hardware and color matching software capable of replacing your system. Several of these are linked on our lab's sponsors page. The companies most likely to have systems with the flexibility you require are BYK-Gardner, HunterLab, DataColor, and GretagMacbeth. Minolta and X-Rite also manufacture instruments, but I am not sure they have software for your applications. Unfortunately, I am not aware of any "consumer's report" on this instrumentation, however I can say that the quality of all of the instruments is very high and likely substantially improved from the time you purchased your system. I would recommend looking at options from all of the manufacturers and making your decision based on the availability of instruments that can measure the variety of samples you have, the capability of their software for your applications, and the ability to transfer any necessary data from your old system. (Back to top)


Would you explain what ΔE is and how is it measured? (243)
ΔE refers to a quantitative measure of color difference that is designed to correlate with our perception of the magnitude of color difference between two stimuli. There are a number of different ΔE formulas that each compute this quantity in a slightly different way. The reason there are more than one is that the prediction of perceived color difference is an imperfect science. There are a lot of physical and perceptual variables involved and research is ongoing to develop improved color difference equations. That said, there are several that are commonly used with significant success. The simplest is the CIELAB ΔE*ab, which is simply the Euclidean distance between the color coordinates of the two stimuli in the CIELAB color space. The coordinates are L* (lightness), a* (redness-greenness), and b* (yellowness-blueness). More complicated color difference equations typically start with these CIELAB coordinates, but are based on more complex distance formulas intended to better predict human perception. The measurements of the stimuli required are either spectral reflectance/transmittance of materials or the spectral power distributions of light sources which are then converted into CIE XYZ tristimulus values and then CIELAB coordinates computationally. A good overview of the development and use of these color difference equations can be found in Berns' introductory text, Principles of Color Technology, 3rd Ed. MCSL also has several related resources available. (Back to top)


What is L-star? (242)
In the CIE system of colorimetry, the Y tristimulus value represents the luminance factor of a sample. This number is intended to correlate with our perception of how light or dark the sample appears, referred to as lightness. Our perception of lightness also depends on our state of light adaptation, which is how our visual system adjusts to various illumination conditions (like automatic exposure on a camera), and is nonlinearly related to the energy in the stimulus (represented by Y). Thus, the L* equation was derived to account for those two factors. It accounts for adaptation by dividing the sample luminance factor by the luminance factor for the illuminant (or a perfect white) and it accounts for the compressive nonlinearity in our visual response through the cube-root and other constants in the equation. Therefore, L* is a computed colorimetric quantity that is intended to linearly correlate with our perception of the lightness of a stimulus. You can find more details on the L* in any modern textbook on colorimetry such as Principles of Color Technology, 3rd Ed. by Berns. L* is also one of the three dimensions of the CIELAB color space where the other two, a* and b*, represent the hue and chroma of the simulus.

The L* equation is
   L* = 116(Y/Yn)^(1/3) - 16
where Y is the luminance factor of the stimulus and Yn is the luminance factor for the illuminant (or perfect white). (Back to top)


What is the established color difference tolerance for the plastics industry? (237)
There are not generally accepted standards for specific industries that suggests a single ΔE value as a standard. The tolerance is usually agreed upon between buyer and seller for the particular colors and applications in question. There are several issues that preclude use of such a simple tolerance. The first is the specification of the color difference equation to be used. Choice of a different equation (e.g. CIE94, CIE2000, CMC, or CIELAB) will result in differences in the perceptual magnitudes and characteristics of the stated Delta-E tolerance. The second is the dependency on color. For example, a CIELAB ΔE of 1.5 for a gray color is perceptually different from the same ΔE for a yellow color. This is part of what the newer equations like CIE94 and CIE2000 attempt to address. The differences between the various color difference equations and techniques for establishing meaningful tolerances with them are described in Chapter 4 of Roy Berns' book, Principles of Color Technology, 3rd Ed. (Back to top)


References

  1. G. Wyszecki and W.S. Stiles, Color Science: Concepts and Methods, Quantitative Data and Formulae 2nd Ed., Wiley, New York, 1982.
  2. R.W.G. Hunt, Measuring Colour 3rd Ed., Fountain Press, England, 1998.
  3. R.M. Boynton, Human Color Vision, Special Limited Edition, Optical Society of America, Washington D.C., 1992.
  4. M.D. Fairchild, Color Appearance Models Addison-Wesley, Reading, Massachusetts, 1998.
  5. R.S. Berns, Billmeyer and Saltzmann's Principles of Color Technology, Wiley, New York, NY, 2000. (from Wiley)
  6. H.G. Völz, Industrial Color Testing 2nd Ed., Wiley-VCH, Weinheim Germany, 2001.
  7. There are several links to GretagMacbeth's Color Conversion Freeware