­¦ý°±·>œ Z\œYÏ€¡©@ •øºYjbjbаа %tˆˆºSl¢¢¢¢¢¢¢ÜÜÜÜÜ íLÜOÍÚÚÚÚÚÚÚÚ ,9 Y^:¢ÚÚÚÚÚ:8¢¢ÚÚÚ888Ú• ¢Ú¢Ú 8hh¢¢¢¢Ú 8Œ8 ¢¢ ÞLå&ŸÜÜ’ V  OO …8… 8ˆ handprint : comparison of hue circles
 

comparison of hue circles

 
This page presents the hue circle as defined in six contemporary color measurement or color description systems:

• MUNSELL

• Swedish NCS

• CIELAB

• CIECAM

• a mixing color wheel (a color wheel defined by the subtractive mixing of paints)

• the artist's color wheel (a color wheel defined by the additive mixing of light, with hue spacing that represents the saturation costs among different mixtures of paints).

The hue positions in these wheels are approximate only, based on the average position of all paints included in my guide to watercolor pigments. Chroma (saturation) has been standardized to place the most saturated colors on the circumference of each wheel.

 

the MUNSELL hue circle

The Munsell color order system is the oldest color model in common use today. The hue circle is divided into five equally spaced basic colors — yellow, red, purple, blue and green, in counterclockwise order — and five blends between them (yellow red, red purple, purple blue, blue green and green yellow). Each of these 10 color sections is divided into ten equal gradations, producing 100 hue intervals; the central or "pure" hue in each category is located at division 5 (yellow is at 5Y, red at 5R, etc.).

The hue spacing was determined through incremental perceptual comparisons between hues around the entire color circle, at different levels of chroma and lightness. Over small hue changes at a constant chroma, equal distances along the circumference represent equal perceived differences between colors. Thus, middle red and deep yellow are perceptually as different as green blue and red blue.

Hues positioned opposite each other on the hue circle represent visual complementary colors, which were determined by visual mixtures of colors on a color top, as described in the book review of Ogden Rood.

Most art instructors today still follow what is supposedly J.W.V. Goetheís definition of complementary colors as yellow/violet, orange/blue and red/green. In part this is based on two misunderstandings of Goetheís color terms, but the Munsell hue circle also shows that these traditional complementary pairs are inaccurate. The specific relationships are more accurately:

light yellow — red blue
middle yellow — middle blue
deep yellow — green blue
scarlet red — turquoise
crimson red — blue green
magenta — middle green
purple — yellow green

The Munsell hue and chroma spacing has been extensively researched and documented over the past century, and the Munsell is now commonly used as a benchmark spacing of hues when assessing the consistency or accuracy of newer color models, for example the CIELAB a*b* plane.

Note the large span between middle blue and blue green, and between yellow green and light yellow. These hues comprise a large part of the Munsell color wheel, yet are infrequently used by most artists. In Munsell:

Blue green is the complement of crimson
Green blue is the complement of deep yellow

See the Munsell Color System for more information.

 

the MUNSELL hue circle

 

the Swedish Natural Color System (NCS)

NCS represents the culmination of the subjective or phenomenological tradition in color research. Color positions were determined as the average of many individual perceptual judgments, using imaginary or ideal color standards as a reference. Test individuals were asked to imagine the four unique hues — a "pure color" yellow, red, green or blue that contains no tint of any other hue — then to estimate the proportions of red+yellow, blue+red, green+blue or blue+yellow required to mix specific hue samples. Subjects could include black and/or white, also judged against imaginary standards, to describe pastel, darkened or grayed colors. Additional experiments demonstrated that this method was about as accurate as judgments based on physical samples of the unique hues.

The NCS is defined on phenomenological complementary colors — that is, they are not complementary colors by the normal definition. Hues opposite each other on the color circle typically do not mix to gray in subtractive or in additive color mixing. In addition, the unique hues themselves are not complementary.

The hue circle is divided into 100 units between neighbor unique hues, 400 in total. However hue spacing is not perceptually equal across each quadrant of the hue circle. The perceived color difference across 10 unit intervals in the red to blue hues is much larger than the perceived difference in the same interval across the green to yellow hues, and differences across the yellow to red hues are much larger than those across blue to green. These perceptual irregularities appear clearly when the NCS is represented in CIELAB, and are necessary in order to make the model geometrically consistent.

Note the expanded spacing of the deep yellow to magenta hues and the large span between light yellow and blue green, with the correspondingly excessive compression of the distance between magenta and middle blue. In NCS:

Blue green is the complement of scarlet red
Green blue is the complement of middle yellow

See the section on the Swedish Natural Color System for more information.

 

the Swedish NCS
the horizontal (red/green) dimension has been reversed, to match the other color circles on this page

 

the CIELAB a*b* plane

CIELAB is a color model proposed in 1976 by the CIE, an international body governing standards and methods in colorimetry. It requires electronic or photoelectric measurements of color and is widely used in the print and imaging industries and in manufacturing that requires high color control (such as automotives or paints).

Color locations are computed from the proportional stimulation to the R, G and B cones, mathematically transformed into a dimension of lightness (L*) and two perpendicular dimensions of hue, which do not match the Hering unique hues or the opponent process dimensions of the visual system (as measured in rhesus monkeys). The hue and chroma of all colors can be represented as a weight or position on these opponent dimensions, labeled a* (where red a+ is a positive value on a* and green a- is negative) and b* (yellow b+ vs. blue b-). Hue locations on the hue circumference can be expressed as a hue angle; the circle is not divided into equal intervals. The a* axis is conventionally shown with a+ (red) on the right; this has been reversed to facilitate comparison with other models.

CIELAB represents visual complementary colors, that is complements defined by mixtures of light or of surface colors visually blended by rotating them on a color top. All complementary colors have the relationship: hue (a*,b*) = hue complement (-a*,-b*). Hue, chroma and lightness are defined so that equal distances between any two colors in the CIELAB color space (over small distances) represent approximately equal perceived color differences in the color samples. However, when compared to MUNSELL, CIELAB appears to exaggerate color differences across green hues and to minimize color differences in blue greens and magentas.

CIELAB is not considered an especially accurate color model by color scientists, but it is quite useful to define color differences and color tolerances and thanks to extensive implementation via automated or electronic color management systems it is currently the de facto color measurement standard around the world.

Note the very small span between deep yellow and magenta (especially between orange and crimson), the large span between middle blue and turquoise, and the large span between blue green and light yellow. In CIELAB:

Blue green is the complement of magenta
Green blue is the complement of middle orange

See the CIELAB uniform color space for more information.

 

the CIELAB a*b* plane
the a+/a- dimension has been reversed, to match the other color circles on this page

 

the CIECAM aCbC plane

CIECAM is the most recent color model proposed by the CIE, an international body governing standards and methods in colorimetry. It requires electronic or photoelectric measurements of color and is at the early stages of adoption in the print, imaging and manufacturing industries that requires high color control (such as automotives or paints).

The technical description of CIELAB (above) apply as well to CIECAM, with a few important exceptions. CIECAM is not based, as is CIELAB, on a "wrong" von Kries transform (adjustment of color appearance in relation to the illuminant color and a "white" surface color standard), and there are several adjustments made to model contrast effects with the surrounding color, the effects of luminance level, the differences between projective or reflective media, and so on. The scaling of both chroma and hue spacing have been improved, and so far as it has been tested on actual visual data, CIECAM performs as well or better than any other color model.

Like CIELAB, CIECAM represents visual complementary colors — in this case, complements defined by cone responses. All complementary colors have the relationship: hue (a*,b*) = hue complement (-a*,-b*). The chroma spacing is also adjusted to account for the decreasing apparent chroma differences in saturated colors. Hue, chroma and lightness are defined so that equal distances between any two colors in the CIELAB color space (over small distances) represent approximately equal perceived color differences in the color samples. Compared to CIELAB, CIECAM appears to have an improved (more accurate) representation of green and blue colors, which alters the complementary color pairings. (Compare, for example, the complementary color for red blue.)

CIECAM is probably the most accurate representation of visual complementary colors available today. The diagram shows that these are not the same as the 18th century complementary color pairs (yellow/purple, orange/blue and red/green) that are still taught in many art courses. The actual visual complement to orange is cyan blue, crimson goes with turquoise, magenta with blue green, and the complement of purple is yellow green. CIECAM is also probably a terminus in the development of "global" color appearance models. Current work is turning in the direction of context dependent color effects, in particular the effects of contrast and spatial size on color appearance.

As in CIELAB, note the very small span between deep yellow and magenta (especially between orange and crimson), the large span between middle blue and turquoise, and the large span between blue green and light yellow. In CIELAB:

Blue green is the complement of magenta
Green blue is the complement of middle orange

See the CIECAM color appearance model for more information.

 

the CIECAM aCbC [chroma based] hue plane
the a+/a- dimension has been reversed, to match the other color circles on this page

 

a mixing color wheel

Mixing color wheels are standard issue among painters, although these hue circles are not based on color perception and cannot be accurately defined with paint mixtures.

Mixing color wheels represent mixing complementary colors, that is, two paints that physically mix a pure gray or near neutral color are placed opposite each other on the circle. The fact that the mixing hue circle shows colors and not paints is the source of mischief. Mixing complements are always imprecise because (1) all paints of a given hue are not effective mixing complements for all the paints of the complementary hue, (2) the mixing complements for a single paint may differ widely in hue (paints obliquely opposite each other on the color circle can also mix grays or neutral colors), and (3) chroma is as important as hue in defining mixing complement pairs. To appreciate the difficulties created by these problems, see the sections on substance uncertainty and mixing complementary paints.

Once opposing paint pairs have been determined through mixing, the spacing between neighbor colors is somewhat arbitrary. It can be determined by proportional mixtures of the three "primary" paints or by equal apparent hue differences (as much as the mixing relationships will allow). Usually mixing wheels are arranged so that the three "primary" colors are equally spaced around the circle (at the 12, 4 and 8 o'clock positions on the circle) and therefore represent paints of equal chroma and tinting strength. This ignores the actual saturation costs between paints that are not mixing complements, including the "primary" colors themselves.

Within these limitations all the mixing complements shown below are based on my own careful measurements of watercolor mixing complements.

Even so, large gaps or crowding in the hues between two "primary" paints are unavoidable, because colors such as yellow have no effective mixing complements and because the neutralizing colors for blue, turquoise or yellow green paints are spread over a wide hue range. In the circle below, this is the reason for the very large gap between blue violet and red blue, and the small span between middle blue and blue green.

Contrary to traditional "color theory" dogma, red is not the mixing complement of green, and purple (violet) is not the mixing complement of yellow. The mixing complement of middle green is magenta, and the mixing complement of middle red is blue green (turquoise or teal); the mixing complement of purple is yellow green, and light to medium yellows have no true mixing complements, but mix very dull greens with red blues and tans with purple. In a mixing wheel:

Blue green is the complement of middle red
Green blue is the complement of scarlet red

See the section on visual vs. mixing complements for more information.

 

a mixing color wheel

 

the artist's color wheel

As we've seen, the different approaches to defining a hue circle are not logically or perceptually compatible. Hue circles must always represent some aspects of color accurately and tolerate some distortion in other aspects.

The first priority in my artists' color wheel was accurately representing visual complementary colors; this was done in 1999 by reconciling the complementary color representations in CIELAB and Munsell. The second priority was representing as accurately as feasible the saturation costs among paint mixtures. This was done by compressing the hue spacing of greens (to compensate for the curved mixing lines across green colors) and by exaggerating the chroma spacing of saturated warm hues, which reduces the curved mixing lines produced on this side of the space by the green respacing. (This also reflects the painters' penchant for emphasizing small color differences among "warm" hues.) Mixing complements are ignored entirely because it is more accurate and convenient to summarize them as a comprehensive table.

In 2006 I reanalyzed all my pigment data in CIECAM, and found that the CIECAM revisions had addressed the problems in the hue and chroma spacing of CIELAB that I had tried to remedy by adjustments based on Munsell. The 2006 version of my artists' color wheel is essentially identical with the CIECAM color spacing.

In this artists' color wheel (as in CIELAB):

Blue green is the complement of magenta
Green blue is the complement of middle orange.

See the artist's color wheel for more information.

 

the artist's color wheel (1999 version)

Last revised 11.12.2007 • © 2007 Bruce MacEvoy
 
ºY˜ OJPJQJIJ°¢´¨“”ÈÄÅÆØ8Q’Ò$ 3 e t § „ “ Ú n } ————————————————————————————1$7$8$H$Õ   ! 6 Æ º “Áquìófj鄹–8W[ôùtÉÈÌ^hnot~ ghmÇ |ÄŒÿ^!b!?"N"Ö"†"æ"B#L#R#S#X#b#t#Å$ã$ë$í$þ$ý$Â$™$q%%''®*¨*- - //Â/Ù/'0\0`0÷0ý0Ê0Á0Ï0–022 2!2n2o2t2œœd} ~ å º Ã Õ   ! 6 Æ º “Áquìófj鄹–8W[————————————————————————————1$7$8$H$[ôùtÉÈÌ^hnot~ ghmÇ |ÄŒÿ^!————————————————————————————1$7$8$H$^!b!?"N"Ö"†"æ"B#L#R#S#X#b#t#Å$ã$ë$í$þ$ý$Â$™$q%%''®*¨*-————————————————————————————1$7$8$H$- - //Â/Ù/'0\0`0÷0ý0Ê0Á0Ï0–022 2!2n2o2t2â23$3Õ4‹4y7}7————————————————————————————1$7$8$H$t2â23$3Õ4‹4y7}7Ò:ž:ª=ø=§>„>Ê>??ò?¢?®?©?Æ?½?š@A A A[A\AaAvAÏA™A•BÙBÅFÖFçIëI@JDJÎKÔK¿M¦MN8N„>Ê>??ò?¢?®?©?Æ?½?š@A A A[A\AaAvAÏA™A•BÙBÅF————————————————————————————1$7$8$H$ÅFÖFçIëI@JDJÎKÔK¿M¦MN8NNGNOOãOèOšOPGPNPQQ!Q"Q$Q&QIQJQZQ[QiQmQÞQýQ”QÊQRRARMRúRûR¢R§R¡R«RÌR˜RBSISbSdShSjSoSvSëSìSæS¢§9:RV“”–¯%&8:fgy{€¶½†”—˜opÇÑÄÆƒ‡ø» ‘­"$79؃‡¡”  r s   î ï  % x y 6eghkléî„…¹Ð–›8>XYöõuvÍÎGI_ajlpruwÅ  iknpÉÖ  vz}~¹÷…_`@Aªº+-CENPTVY[ce¼æÇÑçèìï·“ÊËš›rsÄÑ!!©$$P%Q%Í%Ï%''«'»' ) )Ê)Á)]*^*ø*¡*×**’*”*Ë*Í*Ì*Ô*—*˜*,,,,",$,p,r,u,w,ä,å,--%-)-Ä.ò.‘.¦.Ô.•.z1{1i2k233Ú4Û4-5.5º7‡7€8¶899Å9É9ô9õ9§9¶99¨9Ø9±9¼9ª9¸:œ:; ; ;;];_;b;d;w;y;Ì;Ó;š;;Ò<Ú<˜?™?Ç@É@BBÔB•BéCèCœCCADBDÏEÌE¡G¬G9H:HÆHƒHH»H‹H²H×HHÐHÞHÊHËHqIsI|I~IÇIÑI³IŒI×IIÏIÓIfJgJtJxJvKwKèLêLOOYOZOP…PºP»PÁPËP_Q`QªQ‡Q²Q¹QÞQýQ”QÊQÈQÎQÛQžQëRìRúRûR¢R§RßR©R±R„RWSYSbSdShSjSæS:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: Bruce MacEvoy,Macintosh HD:HANDPRINT:HP:WCL:vismixmap.html Bruce MacEvoy-Macintosh HD1:HANDPRINT:HP:WCL:vismixmap.html Bruce MacEvoy-Macintosh HD1:HANDPRINT:HP:WCL:vismixmap.html Bruce MacEvoy-Macintosh HD1:HANDPRINT:HP:WCL:vismixmap.html Bruce MacEvoy-Macintosh HD1:HANDPRINT:HP:WCL:vismixmap.html Bruce MacEvoy-Macintosh HD1:HANDPRINT:HP:WCL:vismixmap.html@Ädd»þVVdUÔâTºSÝ @GêTimes New Roman5êÄSymbol3ê Arial3êTimes à•­hWRŸÜ9bŸ Þ C "ê Ï•€¿xxÉÄS/•þ Bruce MacEvoy Bruce MacEvoyœ ýÖüÚ˜Oh´ë+'„0`àêú®¿ÃÐ Ù  ( 4@HPX'ososBruce MacEvoyorucNormalaBruce MacEvoyo14cMicrosoft Word 10.1@dâ@äó$'õ»@Óï×µú» Þ Cœ ¹Õ¹ú.ìó+,˜Æ0Œ `hpxÄ àêòÝ ® µ'ê"S÷  Title  !"#$%&'()*+,-./0123456789:œ<=>?@ABCDEFGHœJKLMNOPœRSTUVWXœ›[œœœRoot Entry ¿F7+{ú»]Ä1Table;…WordDocument%tSummaryInformation(IDocumentSummaryInformation8QCompObjXœœ ¿FMicrosoft Word DocumentœNB6WWord.Document.8