The elements of colour

11.17 Design Institute of Australia webinar "The Elements of Colour", May 18, 2023


  • 03:20 Presentation begins
  • 10:40 Hue, lightness, chroma and blackness
  • 14:41 Lightness and chroma vs brightness and colourfulness
  • 23:43 Saturation
  • 25:04 Brilliance
  • 27:07 What is a colour?: Colour stimuli, colour perceptions and colour measurement
  • 28:12 Colours of light
  • 37:55 Colours of objects
  • 39:29 Colorimetry
  • 47:22 Perceived colour and psychophysical colour
  • 48:56 Selected diagrams from the second paper

[03:20] I’d very much like to thank the Design Institute of Australia for inviting me to speak tonight to start off their Colour Series. I’d especially like to thank Catherine Piergrosse for extending the initial invitation to talk, and Nathan and Georgina for organizing the webinar tonight.

The invitation was timely because I’ve just had a two-part paper on the fundamental elements of colour accepted for the Journal of the International Colour Association or AIC. 

Part One, which I’ll concentrate on tonight, presents an extended answer to the question "what is a colour?" by reviewing the connections between colour stimuli, colour perceptions, and colour measurement, while Part Two focuses on an analysis of colour perception, exploring the main modes of colour appearance (as belonging to objects or lights in various ways) and perceived colour attributes (hue, lightness, and so on) and their interrelationships. 

Both parts will appear in a special issue of the Journal devoted to contributions from the ISCC/AIC Colour Literacy Project, very patiently edited by Robert Hirschler, and due out next month. The AIC Journal is open-access and you’ll be able to download all of the papers as pdfs from the link shown here.

It might seem reasonable to assume that most colour educators are in broad agreement about the fundamental facts concerning colour, and would disagree, if at all, only about the best ways to address those agreed facts. But my impression, based on more than twenty years of discussions, both in person and online, is that this perceived agreement is an illusion that would be dispelled very quickly the moment any sizeable group of colour educators began discussing colour in detail. Widespread disagreement exists, not only stemming from simplistic traditional RYB colour theory, but also in relation to our current scientific understanding of many fundamental aspects of colour.

While legitimate differences of opinion exist about some details, my aim in these papers was to provide an account of the the fundamental elements of colour according to the scientific consensus expressed or implied in the definitions of the International Lighting Vocabulary or ILV of the Commission Internationale de l'éclairage  or CIE, except where additional constructs are explicitly added, such as from the Natural Colour System  or NCS. In doing so I’ve drawn on insights, explanations and illustrations I’ve developed over more than twenty years of teaching courses on colour for adults mainly with an art or design background, and fifteen years of creating online resources on colour. The papers build on some earlier presentations including those I gave for AIC 2021 and AIC 2022.

Founded in 1913, the CIE is the organization responsible for the international coordination of technical standards relating to colour and light. In relation to colour, the CIE developed the framework of modern colorimetry from the early 20th century through to the present, and established numerous colour spaces that form the basis of colour management and colour technology, including, the horseshoe-shaped 1931 x,y chromaticity diagram that you may have encountered in various contexts, CIE LAB, the “Lab” space familiar to you from Photoshop, the luminous efficiency function that is the basis of all lighting standards, various standard light sources such as D65, and various standard formulae used for quantifying colour differences in industry. The CIE International Lighting Vocabulary, first published in 1938 and revised periodically up to the current edition published in late 2020, gives definitions of nearly 1400 terms and is by far the most comprehensive and authoritative source available on the terminology of light and colour in science and technology.

The CIE ILV defines the word “colour” in two distinct senses, the perceptual and the psychophysical (or colorimetric). I’ll return to colour in the psychophysical sense towards the end of this talk. “Colour” in the perceptual sense is defined as a characteristic of a visual perception that can be described by attributes of, and I quote “hue, brightness (or lightness) and colorfulness (or saturation or chroma)”.

That might look like they're giving alternative names for just three attributes, but those are actually six differently defined attributes of colour. We’ll look at these definitions individually in a moment but for now I’d just like you to note that colour and its attributes like hue, lightness etc are all defined as characteristics of visual perception and not as physical properties as you might have expected.

These six attributes and their definitions are based on work published in the late 1970’s by Robert W. G. Hunt, and have been essentially stable from the 4th edition of the CIE ILV published in 1987 through to the current edition, published in late 2020. But despite this long period of stability of the standard nomenclature, I’ve observed over many years that the distinctions between some of these CIE-defined attributes are often poorly understood even among very experienced colour educators, and indeed, that these distinctions sometimes seem much easier to grasp for students who are new to the subject than for those with firmly entrenched ways of thinking about colour.

Several factors could have contributed to the delayed reception of these concepts in colour education. The CIE ILV has always been expensive and relatively limited in distribution in hard copy, and its definitions only became freely available in 2012 as the online e-ILV. In addition, the terse verbal and mathematical definitions in the ILV can be difficult to understand for non-specialists, and I’m hoping that the insights, explanations and illustrations presented in my paper will lead to more widespread understanding of these standard attributes.

[10:40] Just three attributes suffice to describe colours as long as we consider only a single mode of colour appearance, for example, the colours perceived as belonging to light-reflecting objects. For these colours the three CIE-defined attributes of hue, lightness, and chroma suffice to specify a colour.

Hue, represented by the position around the color circle, is described in the Munsell system in relation to five what are called principal hues: red, yellow, green, blue and purple, and in various digital colour spaces as a hue angle.

Lightness, which is also known as value, grayscale value or tone, refers to the scale between black and white through various grays, and this is specified in the Munsell system in terms of Munsell value, which ranges from zero to ten, and in CIE LAB by L* which ranges from 1 to 100.

And finally we have chroma for in effect the intensity of a colour perceived as belonging to an object. Chroma has often been equated with saturation, but in CIE terminology saturation is defined as a separate attribute of colour.

An alternative set of three colour attributes that can be used to specify colours of objects are the attributes of hue, blackness and chromaticness used in the Natural Colour System or NCS, as shown on the right. These attributes stem from thinking of colours of objects as having resemblances to, or perceptual components of, black, white and pure colour, adding up to 100. In standard CIE terms, chromaticness, which increases towards the right-hand corner, is in effect chroma relative to the maximum chroma perceived to be possible for object colours of a given hue, but blackness, or perceived black content relative to pure black, which increases in towards the bottom left corner, is an entirely distinct attribute of colour not currently defined in the CIE ILV.

Both hue, lightness & chroma and hue, blackness & chromaticness only really work for colours perceived as belonging to objects. Lightness (or greyscale vale) for example is our perception of how much of the light falling on an object that reflects, compared to a white object. For example, a sheet of white paper is perceived to remain high in lightness (white), even as we reduce the brightness of the light reaching our eyes from it by putting it in a shadow. Colours perceived as belonging to light exhibit brightness, which has an open-ended range, but not lightness (or greyscale value) - OR blackness. For example, think of a view of city lights: we see these as varying in brightness, but we don’t normally perceive the brightest light as being white and all of the other lights as shades of grey, or as having a black component. (It might occur to you that right now we’re perceiving lightness (or greyscale value) and blackness on a light-emitting screen, but the point here is that we normally «read» a screen as an illuminated page rather than as a primary light source, which it physically is). What matters is that we normally perceive the screen as if it was an illuminated page.

[14:41] Both lightness and chroma are defined in the CIE ILV in a way that restricts them to “areas” perceived as being illuminated, that is, as illuminated objects rather than light or light sources. Other colour attributes including brightness come into play when we wish to describe colours perceived as belonging to light itself, including (1) light perceived to be falling on objects, or (2) light reaching the eye, either directly from a primary light source, or reflected or transmitted by objects. Therefore, more than three attributes are needed to fully describe the appearance of illuminated objects, which involves colours in more than one mode of colour appearance.

In this image we perceive a cube having a uniform orange colour belonging to it, as if it were painted all over with the same orange paint having a constant hue, lightness and chroma corresponding to a Munsell notation of about 10R 6/14. For actual objects the attributes of lightness and chroma exhibit a tendency towards stability or constancy under varying viewing conditions, especially varying intensity of the same illumination, which is why they are perceived as belonging to an object.

At the same time, areas A to B to C appear progressively higher in brightness and colourfulness. This is the perceptual result of their sending progressively greater amounts of light to the eye, evoking progressively stronger responses at the level of the light-sensitive retina. Similarly, although we perceive the lighter-coloured areas of the floor as being white things, that is, as objects having a white colour of uniformly high lightness belonging to them, the corresponding areas send light of a wide range of intensities to the eye, which we perceive as a wide range of brightnesses. This variation in brightness and colourfulness is perceived as being imposed by the illumination rather than as belonging to the objects themselves. Brightness and colourfulness describe the appearance of the light reaching our eyes from different areas of an object, in contrast to lightness and chroma, which describe the colour perceived as belonging to the object.

The CIE e-ILV defines an object colour as a “colour perceived as belonging to an object” where the word “colour” links to the entry for colour “in the perceptual sense” or “perceived colour”. So, an "object colour" as defined in the CIE ILV is a visual perception perceived as belonging to an object. Although they are perceptions, we perceive object colours to be located outside us in objects themselves, as in the uniform black, white and orange object colours that we perceive to be located in the tiles and cube depicted in this image, even though, and this bit is in italics, even though these objects are physically non-existent. This last observation can help students to accept that the colours that sure seem to be located in real objects, including the real objects around you now, are similarly not actually located in those objects, but are perceptions that we project onto objects.

However as philosopher Clyde L. Hardin famously said, “Colored objects are illusions, but not unfounded illusions”. If these were actual objects, we would undoubtedly find that they differed in the physical property of spectral reflectance. The spectral reflectance of an object is the proportion of light of each wavelength that the object reflects. When we can freely examine an object in daylight, this colour that we perceive as belonging to the object is usually a very good indication of its overall spectral reflectance. I’ll elaborate on exactly what I mean by “overall” shortly.

So within the rectangular image on the left we perceive a pattern of orange, black and white object colours, which I’ve shown bottom right. Yet within the same rectangle we also perceive a pattern of illumination, that is, a pattern of light perceived to be falling on the objects, comprising areas of light and shadow, as I’ve shown top right. Colours perceived as belonging to illumination, called Illumination colours, can be described in terms of hue, brightness and either colourfulness or saturation, though the illumination in this scene is perceived as achromatic or white light, lacking hue and saturation or colourfulness, and varying only in brightness.

Notice that we experience these perceptions of object colour and illumination colour superimposed in the same rectangle, as if the object colour is seen through the illumination. But how do we perceive, for example, which areas are dark because they are occupied by dark things, and which are dark because they’re dimly lit? It may seem to us that we directly detect these variations in illumination and object reflectances, but there’s no way for our eyes to do that. Instead, our visual system automatically, instantly, and seemingly effortlessly resolves or parses the scene into superimposed perceptions of spectral reflectance (as object colour) and of illumination. It’s only because this remarkable capacity of our visual system exists and is generally very effective that we perceive objects as having a relatively constant and seemingly intrinsic lightness and chroma belonging to them, and not just constantly varying brightness and colourfulness.

This ability to automatically and unconsciously parse a scene into colours relating to illumination and objects imparts a degree of constancy to our perceptions of object colour under coloured illumination. We’re likely to attribute the overall pinkish colour of this image to the illumination rather than to the objects, and so to perceive some white and grey object colours through this pinkish illumination colour. This ability is assisted by the quite distinct process of adapting to the colour of the illumination, which would cause the whole scene to appear less pinkish than it would otherwise. Nevertheless, our ability to distinguish objects according to colour is reduced under illumination that appears coloured, and diminishes to zero under monochromatic light.

Our perception of colour has a multilayered quality that is rarely appreciated. Viewed normally, in what is called distal or constancy mode, we have these two superimposed perceptions of object and illumination colours. But we can also view the scene in what is called proximal or painters’ mode, as a flat array of patches of light, the way an impressionist painter might look at the scene. Viewed in this way the light reaching our eyes throughout our field of view exhibits a third set of colours that can also be described in terms of hue, brightness and either colourfulness or saturation.

[23:43] The sixth attribute listed in the CIE ILV definition of perceived colour is saturation, which is defined as the colourfulness of an area judged in proportion to its brightness.

Now the only way for the light from an area to appear less colourful while maintaining its brightness is for it to appear more whitish, so saturation may also be described as the perceived freedom from a white light component of the light from an area. That is, saturation is the perceived proportion of the chromatic component of a light out of the total of its chromatic and achromatic (coloured and white light) components. So here the light from areas A to C increases in colourfulness in step with its brightness, which amounts to saying that it does not become either more or less whitish as it gets brighter, so it stays the same saturation. This relationship holds within limits when light of the same spectral composition varies in intensity, or as here, where a uniform object is illuminated by the same light at varying intensities.

[25:04] I also discuss one additional colour attribute called brilliance that doesn’t currently appear in the CIE ILV, but which is related to the attributes of blackness and saturation. What happens to the colour of an object in a given setting if we increase or decrease the intensity of the light coming from it without changing its spectral composition? If a light stimulus of fixed spectral composition increases in intensity from zero to very high in relation to its surroundings, it passes through a consistent series of stages, appearing black until a “black threshold” is reached, and then passing through decreasing degrees of blackness to a point of zero blackness and then going on to appear fluorent (fluorescent looking) and ultimately self-luminous. The circle on the right face of the cube illustrates this sequence. At low intensity (the same as for the circle on the left face), the circle on the right exhibits a high degree of blackness. At a certain intensity relative to its white-appearing surround, the circles on all three faces exhibit approximately zero blackness. At a somewhat higher intensity (the same as for the circle on the top plane), the circle on the right appears fluorent or fluorescent-looking. At a much higher intensity relative to the environment, the circle appears self-luminous. The brightness at the point of zero blackness, at which the area appears neither blackish nor fluorent, varies depending on the hue and saturation of the light.

[27:07] This introduction to these attributes of perceived colour was the first and perhaps the most important thing that I wanted to get across tonight. The remainder of first paper explores the connections between colour stimuli, colour perceptions, and colorimetry, to illustrate how colorimetric specifications are purpose-built to capture precisely the perceivable, “overall” spectral properties of lights and objects that we perceive as their colours. This was written partly in response to certain misconceptions that have gained traction in some circles, stemming from demonstrations of colour inconstancy sometimes called “optical illusions”, that human colour perceptions do not relate to spectral properties or to colorimetry. But the discussion will hopefully also be of more general interest for its bearing on human colour vision, the rationale behind colorimetry and the fundamental question of “what is a colour?”.

[28:12] This section begins with a discussion of colours of light. We all know that Sir Isaac Newton showed that the reason why light forms what he named a spectrum when it passes through a prism is because it is broken up into a series of components (we would now say different wavelengths) that appear different colours. Yet when we see a light compounded of different wavelengths, we don’t experience multiple colour perceptions corresponding to these multiple components; we see a single colour.

Crucially, Newton showed that the colour of an isolated light can be predicted from the overall balance or what he called the “center of gravity” of its spectral components in a two-dimensional circuit of directions of bias relative to white light, which he explained with his famous colour circle. The hue of an isolated light can be predicted from the direction of bias of the “center of gravity” relative to white light, and what he called the “fulness or intenseness of the Colour” or its “distance from whiteness”, now called its saturation, can be predicted from the amount of bias.

Another way of saying this is that the colour of an isolated light is the way in which we perceive the overall balance of its spectral components relative to that of light perceived to be white, such as daylight.

So whitish orange as the colour of an isolated light is the way in which we perceive an overall balance of spectral components biased in a certain way relative to daylight, and white as the colour of an isolated light is the way in which we perceive an overall balance of the same spectral components similar to that of daylight (top middle). By this principle, spectral orange, the saturated orange we perceive in the spectrum is the way in which we perceive a very strong bias in the direction of certain wavelengths relative to daylight, rather than a property present in those wavelengths themselves that our visual system “detects”. (A saturated orange light might consist of those orange-appearing wavelengths, but it might also contain little or no light of these wavelengths; for example, the saturated orange spot Y at top right of your screen emits relatively little light of such wavelengths.

Daylight is defined as the visible part of global solar radiation, meaning “combined direct solar radiation and diffuse sky radiation”.  Notice that the spectral power distribution representative of daylight (D65) is not flat throughout the spectrum as you might assume, but has a broad low peak in the short- wavelength part. (CIE Illuminant E on the right, which does have a flat spectral power distribution, appears somewhat pinkish as an isolated light). You might have received the impression that we perceive daylight as white light because our visual system detects equal amounts of all wavelengths; it doesn’t, it perceives daylight as white light because it is attuned to perceive this somewhat uneven spectral distribution as white.

It’s remarkable that throughout our lives our visual system continually attunes itself to compensate for the physical changes occurring in the eye so that typical daylight is perceived as colourless or white light. This adjustment occurs very slowly as our lenses turn brown with age, and relatively rapidly when we get new ones. It makes sense intuitively that it would do this because, as we saw earlier, the more colourful the illumination of a scene appears, the more difficulty we have in distinguishing objects according to their colour, so being attuned to perceive daylight as lacking hue maximizes our ability to distinguish objects by their colours under typical natural conditions.

We now know that Newton’s directions of perceivable spectral imbalance form a two-dimensional circuit because our colour vision depends ultimately on the responses of three receptor types, called L, M and S cone cells. The often-expressed idea that our cone cell classes work by individually detecting long, middle and short wavelengths (or worse, by” detecting red, green and blue”) is incorrect. Our S cone cells respond to a relatively narrow range of short wavelengths, mainly those that appear violet and blue, but our L and M cone cells respond to all and almost all wavelengths of light respectively. Also, the specific wavelength of a light hitting a cone cell affects the probability that the cone cell will respond, and thus the strength of the overall response of that cone cell class, but the cone cell provides no information on the specific wavelength of the stimulus.

What is true is that the three cone cell classes divide the visible spectrum into long-, middle- and short-wavelength bands, in each of which one cone cell class responds more than the other two. Comparison of the responses of the cone cell classes provides information, not about the individual wavelengths present, but about the overall balance of radiant power between these three broad wavelength components of the light. This comparison occurs in the retina by the process of cone opponency (Figure 3, middle and right). Some neurons compare M cone and L cone outputs to form an L vs M cone-opponent signal, while other neurons compare the S cone output to the combined outputs of the L and M cones to form an S vs LM cone-opponent signal. The L vs M and S vs LM cone opponent signals create a 360 degree circuit of possible combinations that we ultimately experience as the circuit of hues, but the red, yellow green and blue hue perceptions mentioned in the CIE definition of hue do not coincide with these cone-opponent axes, and how they arise remains disputed. The role of our three cone cell types is widely covered, if somewhat misrepresented, in popular accounts of colour vision, but the equally important process of cone opponency is much less commonly discussed, leaving a gap that’s currently filled by a collection of inventive but misguided speculations about cone signal processing that I described a few years ago in a piece called “The YouTube theory of colour vision”.

It’s by means of this combined cone and cone-opponent system that we can detect variations in the overall balance of radiant power in relation to a circuit of directions towards long, middle, short, and long and short wavelengths respectively.

So we can now elaborate slightly on Newton’s discovery that “The colour of an isolated light is the way in which we perceive the overall balance of its spectral composition relative to that of daylight; by saying that “overall” here means at the level of its long-, middle- and short-wavelength components, as detected by the human visual system”.

 Of course, perceived colours can be influenced by a variety of factors in addition to the spectral properties of the stimulus, as acknowledged in Note 1 appended to the CIE ILV definition of perceived colour (CIE e-ILV 17-22-040). Nevertheless, despite the importance of these other factors, it would be going too far to deny a connection between colour and spectral properties. It is quite reasonable to say that in many ordinary circumstances, variations in the spectral composition of light at the level of its long-, middle and short-wavelength components are detected by the human visual system and perceived as different colours. This of course is why we’re all looking at machines that work by emitting different mixtures of long-, middle- and short-wavelength light.

[37:55] So, what about object colours? As with colours of lights, our perceptions of object colour are not based on instrumental measurements but on the responses of the human visual system and are therefore shaped in part by the characteristics of that system. Since we can only perceive spectral reflectance at the level of its three main spectral components, the colour we perceive as belonging to an object when it is freely examined in daylight, is the way in which we perceive its “overall” spectral reflectance, “overall” again meaning at the level of its long-, middle- and short-wavelength components, as detected by the human visual system. The hue of the colour is our perception of the direction of bias of this “overall“ spectral reflectance, the chroma is our perception of the amount of bias of this “overall“ spectral reflectance, and the lightness is our perception of the proportion of light that the object reflects, as always subject to the usual individual and environmental effects on colour perception, and the workings of the human visual system. (Notably, the response of our visual system tapers towards each end of the spectrum, so wavelengths near these extremes have a decreasing influence on our perception of colour).

[39:29] Colorimetric specification of lights and objects can be an area of confusion in the broader colour community, and I’ve sometimes encountered the view that colorimetry is all very well for technology but has nothing to do with human perception. In fact, colorimetric specifications of lights and objects are purpose-built to represent for practical purposes just those physical differences that we perceive as colour differences by ignoring physical differences that are not perceivable to human colour vision.

You may have encountered the CIE x,y diagram previously in various contexts, especially in comparing the gamuts of different RGB colour spaces or different devices. It’s not the latest but is still the most familiar descendent of Newton’s colour circle. As was already implicit in Newton’s circle, physically different mixtures of spectral components can evoke the same perceived colour if they have the same “center of gravity”, or overall balance of spectral components. The three spectral distributions on the left appear white as isolated lights: CIE illuminant D65, representative of noon daylight, CIE illuminant F7, representative of a fluorescent illumination that matches D65 in colour, and a specific white LED screen adjusted to match these illuminants. Despite their physical differences, these three spectral distributions match as isolated lights because they have the same overall balance at the level of their long-, middle- and short-wavelength components, as detected by our combined cone and cone-opponent system.

Location in the x,y chromaticity diagram represents this “overall” balance of wavelengths, called chromaticity, given the necessary assumption of a mathematically defined “standard” human observer, and our three lights all plot at the point marked D65.

A colorimetric specification of a light represents a class of physically varied lights having a common disposition to evoke a perceived colour, in the sense that physically different lights having the same colorimetric specification would be expected to match in perceived colour when viewed under the same conditions. Thus, a colorimetric specification of a light may be said to quantify what Newton called colours “in the Rays”, meaning the light’s “Power and Disposition to stir up a Sensation of this or that Colour”; our three lights have a common disposition to evoke this or that colour.

A colorimetric specification does not however correspond to a single perceived colour; our three lights should match in perceived colour in the same viewing environment, but this matching perceived colour would differ as this viewing environment differs. For example, all three lights of our might look slightly pinkish in a predominantly green viewing environment.

As Note 1 to the definition of perceived colour implies, it is impossible to predict the perceived colour of a stimulus based on its spectral properties, irrespective of other conditions and considerations. Colorimetric specifications are not intended to predict perceived colour irrespective of viewing environment, and their failure to do so is not a failure of colorimetry if its rationale is properly understood.

Luminance is a colorimetric measure of the intensity of a light from the point of view of the human visual system, but like chromaticity, luminance tells us what a light will match, but not how that light will appear. The curve on the left showing the responsiveness of the visual system of a standard human observer to different wavelengths is called the luminous efficiency function. Luminance is the physical power of the light weighted wavelength-by-wavelength by this function. Physically different lights of the same luminance would be expected to match in brightness if their brightness is compared in certain ways, notably by showing no flicker when alternated very rapidly or by finding the point at which they exhibit a minimally contrasting border. But luminance does not tell us how bright a light will appear; these matching lights would vary in brightness depending on the viewing environment, appearing brighter in an environment of lower luminance.

An important complication to the concept of luminance that I discuss in the second paper is that if two lights agree in luminance but differ in colour, they might be perceived to differ in brightness when compared by other methods.

Colorimetry of objects involves colorimetric specification of the light reflected by the object under a specified light source, normally a standard daylight illuminant. For an object that reflects all wavelengths equally, the x,y chromaticity of the light it reflects is the same as that of the illuminant, and other chromaticities result from various directions and degrees of bias of the spectral reflectance at the level of its long-, middle- and short-wavelength components. Specifications of this reflected light can be plotted in the three-dimensional colour space CIE xyY, formed by adding to the x,y chromaticity diagram the third dimension Y, representing the relative luminance of the light reflected compared to a reference white. Granted the necessary assumptions of a standard observer and a specific daylight illuminant, CIE xyY values quantify for practical purposes the human-perceiver-dependent “overall” property of an object’s spectral reflectance that we perceive as the colour of that object in daylight, which is the colour that we tend to think of as the (seemingly) intrinsic colour of the object.

Each chip in the Munsell system, representing a specific hue, lightness (value) and chroma, is manufactured to embody a CIE xyY specification of the light that the chip would reflect under a specified daylight illuminant. Each of these xyY values represents a class of physically different spectral reflectances of objects that would match in colour to the standard observer under this Illuminant, and for practical purposes would be expected to appear essentially the same hue, lightness and chroma to a colour-normal observer when freely examined in daylight.

[47:22] I mentioned at the start that the CIE ILV defines the word “colour” in two distinct senses. Along with “colour” in the perceptual sense or “perceived colour”, the ILV also defines “colour” in the psychophysical sense, as a specification of a colour stimulus in terms of colorimetric values. When we speak of “colour measurement”, “colour difference formulae”, many “colour spaces” and the 16.7 million RGB “colours” on our screens, we’re using the word “colour” in this second sense.

In defining two senses of the word “colour” the CIE ILV in effect acknowledges that we may wish to use the word “colour” either for our perceptions of colour or for the measurable properties of lights and objects that dispose them to appear this or that colour. Colorimetric values specify the perceivable ‘overall’ spectral properties shared by physically varied lights and objects that dispose them to match in colour to the human visual system. In terms of Hardin’s quote – “Colored objects are illusions, but not unfounded illusions” - “psychophysical colour” specifies for practical purposes the human-perceiver dependent properties of lights and objects that our “illusions” of object colour and other perceived colours are founded on.

[48:56] Part Two focuses on colour perception in greater depth, exploring the main modes of colour appearance and the various perceived colour attributes and their interrelationships. Part 2 is the longer of the two papers and I’ll only try to show a selection of the illustrations tonight. Here we see examples of some of the main modes of colour appearance. Here we see various object colours, which can be subdivided into the surface colours of opaque objects and the volume colours of transparent and translucent substances such as seawater. And here [49:40] we see various illumination colours of light perceived to be falling on an object with a uniform object colour. This one [49:50] illustrates interconnections brightness, colourfulness and saturation and chromaticity. In this illustration [50:03] I’m addressing the distinction or relationship between chroma and saturation, which are very often confused. A column of digital swatches of uniform chroma in the Munsell system can be seen to range from lighter swatches emitting a large amount of relatively whitish light to darker swatches emitting a smaller amount of less whitish/relatively more chromatic and thus more saturated light. Swatches exhibiting similar saturation lie along lines that radiate from near the zero point on the value scale, in contrast to the vertical lines of uniform chroma. For example, swatches A, B and C all exhibit similarly high saturation – they all emit similarly pure orange light - but vary considerably in chroma. The dimension called “saturation” (S) in the widely used digital colour space HSB, which you’ll know from the colour picker in Photoshop, is a simple measure of saturation in the sense defined by the CIE, relative to the maximum possible for digital colours of a given hue. Image areas A, B and C here have the same “HSB saturation”, and that relationship was used of course to paint the cube here.

The diagram on the left [51:43] shows how fields of uniform hue and saturation and varying brightness are readily perceived as uniformly coloured objects under varying illumination, provided that their arrangement is consistent with this interpretation. These sequences of colours follow vertical lines in CIE xyY space but radiate from near the origin in hue-value-chroma spaces such as Munsell and CIE Lab. Somewhat surprisingly they seem to radiate for most of their length from a point a bit below zero in these spaces, which is related to the way the lightness scales in these spaces are defined.

This diagram [52:33] is from a discussion of how colours of zero blackness, which trend close to the top of the Munsell colour solid, vary greatly in lightness depending on hue and saturation, being high in lightness for all whitish colours, much darker for saturated blues, and in between for saturated yellows. And this one [53:02] illustrates the complication to the concepts of brightness and lightness related to chromatic intensity that I mentioned earlier. In both the Munsell system and CIE LAB, the Lab space you know from Photoshop, swatches of the same lightness by definition have the same luminance. So, if we look at any horizontal row of swatches such as the lightness 50 row here, they all have the same luminance, but we might easily judge the high chroma red to be brighter than the grey of the same L*, and we might judge it to match the lightness of the 60 grey or even higher. And yet if we place the L 60 grey alongside the high chroma red, I think we can see that it would need to go darker to reach the point of minimally distinct border with the red, even though the red has a sort of glowing appearance. I discuss in the paper how this perception of extra brightness and lightness associated with high chroma swatches may relate to them being unusually high in relative luminance for their hue and saturation in relation to the gamut of familiar object colours. And finally these illustrations [54:25] are from the discussion of the attribute of hue, which includes an account of various simple hue circles based on different organizing principles.

I’d like to thank my students, and my many colleagues over the last twenty years including the Colour Literacy team for their feedback on my evolving attempts to explain the concepts discussed in the two papers. I’d like to especially thank Robert Hirschler for his feedback and his patience as editor and Professor Mark Fairchild for very kindly agreeing to provide comments on both papers in addition to those of the anonymous reviewers.

So just a few things to finish up. As I said my two papers will be available in the forthcoming Special Issue of the Journal of the AIC on Contributions from Colour Literacy Project team, so you’ll be able to download those if you’d like to read more about any of what I’ve discussed tonight. You can see the work of the rest of the team in the other papers in the volume, and on other things including forums, web pages on various colour concepts and misconceptions and a series of eye-opener exercises for teachers, at the link shown lower right.

You can see more of my own work on my website the Dimensions of Colour, which I worked mainly up until 2017, and there’s an index of links to my more recent work on a page towards the end of that site. Colour Online is collection of more than 500 links to useful websites, texts, videos, software and forums, and along with the two papers is my other main tangible contribution to the Colour Literacy Project so far.

My boss at the National Art School would kill me if I don’t mention my online colour course there, which consists of eight 3-hour live online classes, with time for questions and discussion, plus optional practical exercises that can be done in between classes. There are two more terms left this year, each with a limit of 12 students, and enrolments are already open on the NAS website.

And finally I’d like you to know about the AIC Study Group on Arts and Design, which you are more than welcome to follow on our new Instagram account. And of course there’s the Colour Society of Australia, of which I’m Past President, Vice President and NSW Divisional Chair. We’re a 100% volunteer-run not-for profit organization with an active program of live events and free webinars on all aspects of colour including various kinds of design, art, science, philosophy and education. And I’ll finish by mentioning our national conference coming up on October 12-14 in Western Australia, which will be a hybrid in-person/online event. So stay tuned on the CSA website or social media for more information on that and on our other upcoming webinars and events.


David Briggs, 2023. "The Elements of Colour. Part One: Colour Perceptions, Colour Stimuli, and Colour Measurement. Part Two: The Attributes of Perceived Colour". The two papers are now available as free pdfs at:

Colour Literacy Project


My websites
"The Dimensions of Colour":
"Colour Online":
Index of works:

"Understanding and Applying Colour", online Public Programs course, National Art School, Sydney. Full course outline at

International Colour Association (AIC) Study Group on Arts and Design Prof. Dr. Maria João Durão | Chair; Dr. David J. C. Briggs | Co-Chair

Colour Society of Australia

This page published February 7, 2024.
< Shillito Portfolio Colours of Objects and Colours of Light >