My two previous the Posts provided links to Chapter 12, “Local colour interactions” and Chapter 13, “Colour constancy”, from my book “What Scientists can Learn from Artists”. ThisPost provides a link to Chapter 15, “The other constancies”. Below the two images you will find an edited version of the “Introduction” to this chapter. As with other Posts, if you find that the subject matter interest you, you can click on the link below to the .PDF version of the chapter as a whole. The images illustrate two of the visual perceptual problems with which artists have had to come to terms.
Introductory to Chapter 15
Colour constancy is by no means the only constancy of visual perception. There are many other constancies and all are fundamental to the ability of the eye-brain to make practical use of visually acquired information. Paradoxically, although their name suggests stability, they are responsible for the veritable “shifting sands of appearance” which, in its various guises, constitutes one of the main problems for artists seeking to obtain accuracy in drawings or paintings from observation. This is because they ensure that, when we look separately at any two similar features of appearances whether they be whole objects, parts of objects, sections of contour or colours, there is a very strong tendency to see them as being more similar to one another than objective measurement would dictate – often a great deal more so. Our visual systems upset the measured parameters of external relationship by relentlessly forcing them towards normative dimensions and values. As a result, the constancies involve enlarging and diminishing, squashing and stretching, revolving, darkening and lightening and modifying colour. Any list of the constancies of particular interest to the artist should certainly include (a) size constancy, (b) shape constancy, (c) orientation constancy (d) lightness constancy and (e) colour constancy.
Introduction to the Post on “local colour interactions”
This Post is the second that offers a link to a .PDF version of a chapter from “What Scientists can Learn from Artists”. The purpose of making this more scientifically oriented information available on this website is to encourage readers to go deeper the ideas presented in my practice-oriented books on drawing, painting and creativity (see “Contents List”on the “Posts Page”). Chapter 11, the chapter featured here, focuses on new and unfamiliar things of potential value to artists that can be said on “local colour interactions”, a subject that has featured widely both in books and in the classroom. Its father figure is Eugene Chevreul, the Chemist at the Goblin tapestry works, who was responsible for the phrase “simultaneous colour contrast”, and the best known publications on the subject are by Johannes Itten and Joseph Albers, both teachers at the Bauhaus. Of more recent books covering the subject, I can recommend “Colour : A workshop for artists and designers”, by David Hornung.
With so many authoritative writings on the subject, it might be supposed that I would have little to add, particularly since, as a general policy throughout my books, I have done my best to avoid wasting time on subjects that have previously been exhaustively covered in convincing ways. It is for this reason that my chapter on “local colour interactions” concentrates on subjects that do not appear in the publications of Chevreul, Itten, Albers, Hornung or, as far as I know, of anyone else. What I have to say is based on research triggered by the excellent teaching I received at my art school and issues arising in my own paintings. Its novelty comes either from original or less well know scientific research that deals with matters of potential interest to artists.
It was unequivocal evidence of “induced colour” and “colour constancy” that triggered the realisation among scientists towards the end of the Eighteenth Century that colour is not a property of surfaces in the external world but phenomenon that is made in the head. Once this idea had been digested, it gained momentum and evidence began to pour in to suggest that all visual experience is a creation of the eye/brain combination. This game-changing paradigm shift was to lead, not only to the birth of the science of “visual perception”, but also to fundamental changes in the practice of artists, either when drawing or painting from observation or when seeking control of pictorial dynamics. This is why the “constancies” and “simultaneous contrast dynamics” play such an important role in my books on the practice of painting and drawing. It is also an important part of the reason why I have written “What Scientists can Learn from Artists”, the last volume of my four volume series that explains the science behind so many of the ideas elaborated upon in the remaining three volumes. In going more deeply into the subjects that play such an important role in these books about artistic practices, it plunges us deep into the astonishing nature of the working principles of visual perception. Apart from the sheer wonder this must surely generate, knowing about the ways these determine how we “look” and how we “see” should have a significant benefits for artists: The deeper understanding and appreciation of the extraordinary things that are happening in our heads should help artists to:
Deal with the many practical problems that invariably face them when drawing or painting from observation
Make more creative use of their physical and conceptual tools.
The next Posts I will be chapters from the science volume.
A life changing event
This Post on “colour constancy” is the first from “What Scientists can Learn from Artists”. Its inspiration derives fromEdwin Land’s irrefutable demonstration of the phenomenon of “colour constancy”, which proved to be a milestone in the search for an understanding of a subject that turned out to be of key importance to the understanding of how we “perceive surface”, “sense space” and are “aware of of lighting conditions”, all subjects of key importance to the ideas presented in “Painting with Light and Colour”.
A photo of the equipment used by Land for his epoch making colour constancy demonstration.
A reprise of the “Introduction” to the chapter and a link to a .PDF version of it (no need to read it twice: if you read it below, you can skip it in the chapter)
Links to Posts from “Painting with Light and Colour”, all of which (particularly chapters 7 to 11) have a debt to research that grew out of the colour constancy demonstration.
As explained earlier, a key event in my life was the encounter with Professor Marian Bohusz-Szyszko. The ideas he shared set me off on a lifelong journey of discovery. My first step was to set about testing his seemingly extravagant assertion that it is only necessary to follow two rules to guarantee a good painting:
There must be no repetition of colour on the same picture surface.
All the colours used must be mixtures containing at least a trace of complementary.
After four years of experimenting, I proved, at least to my own satisfaction, that there is a special quality in all paintings that abide by these two rules. It is difficult to describe, but it involves the creation of a sense of pictorial space and harmony.
Fortunately, a troubling paradox arose that would eventually have a profound effect on the development of the ideas presented in this book. It concerned the Professor’s physics-based proof of the invariable variability of colours in nature. This asserted that no two parts of any surface will reflect exactly the same wavelength combinations into our eyes due to:
The complexity produced by the inter-reflecting surfaces
Variations in viewing angles and distances
The paradox is that, if the light reflecting from two parts of a surface can never be characterised by the same wavlength combination, how could artists repeat colours on a picture surface? Even if two regions were painted with exactly the same pigment-colour, how could these appear as the same?
Other people might already have known how to resolve this mystery, but for many years I had no idea how to do so. My first inkling of a solution came after many years, as a result of reading a paper by Edwin Land, the inventor of the Polaroid camera. In it was a powerful demonstration of the phenomenon of “colour constancy” and an attempt to explaining it. What the demonstration showed was a region of colour within a multicoloured display (henceforth referred to as the MCD) being perceived as remaining the same, even when the experimenter changed the combination of wavelengths being reflected from it. I was excited because here were two colours being perceived as the same despite reflecting different wavelength combinations into they eyes? For me it was a eureka moment. However a big problem emerged for it was soon clear to me that the explanation of the colour constancy demonstration suggested by Land was not neurophysiologically plausible. An alternative had to be found. I could never have guessed at the treasure trove of discoveries that would come out of my struggles to provide it. This chapter describes Land’s demonstrations in the context of an earlier attempt at explaining colour constancy. The next chapter introduces our neurophysiologically plausible colour constancy algorithm.
This Post provides a link to Chapter 14 from my book“Painting with Light and Colour”, which is the fourth of the five chapters devoted to colour mixing. Its purpose is to show that all the complications of colour theory proposed and explained in the previous chapters, need not be a barrier to our creativity when it comes to their practical application. Quite the reverse. Below the image of a student at work, is a reprise of the “Introductory” to Chapter 14. If its claims make you want to find out more, click on the link beneath it to obtain a .PDF version of the chapter, which will explain how it can be made easy (a) to mix and (b) to make use of any of the thousands of subtly different, complex colours required for exploring the full extent of colour space.
This Chapter suggests a practical way of getting around the seeming obstacles discussed in the previous chapters. At first sight the method proposed may appear to involve important sacrifices, but upon further investigation it turns out that even its shortcomings can be interpreted as powerful advantages.
This is the second of five posts on colour-mixing. It continues the process of preparing for three more practical chapters that follow. It role is to confront a number of commonly held misunderstandings relating to the ‘colour circle’.
Following the image of a student getting ready to mix a colour, there is a slightly edited version of the Introductory to the chapter and a link to it the whole text.
At my painting school, I give talks on colour-mixing. The chapter which I am publishing in this Post as well as the next three chapters (all of which I will publish in the near future) are based on these. Between them they both flesh out some of the claims made in the previous chapter and provide a sound and practical approach to colour-mixing. My purpose is to provide help with:
Finding a maximum of colours in any part of the colour sphere (as described in the last chapter).
Creating a sense of light, space and harmony in paintings.
The first of my talks concerns colour-mixing by stirring (as opposed to colour mixing by layering, which I will deal with later). Experience has shown that for many people coming to my Painting School for the first time, it is necessary to start my explanations at the most basic levels. Accordingly, I introduce my talk by apologising in advance for going over ground that may already be familiar to some, but suggest that it is better to be absolutely sure of building on common and solid foundations. Also, my decision to repeat the structure of my Painting School lesson means that some of the material that appeared in the previous chapter, is presented again in this chapter, although in slightly different ways.
Recently I was asked if I could post the five colour mixing chapters from my book “Painting with Light and Colour” (Chapters 11 – 15). I will be surprised if you do not find that many of the ideas in them are new, interesting and practical. At the bottom of the page is a link to Chapter Eleven, the first of the four chapters, whose title is, “Colour mixing – definitions and misconceptions”. To whet your appetite (below the image) I have included a slightly edited version of its “Introductory”.
Introductory to Chapter Eleven
Introductory At the outset of my life as an artist, my conception of colour-mixing was of a dry and mechanical subject. I thought of it as no more than one of those necessary basic skills that could easily be picked up along the road. To my surprise, nothing turned out to be quite so routine as it had seemed, and one line of enquiry led to another in a most seductive way. Each new development plunged me deeper into the history either of science or of art, until an engagingly coherent story emerged. The result was a practical understanding of a kind that might be difficult to find elsewhere.
“Most how-to-do-it art books have sections on colour-mixing and there are a number of tomes that offer technical information for professionals. These latter tell us that scientists have understood the physics underpinning colour-mixing theory for a very long time: Certainly they have done so since James Clerk Maxwell’s lecture on colour vision, given at the Royal Institute, two years before the First Impressionist Exhibition in 1874.
In view of the availability of all these sources of information, it might be thought that there is nothing left to add. Unfortunately, this is far from the case. The problem is that:
Too many painters are being seriously misled by the half-truths and even falsehoods which have entered into the stock in trade of popular colour-mixing theory.
Science has far from stood still since the 1870s. Particularly since the 1970s, scientists have been finding out a great deal of new information about how eyes and brains work and, as a result, have arrived at a number of new understandings that could help artists in practical ways, which are not being made use of by the artistic community.
For these reasons and others, it is clear to me that there is a need for the up-to-date approach to practical colour mixing that is supplied by the next chapters.
One approach to clarifying matters is to place the information presented in an historical context. Doing so reveals that:
Some of the best of ideas have been obscured by the passage of time.
The evolution of colour-mixing theory, owes much to parallel development of the histories of science and of art.
The story of when, how and why artists adopted new colour-mixing practices, provides many insights into their potential uses in painting.
With respect to the links between the discoveries of the scientists of visual perception and the practice of the artists, the evidence is usually sparse and often ambiguous. To compound the problem history (not least the history of science) becomes distorted because it is told by people who write with the benefit of hindsight and sometimes from the perspective of a particular prejudice.
It may surprise some people to find how many famous scientists are credited both with more originality and much more fully developed and rounded versions of their ideas than they actually had. A mismatch of this kind may be suspected in the relation between the confusions inherent in the early development of the ideas developed by Seurat and Cézanne and the neat synthesis of them by Professor Bohusz-Szyszko. Similarly it is unlikely that any of the early Impressionists had as clear a conceptual framework concerning the real surface/illusory space dynamic as was eventually to evolve from their pioneering ideas. While these are very interesting areas for discussion, the process of trying to unearth and pin down exactly what the early pioneers had in mind is a work for scholars. The focus of this book is artistic practice and it is the more refined picture as developed by the more recent artists and theorists that are the most useful in terms of their practical value.
We start a short survey of these by providing some basic definitions as used in this book:
Today, I have been editing the entry for aerial perspective in the Glossary for my books. As I was making my corrections, I had the idea that readers of the Posts Page of this website might like a preview of this and future Glossary edits that I feel might interest them. So, to start the ball rolling, here is a slightly expanded version of the one on aerial perspective, with four images added.
Aerial perspective: Between any viewer and the surfaces of the objects at which they are looking lies a portion of the earth’s atmosphere. In addition to the transparent gases that make up air, this contains quantities of dust and other particulate matter (such as the water droplets in mist and, more evidently nowadays, various kinds of pollution). The effect of the intervening atmosphere on the appearance of distant hills and objects seen is well known to us all. We all perceive distant parts of landscapes being bluer and/or greyer and lighter than nearer parts, and objects seen through mist or fog appear progressively greyer and lighter as the distance between them and us increases. Many artists dating back to the Italian Renaissance, most famously Leonardo da Vinci (1452 – 1519) and Claude Lorrain (1600 – 1682), have demonstrated the value of applying these principles in paintings. So convincing was their effect, that they were adopted as “rules” by the French Académie Royale de Peinture et de Sculpture soon after it was founded in the mid seventeenth century. No one would dispute that the images in Figure 1 and Figure 2 produce a sense of progressive distance.
Although the theory explaining aerial perspective is scientifically sound and although it virtually always has an important effect on how we perceive both distant parts of landscapes and objects on misty days, it has no discernible effect on how we perceive objects within an arms length. So at what distance does its influence become apparent? The answer is that it varies according to the composition and density of the particles floating about within it. Thus, on dry, bright days of the kind that often follow abundant heavy rain, when the air has been washed clean, the visibility of atmospheric intervention is minimised, whereas on a hot sultry or misty days, when the air is at its fullest of dust and pollution, it is maximised. In practice, this can mean that on a clear day its effect on the appearance of objects hundreds of meters away may not be discernible.
Despite these facts of appearances, over the years, I have found that many students, when they first come to my Painting School, have been in the habit of adding blue to objects much nearer than that (in one exceptional case, a newcomer, when painting a bunch of flowers in a vase, added blue to the colour of flowers and foliage at the back of the arrangement, arguing that it made it look further away). When I see this being done, it tells me is that the student in question cannot have been looking at the near/far colour/lightness relativities. If so, how can they appreciate the amazing riches of colour relations in nature? They need to learn that rules are not for following blindly, but for testing, a process which will always open doors of awareness.
To further complicate the situation, there are a number of other variables that can result in people perceiving more distant surfaces of a particular colour as brighter and more fully saturated than nearer surfaces of the same colour: In other words, the opposite of the aerial perspective rule. For example, in summertime, the green canopies of distant oak trees that are illuminated by bright sunlight will look lighter and brighter than those of nearer oak trees should they happen to be situated in the shadow of clouds. Also, a boat on a lake that is painted with a fully saturated red that is actually further away from the viewer than a boat painted with a desaturated red will still look further away. If we made a painting of them, matching as best as possible the colours as we see them, the laws of aerial perspective would predict that the further boats would be perceived as being nearer than the nearer boat. Clearly there needs to be a way of depicting distance that has nothing to do with the representation of atmospheric intervention. Luckily there are several of these, including overlap, relative size and texture cues, but only one of them necessitates the use of colour. Unfortunately, this colour-dependent way of enhancing illusory pictorial space appears to be little known, despite its solid foundations in well known history, its sound scientific underpinning and the ease of its practical application. Much of my book “Painting with Light and Colour” is devoted to giving it new life. If you want to know more, read chapters already published on this website and watch for later Posts.
Two examples of minimal effect of aerial perspective, containing contradictions to the laws as exemplified by Clause Lorrain.
John Constable (1776 – 1837) was a great admirer of Claude Lorrain, but he looked more carefully at nature. Figure 3 and Figure 4 are images of two of his paintings that contain elements that are not consistent with a rigorous interpretation of the laws of aerial perspective. See how many you can find?
Now look at paintings by the Impressionists – Monet, Renoir Pissaro, Cezanne, Gauguin, Bonnard, etc. – to see how much they make use of the rules of aerial perspective. Where they do make use of them, was this the result of applying the rules or of looking carefully at nature? According to what is written above, far distant hills should always actually look bluer or greyer, but what about landscapes representing the kind of distances depicted by Constable or shorter ones?
Effect of patchy cloud cover on relative brightnessess
Finally to ram the point home, here are three photographs that illustrate how patchy cloud cover can produce contradictions to the laws of aerial perspective.
Figure 5 illustrates an exception to the law. Due to their being brightly illuminated by sunlight, the walls of the distant church tower are much brighter than those of the house in the foreground, which is in the shadow of passing clouds.
In Figure 6 the situation is reversed. The walls of the house in the foreground are now brightly illuminated by direct sunlight and are much brighter than those of the church tower, which is now in the shadow of passing clouds.
Figure 7 shows:
The far house,
The strip of green field in front of it,
The sunlit patches of brown earth in the ploughed field,
as being brighter than,
The near house,
The ribbon of green field in the bottom left of the image,
The area of brown earth immediately above it.
In all three images atmospheric intervention is playing a part, but in Figure 5 and Figure 7, its effects are being obscured in the ways described.
The purpose of this Post is to provide a link to Chapter 10“Illusory pictorial space and light”, from my book, “Painting with Light and Colour”. This provides a simplified explanation of the science behind the ideas developed in earlier chapters concerning ways of creating and/or enhancing effects of illusory pictorial space by means of using mixtures containing small proportions of complementary colours. In the process it explains why the same method can be used to create harmony in paintings. It also explains why colour repetition has the potential, not only to produce visual discord, but also to generate optical excitements.
Members of the University of Stirling Vision Group
In many places in my books, I acknowledge the importance of the role of colleagues from University of Stirling in the development of the new science-based ideas put forward in them. In particular I mention cooperations with scientists from various departments who later were to join me in the University of Stirling Vision Group. The most important of these were:
Alistair Watson (Physics, psychology and computer imagery).
Lindsay Wilson* (at the time working on aspects of visual perception).
Also, although Peter Brophy* did not join our group, he was an ever-available and important source of information on the biochemistry of the brain.
The founding of the Vision Group.
It was in the Autumn of 1984 that Alistair, Leslie and I took the first steps in the setting up of the University of Stirling Vision Group, which was to have many meetings attended by the above named colleagues and other members of the various interested Departments. Its starting point was a package of ideas developed by Alistair and myself, and two core algorithms based on them, produced by Alistair. These were:
A “colour constancy algorithm“, capable of modelling both spatial and temporal colour constancy, which was inspired by our interpretation of how this phenomenon is achieved by human eye/brain systems. As a preliminary step to achieving this main objective, the algorithm has to pick off the information about surface-reflection. Since it was obvious that the reflected-light contained information, we speculated upon how it might be used by the eye/brain. Due to my interest in picture perception, we focused on its potential for computing surface-form, in front/behind relations, and the wavelength composition of ambient illumination.**
A “classification/recognition algorithm”, based on our interpretation of how human eye/brain systems achieves their primary task of enabling recognition.***
We could not help being excited by the early tests of these algorithms and the speculations concerning their potential. In our enthusiasm to push matters further, Alistair suggested we should seek the help of other researchers, particularly ones with expertise in:
Mathematics and computing.
Visual perception with special reference of visual memory.
It was at this juncture that, having decided on a name for what we were hoping would become a collaborative group, we contacted Leslie Smith for his mathematical and computing skills. But this was only a start. Once Leslie was on board, we approached Bill Phillips, whose long standing interest in visual memory had led him to take the plunge into the recently emerging domain of neural networks and learning algorithms. After many Vision Group meetings, much sharing of ideas, many hours spent working on implementations of algorithms, and the writing of a number of working papers, we decided to submit a suite of five grant applications to the Science and Engineering Research Council, who had let it be known that they were looking for groups of researchers working on the use of computers to model the functional principles of neural system. The stated aim of the SERC was to set up a small number of “Centres of Excellence” in this domain. Not only were two of our grant applications accepted (one submitted by Bill Phillips and one submitted by Leslie Smith), but also our university was encouraged to create a brand new Centre for Cognitive and Computational Neuroscience . This empire absorbed the University of Stirling Vision Group which ceased to have an independent existence. Its coming into existence also coincided with my departure from Stirling on my way to founding my Painting School of Montmiral, where I intended to put theory into practice both in my own work and in my teaching. I also had hopes of confirming and, with any luck, extending the theory. Also after leaving Stirling University, Alistair and I were founder members of a small software development company which used ideas developed within the Vision Group as a basis for creating an image manipulation tool. ****
* The links to Bill, Leslie, Lindsay and Peter relate to their current status. Alistair, Karel and Ranald all retired or died before the Internet became the essential information source it has since become.
** My book is full of examples of how fruitful this speculation proved to be.
What we see is not light, but experiences created by neural networks within the eye and the brain (what I often refer to as “eye/brain systems”). Although it is true that the visual world that we know could not happen in the absence of the patterns of light that enter our eyes, it is only made manifest to us as a result of what is going on inside our heads. This Post provides a link to Chapter 9 of my book “Painting with Light and Colour”, which describes two demonstrations that show just how great can be the difference between an image predicted on the basis of readings from a light meter and the one we actually experience.
Edwin Land’s demonstrations
The demonstrations have a personal importance because they played a key role in the story of my quest to explain the paradox inherent in the dogmas of Marian Bohusz-Szyszko(explained in Chapter 2). They were devised by Edwin Land, the famous inventor, as a part of his investigation of the phenomenon of “colour constancy”.
For another relevant source of information on Land’s demonstration please consult “Land’s colour constancy demonstration”, an edited version of a chapter from my book “What Scientists can Learn from Artists”. You might also want to read the original article in “Scientific American, December 1977”. In this Land explained what he described as his “Retinex theory of colour vision”.
Other chapters from “Painting with Light and Colour”: