Color, Vision, and Art

[Synthetik]

I thought i'd pass on this link to a web exhibit on Color, Vision, and Art. It gets into some details on vision science and visual perception relating to artistic trends in the emergence of modern art. It's a pretty good introduction to these topics and i think would be interesting to Studio Artist forum members. Some of the color theory ideas in particular can be replicated using different settings in the Paint Color Source control panel in the paint synthesizer. I've recently been revisiting the notion of examining the raw statistical components in imagery that work or don't work for abstract art. Combinations or interactions of color, spatial relationships, spatial frequency, texture, etc. A deterministic way of thinking about it would be what are the good perceptual or statistical recipes or ingredients that work well in creating compelling abstract art. The end goal is to have a better understanding of how to specify these sorts of things for generating procedural abstract art in Studio Artist. Anyway, i thought i'd throw this out for some discussion here in addition to anything you may find interesting in the web exhibit.

[Jean Detheux]

John,

What "works in abstract imagery" for one person does not work for another.

In fact, what seems abstract to one person may be a faithful "rendition" of "reality" for the one who did it (Giacometti's "If I could draw what I really see, nobody would know what it was I was looking at." I totally subscribe to that.)

I glanced at the article you pointed our attention to, and immediately was struck by how much its approach is rooted in what Husserl called "ontologizing."
It's fallacious to "explain" perception as if we could "see" its workings without "using" it.
What the "objective world" science attempts to "explain/describe" is always the "result" of a highly -and unavoidable- subjective process, perception.
"Perception is constitutive" said Merleau-Ponty, something too many scientists ignore, willingly or not, and thus fall victim to their own ontologizing (the "objective world" they purport to study is always of their own making, through their perception/intentionality/Weltanschauung).

And if we enter the world of art, we add yet other "subjective" layers to the equation, it seems there are as many color theories as there are painters (nothing wrong with that in my book).

There is however one aspect of color, vision, and art, that is to often ignored, even if it gets touched upon peripherally once in a while. Not only is a color modulated by its surroundings (some of that is shown in the above article), but in an image (a "framed/contained event"), the color is also greatly modulated by its sheer placement on the picture plane.
The same applies to the "impact" of a shape, it is also determined by its placement in the picture plane.

I did talk about that in this exchange a few years ago.

I also wrote a great deal about perception, and in some detail, in my Notes from the Underground articles, most specifically in parts three and four
Part four especially attempts to make clearer the immense impact our "intentionality" has on our perception. The full title of that article is "Knowing enough about seeing to let 'the other' do the drawing." (It was cut off because of formatting limitations at AWN).
Can't spend much more time on this (really busy right now), but I'll add this: Philip Guston once remarked that he could "at a glance tell what any of Rembrandt's students was about, but had no clue whatsoever what he, Rembrandt, was actually doing."

Being a graduate of an excellent Belgian art school (Académie royale des Beaux-Arts de Liège), where we were spoon-fed all sorts of theoretical stuff (composition, anatomy, etc.), I can vouch for what Guston is implying: the masters have it in common, this "not fitting within what theories can define."

It all seems to boil down to "groping in the dark," working intuitively, sometimes clutching at a few theoretical bits, yet often, if not always, needing to drop even that feeble help if one is to make any "progress."

The key seems to me not based on what one knows, but on how well one can function when no longer knowing what to do.

[Synthetik]

While it may be true that what works for one person does not work for another, i think you can learn something by understanding what is working or not working in an abstract image when it does or doesn't work. And if you are going to be generating procedural art, then at some level in the procedural system you are working with that knowledge is going to have to be embedded in the system in some way.

And the aspects of visual perception and visual processing discussed in the exhibit are very real. The human visual system processes imagery in some very specific ways. Everything you see is a function of the processing going on in the visual cortex of your brain. Ultimately the emotional quality you derive from an image ends up being influenced by other aspects of your memory, personality, life experience, etc. But the raw visual components that make up the perceived image are processed in some very specific ways in your visual cortex. Having some understanding of what is going on in that processing can be a powerful tool to an artist.

I think the same can be said about having some understanding of the technical details behind certain works. If you like the coloring of a particular painting understanding how the artist achieved those effects is something you can then use in your own work. Or you can choose to ignore it. But knowledge really is power. Whether you are talking about composing music or creating visual art, understanding the history of what came before, the technical details of how to achieve certain effects, and how people perceive the world are very useful things.

Some of the technical details associated with the color combinations used by the Impressionist painters are directly related to how color is processed by the visual cortex. There's a lot of work that's been done over the years in neuroscience studying the visual cortex in humans and animals to understand what is going on in that processing.

Sure, when you are in the moment of creating something what you are talking about with regard to not getting caught in left brain thinking and being free in the moment makes a lot of sense. But when you are in that moment you are drawing on knowledge that you have learned in the past.

[Dennis Miller]

This all relates to music as well, of course. And I agree that there is great benefit for composers in analyzing the music of others to figure out aspects of "how it works." And the "masters" did that as well (Brahms studied counterpoint his whole life). But of course, that knowledge only takes a composer so far.
A great book to follow up the discussion of the visual system is Margaret Livingston's Vision and Art. She does a lot more with the science than with the art, though.

Ultimately, I assume you can write algorithms to generate color patterns based on various types of relationships (complementary or otherwise). But my guess is, you won't create a masterpiece but one in a few million times. ;-)
Best,
D.

[Synthetik]

Yeah, quite a bit of the presentation in the web exhibit is based on Livingston's Vision and Art book. Semir Zeki is another neuroscientist who has written about art and visual perception in the brain. He also started the Institute of Neuroesthetics which is devoted to the study the neural basis of creativity and aesthetic appreciation of art.

[Jean Detheux]

While it may be true that what works for one person does not work for another, i think you can learn something by understanding what is working or not working in an abstract image when it does or doesn't work. And if you are going to be generating procedural art, then at some level in the procedural system you are working with that knowledge is going to have to be embedded in the system in some way.

John, I think that learning what works, or not, in an image, abstract or not, is done quasi instantly.
What can be "learned" anyway (Guston's quote about Rembrandt and his students).

And the aspects of visual perception and visual processing discussed in the exhibit are very real. The human visual system processes imagery in some very specific ways. Everything you see is a function of the processing going on in the visual cortex of your brain. Ultimately the emotional quality you derive from an image ends up being influenced by other aspects of your memory, personality, life experience, etc. But the raw visual components that make up the perceived image are processed in some very specific ways in your visual cortex. Having some understanding of what is going on in that processing can be a powerful tool to an artist.

That is a prime example of ontologizing, only our culture places the "center"of it all in the brain.

How could we ever "see" the "raw visual components that make up the perceived image" without already being immersed in perception, therefore in subjectivity (to be sure, "subjectivity" is far from being a dirty word in my book).

The "raw perceptual components" are not what we start with, they are what we arrive at, by choice, when we decide to "look at" the situation following specific choices/intentions, conscious or not.

"What do I see before knowing what it is that I am looking at (or 'for')?"

Or again: "How do I see? Let me count the ways."

I do know a bit about the physiology of the eye, I used to lecture on that, but I also know that what it shows is very very "thin" compared to the actual experience of seeing.
And as we (always) start in experience, I believe it is important to keep the examination of the perceptual processes rooted in that realm of experience, especially if we are looking at the relevance of the knowledge of those processes as it pertains to art, especially to art "making."

I think the same can be said about having some understanding of the technical details behind certain works. If you like the coloring of a particular painting understanding how the artist achieved those effects is something you can then use in your own work. Or you can choose to ignore it. But knowledge really is power. Whether you are talking about composing music or creating visual art, understanding the history of what came before, the technical details of how to achieve certain effects, and how people perceive the world are very useful things.

If power is what matters, count me out, am "after" something else.

As for knowing how it was done, there is a totally different approach: focus on "your" perception of your "ordinary" reality, and as you open up to the elusiveness of what you took for granted, you will need to try out all sorts of things to keep with that elusiveness -as elusiveness-, as it gives itself, and when doing so, you'll find yourself in the company of many of our masters, who were not doing anything else.

Let the technique do the driving, and it will shape your perception, irremediably cutting yourself off from your best resource, your own "not-knowing."

That is the major error made by most art schools, especially when they entered the university system, they went the "power"way, privileging the "conceptual" over the experiential.

Talking about "how people perceive the world," there is an absolutely fundamental, basic process in/as consciousness: we constantly shift from "Where is it?" to "Aha! There it is!"
From stasis-reclaiming to flux-reclaiming.
Consciousness works by either "entering the world," or stay short of it.
Can hardly go more basic than that.
Where are the features in software or, for that matter, where are the art classes in art schools, that would help people become aware of that most basic of processes, and "ride" them?

Some of the technical details associated with the color combinations used by the Impressionist painters are directly related to how color is processed by the visual cortex. There's a lot of work that's been done over the years in neuroscience studying the visual cortex in humans and animals to understand what is going on in that processing.

Look at the series of paintings by Monet (there are several) and you'll notice that, unlike Cézanne, he kept the "grid" of his motif constant, that no matter what the colour changes were, they never affected the actual "structure" of his motif.

Cézanne, on the other hand, knew/saw that a shift in colour is a shift in shape, and he could not (to his great credit) dissociate colour from form.

It may suit the scientific minded to dissect experience, but artists are (or at least, should be) about the wholeness of it. Once again, let me repeat what Guston said: "I can see at a glance what his students are about, I have no clue what it is he was doing."
And Guston was one of the most intelligent, articulate and learned artists there ever was (and one hell of a painter to boot).

Sure, when you are in the moment of creating something what you are talking about with regard to not getting caught in left brain thinking and being free in the moment makes a lot of sense. But when you are in that moment you are drawing on knowledge that you have learned in the past.

Precisely, the "phenomenological reduction" is one way of "bracketing" all that one knows, so that one can cater "to the appearing as it appears."
If what I "understand" of perception is valid, then we never cease from being "in the moment of creating something," usually calling it "reality."

I have learned and taught artistic anatomy (morphology), linear perspective, classical composition, and I can attest to the fact that the only thing I can derive from all that which has any worth for the continuation of my work is that, at least, I know what not to "use."

I understand you are coming at it from a different angle, you are creating this magnificent application, Studio Artist. But from the point of view of this practicing artist, knowledge is not necessarily power, or power is not all that desired.
Art is also (about) a process of attrition, just like "real" life.

[Synthetik]

Maybe a better way for you to think about it is to split tool making from the use of tools. So view the discussion i'm really interested in pursuing as one of looking at different techniques that could be used to make interesting paint tools. Or in extension, tools that make certain kinds of texture, line, shape or form, color relationships or other kinds of statistical or perceptual properties of imagery that can be quantified in some way so that those desirable characteristics can be generated by the tool. I kind of think you are more focused on the use of the tools in the creative moment in your discussion, which is a whole different thing.

When viewed in the guise of putting together interesting paint presets there's a wealth of information in the web exhibit i mentioned that could be utilized to put together interesting painting tools ( i think). Studio Artist has a lot of different controls that let you try to reproduce some of the different marking or color combination techniques discussed in the exhibit. Even though i'm familiar with a lot of the material every time i read through something like this again i get new ideas to try to add to Studio Artist. Or new things to try to program to build new tools based on features already in the program.

And then when you start to talk about constructing procedural art systems the whole name of the game is that there needs to be something driving the construction of the system since the procedural process is not interactive when it is generating the imagery. It's not like interactive drawing or painting where a human viewer is in the physical loop of the image construction, you setup a procedural system that generates imagery and then let it run automatically. So somewhere in the system there needs to embedded some kind of statistical relationship associated with the characteristics of the imagery it generates so that people will find the results visually interesting. Either that or the system itself needs to understand enough about visual perception so that it can make the distinction as it is running. Or some combination of the 2. I hope the distinction between this way of generating art vs live interactive painting makes sense, and why some kind of concrete analysis of what is visually interesting or compelling vs not makes sense and indeed is essential when designing this kind of system.

To give a real specific example, in my own work recently i'm wrestling with trying to come up with some new approaches for coloring abstract procedural art. One technique i fall back on a lot is grabbing an existing image i like the coloring in and auto-generating a color palette from that and then using random palette source mode. But in some sense that's kind of cheating. Another approach is to build a palette from a specific color using color theory ideas like opponent or analogous coloring.

When i build MSG presets i often work on different patching techniques that introduce 2 or more different palettes into an image so that rather than the entire image being a homogenous set of colorings the processing works to break the image up into spatial regions that are colored differently. Again, this is all about trying to introduce certain kinds of spatial or statistical properties into the final imagery generated from the procedural system that has characteristics i think are going to lead to visually interesting images. And you can extend this kind of discussion to all of the other visual aspects (like shape or texture or line flow) that are components of that image. Of course when you run procedural systems like this the end result may or not be something interesting, and typically the system is constructed so it generates an infinite variety of different random images in the statistical style so you can sift through them until you get something that really works well.

I'm not sure what you mean when you refer to the grid of the motif for Monet vs Cezanne, so if you could explain that a little better i'd love to understand what you are talking about here.

[Jean Detheux]

Maybe a better way for you to think about it is to split tool making from the use of tools. So view the discussion i'm really interested in pursuing as one of looking at different techniques that could be used to make interesting paint tools.

True, but if we are talking processes, the shift in modes of perception are even more "real" than texture, line, shape and all. They are the very means by which we can grasp texture, line, shape, etc..
And this is where things get really interesting: the shifts point to a constantly dynamic process/experience, "the world won't sit still, neither can we" and yet, we build theories -and therefore tools- as if it/we were "still."

When viewed in the guise of putting together interesting paint presets there's a wealth of information in the web exhibit i mentioned that could be utilized to put together interesting painting tools

When I (try to) read that kind of paper, I find myself facing a lot of material that is extremely limited in dealing with human experience, and yet, it purports to describe/explain human experience. It is riddled with "ontological assumptions," "averaging" the mechanics of experience to the point where that average no longer resembles anything human.

Let me use the following example borrowed from Sufism: one evening, an old man is walking home, drunk. While approaching his house, he pulls out his keys and drops them. He starts looking for them, but it s night, and very dark. Later on, a friend walks by, an seeing the old man on his hands and knees, he asks him what he is doing there. The old man replies that he is looking for his lost keys. The friend joins him in his search, and after a few minutes, he says: "Are you sure you lost the keys here? I can't find anything!"
"No" said the old man, "I lost them over there where it is very dark, so I am looking here where there is some light."

This is not said in jest John, art history shows us that much of the greatest figurative art was done by cultures who knew nothing about anatomy (they "fed" on something else). I believe we do not need to "know" about "texture, line and all" in such a hardcore way in order to make art.
In fact, even with my baggage in art, even "academic art," I suspect it is probably the opposite.

I want to clarify that: it is probably OK to know about the workings of the eye, the brain and all, providing one is rooted in personal experience ("know thyself first").
But it is erroneous to believe that one can connect with personal experience by way of that knowledge: it fills the emptiness we need to enter fully -as emptiness- in order to connect with that level of experience.

"To acquire a certain kind of knowledge, you have to become a certain kind of man." (Husserl)


And then when you start to talk about constructing procedural art systems the whole name of the game is that there needs to be something driving the construction of the system since the procedural process is not interactive when it is generating the imagery.

I do understand the distinction, and I already said so in my previous post when I said "I understand you are coming at it from a different angle. . ."

But given the way Studio Artist functions, the fact that you have already integrated in it a form of artificial intelligence based on cognitive science, makes me say that I do not know why that would limit itself to "deterministic science?" Why can't it also feed on what we already know about the (for example) "valenced" nature of lived experience?
Couldn't the "action" button trigger two modes of intelligent drawing, with any preset opened to the automatic action, based on a choice made by the user? ("From the inside out," and "from the outside in," just to name two such opposite-complementary modes.)

To give a real specific example, in my own work recently i'm wrestling with trying to come up with some new approaches for coloring abstract procedural art. One technique i fall back on a lot is grabbing an existing image i like the coloring in and auto-generating a color palette from that and then using random palette source mode. But in some sense that's kind of cheating.

Why would that be cheating?
Nobody starts as a blank slate. We are, whether we know it or not, the "product" (or continuation) of all that happened before us.
Marcel Duchamp demonstrated, with his "Readymades," that "to notice is to create" (paraphrasing Piaget here who said that "to understand is to invent").
Surely, picking this or that image to be used as a palette, and to do so amongst the billions of available images, is hardly cheating, or accidental!

Color is such a broad field, it's impossible to deal with it in depth in such a place.
However, here are a few things: my students, especially if having problems with color, were often encouraged to first use black and white only, and to introduce "color" only when they would notice that two shapes would be seen as having the same tonality, but would not merge into one as they would read at a different "temperature" (cool<->warm).

For a couple of years, several decades ago, I limited my palette to black, white, and red. It was amazing how much that limited palette allowed me to do, as long as I could suggest warm and cool "colors," I was able to paint without feeling limited at all.

This above escapes a deterministic view of color, it is not to be found in the rods and cones, nor the optical nerve, not even in the cortex, it is happening at the level of "association" or better yet, "meaning."
It won't show on a graph.

When i build MSG presets i often work on different patching techniques that introduce 2 or more different palettes into an image so that rather than the entire image being a homogenous set of colorings the processing works to break the image up into spatial regions that are colored differently. Again, this is all about trying to introduce certain kinds of spatial or statistical properties into the final imagery generated from the procedural system that has characteristics i think are going to lead to visually interesting images.

You know how much I like Studio Artist, with it and Final Cut Pro alone, I could do most of my work without missing much. But I have not had any use yet for MSG.
The input of the hand is to me essential, even if I use a lot of the "action" function.

However, I wonder if you couldn't build into MSG two opposite-complimentary modes like the ones I mention above?

I'm not sure what you mean when you refer to the grid of the motif for Monet vs Cezanne, so if you could explain that a little better i'd love to understand what you are talking about here.

That's a very vast subject to be dealt with here this way, but I'll give it a first shot.

In the examples below, you can see that Monet approaches a subject matter (a "motif") and first draws it, then "colors" it later.
When he paints a scene at different times of the day, he draws an almost identical drawing on each successive canvas, his colors change with the evolution of the light of day, and for him, that does not imply/need a change of "composition."
His driving idea is based on a taken for granted external world, the light is changing, yes, but the world is not.

Cézanne, on the other hand, is totally immersed in the "oneness of form and color," to him a change in color requires/imposes/creates a change of form.
He almost never draws his motif first, to be colored later, he simultaneously draws and paints: color is form, form is color.
His driving idea is that a change of perception creates a totally different whole (he said some fascinating things about that), and that whole is what he is after. (I did try to explore that in my 4th Notes from the Underground paper)

You can see examples of both painters' approaches below:

Monet:

Rouen

Poplars

Haystacks

Parliament (he lower three images)


Cézanne:

St-Victoire

especially:

Lauves

more examples:

Sous-bois provençal


Riverbanks

Best,

Jean

[Synthetik]

Thought you would appreciate this blog post Jean on artistic creativity and the brain.