Tuesday, December 21, 2010

Some thoughts on eye control, eye path tracking and design

image above from artrenewal.org

James Gurney did a post not to long ago wherein he collaborated with a company that specialized in tracking viewers eye movements. Usually this is done for reasons other than the fine arts What they found was surprising. You can read James post here. There are several other posts from the same week on the site about this experiment.

They found that the viewers eye traveled across images in a way that had little or anything to do with the pathways that artists have imagined they were establishing with their designs. All of the compositional devices used through our art history and those I have espoused on this blog may be called into question by the results of eye tracking. At least in this experiment it seemed that the human eye tracked about images in a much different manner than previously supposed. When I read the post I began to mull over what all this meant.


I believe that the traditional designs do work on a viewer, but they may perhaps do so for different reasons than we have previously thought. Now, I am just positing this idea, I am not so sure I am right , but here it is anyway. (Disclaimer: next week I may think something entirely different, or perhaps not think at all)

Perhaps the design structures that we as artists build under our paintings may not lead the eye about the painting (though perhaps it does ) but it may present an attractive geometric armature that serves as decorative reinforcement to the flow of the painting. The existence of subtly expressed geometry running beneath the image itself may give a "humanness" that is a counterpoint to the randomness of nature. The existence of a mathematically ordered structure may play counterpoint with the naturalism of the artists subject. We may find our desire for order secretly pleased by the imposition of geometric rationality on the chaos of unselected nature.

I know design works, I am less sure of why it works in the light of the new information coming from eye tracking. Perhaps it is no more than a cultural agreement on what geometrical skeletal substructures should please us and which make us feel awkward.

Either way, for now I will go on using the traditional design ideas because they seem to work both in historic paintings and when I apply them to my own work. But it is interesting to think that they may be cultural conventions and not actually scientifically justified.

42 comments:

sharprm said...

They should rig that device and see how the eye moves as someone draws a simple curve. If there is still jagged movement of the eye, may mean the 'mind's eye' also moves differently (which design caters to).

Also didn't these kinds of studies say guys consistently check out the crotch of guys in all photos without realizing it?

Andreas Birath said...

Very interesting. I would even say that I am more attracted by the idea of design and composition as cultural conventions. If the purpose of these ideas were purely scientific i would stop using them. Our job is not to make documentaries, our job is to make poetry.

Deb said...

fascinating stuff... I'll be interested in what you have to say tomorrow.
Would be equally interesting to see if gender, as well as cultural norms, might play a role in this. For example, would women tend to focus more on faces or figures, and men on architecture in the example that James give on his blog?

Philip Koch said...

Exactly what an eye motion tracking experiment shows us isn't so clear.

Over the years I've always found the viewer's eye has a mind of its own in choosing where to look when in a painting. But I totally think the abstract structure of a painting (what we usually call "composition") is critical to the painting's success.

Unconsciously we take in much more than we're aware of at any given moment. The abstract composition is often just sensed by the viewer. It acts like a background rhythm, not unlike an drum beat, energizing and organizing the painting. Some may feel this more a culturally learned response, other's may feel it matches some little understood aspect of our psychology (Carl Jung's idea of archetypes).

What's certain is if you see a painting that grabs you, it has a powerful composition.

JonInFrance said...

Hi Stape, there been a discussion on this very subject over at Paul Foxton's blog - "Learning to see"!!! (in which I mentioned your emphasis on "installing design").

I've been mulling it over - and came to the same conclusion as you present here - the mind likes order. The problem then may be - what kinds of "order" are best..... back to square one as usual, sigh.

Michael Chesley Johnson said...

Arthur Wesley Dow, in his book COMPOSITION: UNDERSTANDING LINE, NOTAN AND COLOR, says that design can't really be taught; it must be absorbed by looking at lots of good art. I'm paraphrasing, but it's an interesting idea. I think your idea is right on; a hidden, mathematical armature might be what Dow was sensing but didn't see.

Ayal Pinkus said...

Actually, I have a theory about this. I think it has to do with primal urges like survival or procreation. Your eyes will rest on a dinosaur in the shadow in a painting, or on the breasts of a woman (even though all the detail and contrast is in her face).

It might also explain why dinosaurs are popular with children, and skinny mags with grown-up men, and why monsters and scantily-clad women are popular in fantasy art.

I had an interesting experience once. It was a busy day on the streets of Amsterdam, and a friend of my sister put huge fake breasts underneath her dark sweater. We were sitting by the wayside laughing at all the people who could not keep their eyes off of her breasts. Now the interesting thing was that all the detail and contrast was in her face, not in her sweater...

mariandioguardi.com said...

Well, mathematic expression is an expression of a proof..one might say a truth.

And so when an underlying truth presents itself in a composition it is re agonized on many levels. But certainly it's not the whole story .

Maybe the eye wandering...and the composition are actually separate issues. And then there is effects, high contrast , warm -cool colors - etc.. Maybe created effects are are the third leg (component) of the stool (painting).even Jackson Pollack worked with in an order.

The real test in Amsterdam would have been to paint the woman's face red and see where eyes went.

Paul F said...

I think there's two separate issues that are perhaps getting confused a little here: one is leading the eye through compositional devices, the other is good design.

What Jim Gurney's posts showed primarily is that the eye will first scan around to find the overall context of the picture, the gestalt if you will, and will then focus in on the elements that the viewer finds psychologically interesting - faces and figures in particular, or stuff you can hold and use, meaningfully interact with.

The part of traditional thinking about composition that this calls into question is whether we can lead the eye of the viewer in a track that we'd like it to follow. Apparently, we can't.

The experiment doesn't call into question the role of good abstract design in making a picture work though, that's a different issue.

Some really interesting experiments have been done previously on this by a psychiatrist called Yarbus. There's a pdf translation of his book on the subject, with eye tracking experiments on a Repin painting here.
I must credit Jim Gurney with putting me on to those in the first place. He's obviously done quite a bit of research himself on this.

Michael, with respect I think you've missed the point of Dow's book. It's actually a series of practical exercises to be worked through which are designed to develop a sense of design and 'good spacing.' He does talk about learning from masters of design, but his emphasis is then to make variations of these in order to learn yourself what makes them work. I like his very practical approach.

Dow inherited the bias of the Aesthetic movement that stressed irrationality and intuition. He doesn't miss the existence of geometry, he comes out very strongly against it in design and champions the more 'organic' Japanese art, also inherited from the Aesthetic movement. I'm not saying he's right, but like everything I think it's a case of taking what is most relevant to you and working with it. For what it's worth, I have a feeling that there's a lot of merit in his approach and I intend to do some practice with it. But that doesn't mean that I'd necessarily eschew any kind of geometrical design, I don't think the two approaches are mutually exclusive.

Plein Air Gal said...

Fascinating stuff!
What I find most interesting is that MY eye didn't follow the track of EITHER viewer, although I did eventually reach each of the same points. I think taking in the overall "lay of the land" first is a given and where the person's eyes go next in any painting depends on the person and their own preferences ... eg those who prefer scenery and vistas would go to the distant mountains, those who love architecture would be drawn to buildings, and those who are "people persons" would go to figures first. The trick is to keep the eye moving and to explore outside the area that most attracts any given viewer ... and that's where installed structure plays the important role. I'm thinking it actually doesn't matter much WHERE I look first in the painting and in what order, as long as I (eventually) see all of the elements and can recognize the main subject in its context.

Susan Roux said...

It's very interesting. I wonder if this visual path differs greatly between artists and non-creative people?

I have to say, I used to ask my mom where her eye went when she looked at my work. She couldn't identify it so I began to tell her exactly where I thought it traveled. She agreed and laughed wondering how I could describe her visual path. I was purposefully painting in connectors while tweaking it in the end. I don't always do that anymore, but it would be interesting to have this test run on some of those paintings where I imposed a path. It's very interesting to think about...

Antonin said...

I mean it is just obvious that a painting need design to work. Maybe the visual path is just a poetic concept. Poetry has a "mathematical" rhythm design but is meaning and understanding goes beyond it. I think that is the same with a painting. We are poets !

mike rooney studios said...

i saw that tracking study on gurneys blog too, and totally believe it. the thing was basically saying we look for the biggest contrast first, the second next, third etc. i believe that and not alot of "composition/eye movement" do's and dont's we've been told for years. i personally dont believe because the bow of a boat is pointing off the canvas that my eye goes off the canvas. c'mon. some of that stuff sounds good in theory but the eye tracking studies refute it.

Rubysboy said...

Perhaps it's the 'mind's eye' that artists are leading. Viewers construct a mental image based on successive fixations that give only a partial view. Doubtlessly different viewers construct their image in different ways (with different fixations). In what order they do this may not matter. So perhaps it's not the eye, literally, that painters lead, but the visual system that makes sense of these snapshots.

robwood said...

Great subject! A related topic that Duke researchers studied recently is the manner in which the brain sorts out what it is looking at. They found that the visual part of the brain is constantly comparing its assumption about what it is "seeing" to the image formed by the eye on your retina. It goes something like this: you are gazing at a vertically oriented, top-heavy object. Your brain says, "building? human? Light pole? etc?" Within milliseconds, it compares the object to stored versions of these guesses, and quickly lets the brain "see" that you are in fact looking at a tree. This would suggest that the way a viewer wanders through a painting may be partly determined by their experience. Do they have memories for comparison? Can they put this in meaningful context?

It may also relate to why, when a passage in a painting doesn't "read right", it throws everything off balance, because the visual brain is stuck in a sorting mode and can't move on to a "wandering" mode.

Or maybe I'm just reading way too much into the Duke work! You may wish to decide for yourselves.
Here's a link to an article:

http://www.newsobserver.com/2010/12/20/871665/science-briefs.html

Mike Thompson said...

There are enough successful paintings out there that violate major rules of composition that it is obvious that there is more to it than X = A + B + C. It is more probable that it is more like a root mean square rule where X = SQRT(A*A + B*B + C*C) or some other equation that has a series of important factors and coefficients. If the final result is greater than an ''OKness'' threshold, we say it is a good painting.

Give 100 artists the same five items to paint in a still life and you end up with dozens of basically different compositions but each one is better than some threshold that makes it pleasing. Most of those painting will probably use some standard compositional rules but the odds are that there will be some good pictures with pretty blatant violations of the rules.

What I found interesting was not that the eye scan paths were so different but that the hot spots were statistically important regardless of how the eyes got there. Maybe the new thinking should be how to make those hot spots a little hotter - a skimpier bikini or a juicier apple, if you will.

If I recall correctly there used to be a book about subliminal images used in advertising. The teaser to get you to buy the book was a picture of a glass of melting ice that was supposed to be sexually suggestive to your subconscious mind. How many hours did the photographer spend moving those ice cubes around to control the movement of your eyes in the ad? More likely he poured out two fingers of scotch, drank it down, and snapped a couple dozen pictures while the cubes melted until he got a good one.

Roberto said...

Dearest Stapey-

Personally, I don’t think eye tracking has anything at all to do with design, composition or how we ‘perceive’ an image. If we had little scanners in our heads, instead of eyes, our scan tracks would start at the top of the image and scan across from L to R, and move down the image to the bottom.
Our amazing eyeballs, and their scanning-aps, receive the input and our lump of sentient-meat creates a virtual image, which we experience as a beautiful (or not) painting, photograph, or dinosaur in the bushes. Part of what our brain/minds do is recognize objects and faces (and crotches and breasts); We then have an experience about that (which is where the aesthetics or design comes in), and then make up a story about it that we believe; (and not necessarily in that order).
If design weren’t important we would not recognize it in nature, or compose it in a camara’s view-finder, or impose it on a painting. -RQ

Lucy said...

Speaking of beauty, I don't care how my eye is scanning it, but the snow nocturne is sublime!

The study fails to disclose how the viewer FEELS when looking at the painting. However the eye scans it may have nothing to do with how the design of a painting impacts visual pleasure.

John D. Wooldridge said...

I'm very much in line with Mike Thompson on this topic. I think the most telling post in James' series is the third where the scanpaths of multiple viewers were agglomerated to clearly display the hotspots. In these, I think we see the abstract design that James imposed early in the proces or at least glimmers of it.

Yes, I think it is interesting that statistically speaking, the viewers fixated on similar regions and appear to move between these regions. The order of movement is not identical but you can clearly see statisical "highways" of movement between the hotspots. Perhaps we are not designing for the individual experience but more for the statiscal experience of a multitude of viewers.

I would also find it fascinating to see this same experiement carried out over longer durations. I suspect that the early time frame eye jumps are the act of establishing the order, finding those important regions. My hyposthesis is that as viewing time increases the viewer's eye will calm and flow much more seamlessly between the hotspots in a fashion more akin to what was planned in the abstract design.

Stapleton Kearns said...

sharprm;
I do NOT!
.................Stape

Stapleton Kearns said...

Andreas;
They are interesting on their own to me.
...........Stape

Stapleton Kearns said...

Deb;
That is a good question.
..................Stape

Stapleton Kearns said...

philip:
The abstract composition is often just sensed by the viewer. It acts like a background rhythm, not unlike an drum beat, energizing and organizing the painting.
Love that! (Jung is a weenie though)
..............Stape

Stapleton Kearns said...

Jon;
Agreed the problem remains the same doesn't it?
................Stape

Stapleton Kearns said...

Michael;
Dow hung out in Ipswich Mass, that's a place I know well. Some of his painting locations are there.
................Stape

Stapleton Kearns said...

Ayal;
I saw that girl today at the mall, and in the grocery store!
..............Stape

Stapleton Kearns said...

Marian;
More interesting still to use a girl with no head at all, and see if any man noticed.
...............Stape

Stapleton Kearns said...

Paul;
Thanks for that. I err towards the geometric, I am not sure why but it makes sense to me.
................Stape

Stapleton Kearns said...

Plein air gal;
The lay of the land search could explain a lot too.
...............Stape

Stapleton Kearns said...

Susan;
That too is an interesting question.
...........Stape

Stapleton Kearns said...

Antonin;
Design is poetic, transcription literal.
.............Stape

Stapleton Kearns said...

Mike;
I can draw plans for some really weak designs that do send the eye out of the canvas, at least I used to think so!
................Stape

Stapleton Kearns said...

Rubysboy;
So perhaps it's not the eye, literally, that painters lead, but the visual system that makes sense of these snapshots. I like that!
..................Stape

Stapleton Kearns said...

Robwood;
I will check that out. Thanks
..............Stape

Stapleton Kearns said...

Mike;
Perhaps but there are lots of badly designed pictures out thewre. I is possible to make them.
................Stape

Stapleton Kearns said...

Robbie:
If design weren’t important we would not recognize it in nature, or compose it in a camera’s view-finder, or impose it on a painting.
That is a good argument. If it didn't matter it wouldn't be so obvious when it was done badly.
................Stape

Stapleton Kearns said...

Lucy:
Thanks. it is also subzero.
..........Stape

Stapleton Kearns said...

John;
Like the bedbug has no wings, but it gets there just the same.
...............Stape

James Gurney said...

Great topic, Stape, thanks--and fantastic comments.

As Paul F suggests, eyetracking discoveries don’t negate traditional compositional theory. Everything we learned about the Golden Section and leading lines and the Steelyard Balance still is useful for making a pleasing picture.

What the evidence questions are categorical statements about how the eye moves in a picture, especially statements that suggest one can control the eye’s movement along contours, or that the eye will certainly enter or exit a picture in a particular way.

Without eyetracking analysis it’s hard to know how we really look at pictures. People are notoriously inaccurate in self-reporting their eye movements, and that may be one of the reasons so many unscientific notions have developed.

There’s still a lot to be learned by scientists in this area. I think what’s missing in the eye tracking pathways is a deep understanding of the role played by peripheral vision.

I would have a hunch, for example, that artists are uniquely skilled at shifting from spot-focus perception to peripheral perception consciously, without significantly moving the position of the fovea. But to my knowledge, this hasn’t been tested or proven yet.

Some of the most exciting studies were done by the Russian scientist Yarbus way back in the 1960s. For those who don’t know these studies, Yarbus showed that viewers look at paintings very differently depending on what leading questions they are presented with at the outset. This has important implications for people writing captions to go with pictures in books or in museums.

I agree with John D. W. that the heatmap information is probably even more useful to us than individual scanpaths. There are other way to crunch the data, to show for example, what most people look at first, second, and so on.

Regarding Susan’s question, yes, artists do see differently, and I reported on this in another blog post: "Do Artists See Differently?"

James Gurney said...
This comment has been removed by the author.
James Gurney said...
This comment has been removed by the author.
James Gurney said...

Sorry--removed duplicate posts.