Pareidolia for the Modern Painter

This is another cross post from our little collaborative project at Pixeldiggers. If you get a chance hop over there and see what everyone else is up to.

Pareidolia (that something like pear-e-DOLL-i-ya) is one of those big fancy doctor words that every artist should know. It refers to the phenomenon of seeing something meaningful in what is in reality completely random. Think: seeing a bunny in the clouds, or a man's face in the moon, or the disapproving face of the Viking God Woden in the scorch marks of your morning toast.

For most people pareidolia is little more than an interesting trick of the eye, but for an artist it's a fantastic way to generate ideas. If you've ever had the experience of sitting down to your instruments and thinking "I just don't know what to draw", pay close attention.

Pareidolia works because your brain is a pattern matching machine, especially for shapes that resemble faces. (If you're interested in how this works, I highly recommend Jeff Hawkins book On Intelligence)

It's so good in fact that when you give it something vague and formless to look at it can't help but interpret what you see as something it's not. This can happen by accident as in the examples above, but you can also provide your eyes with something random to look at on purpose and then coax an image out of the chaos.

As part of a digital painting assignment this last semester we tried this method out as an exercise. It was so much fun that I've used it a few times sense and I thought I would share a little about what I've learned so far.

Your first task is to make some kind of visual noise to work with. One approach, and the one we talked about in class, comes from Scott Robertson on his Gnomon Workshop DVD, Creating Unique Environments. Scott uses large greyscale markers to make interesting organic shapes on a dozen or so sheets of paper, then scans them into photoshop layers where he overlays and blends them with different blending modes until he sees something interesting.

Another approach, this one from Chris Oatley, is to take a random photo from the web and then zoom WAAAAAAY into small sections of it looking for something interesting.

For my first attempt I dug through my art drawer and pulled out every used up charcoal stump and mislaid tube of paint and spent about an hour randomly smearing things on different scraps of paper. Then I scanned them all into layers of a single photoshop file at very high resolution (600 dpi). Then I spent about an hour zooming in to different portions looking for interesting shapes. Here's one of about 6 sets I came up with:

The first thing I realized after doing this is that not all random stimulus is created equal. I had a lot of trouble finding interesting shapes, but on Scott's DVD he picks them out one after the other. The reason is that even though the shapes he was making were random they all had organic structural shapes, so when he went looking for landscapes and buildings the building blocks were there.

Still, I did find some fun things. I especially liked that one on the bottom left so I opened it in a new documents and started laying things over it, painting into it, running filters against it, and after about an hour came up with this:

And then this:

It's nowhere near finished, but it's certainly a good start on something.

That first session was so much fun I thought I I would give Chris's technique a try, so I zoomed way in on a section of that painting, then started layering, cropping, and painting again. Here's the result of that:

So you can see how things can change quite drastically.

One thing that was bugging me about this process was that the results were so random. I wanted to see if I could control things a little more. This next time before I started working I thought about the kind of image I wanted to create. I've been looking at a lot of ink work lately, people like Mike Mignola, Ale Carloni, Alex Toth, and Francis Vallejo so I thought this might be an interesting target to aim for. I also had just finished watching China Town, a really fantastic Film Noir, and I had the idea of those 1920's cars with the giant headlights in my mind.

I got out my bottle of ink, some bristol board scraps, and some masking tape. I put strips of tape down at random on the paper and then made a mess with the ink. Heres a scan of some of the results:

So just like before I scanned these into the computer at high rez, but this time instead of zooming and layering, I used the lasso tool to cut out random shapes which I pasted into a new file. Then I started moving the shapes around like puzzle pieces until something started to emerge.

After a few hours of moving and only a little bit of painting, here's what I got:

Other than some of the fine details like the face, the car's grill, and the skull emblem on the door, everything here was pieced together like a collage with only a minimum of straight painting. Here's the final version:

I'm still not confident that I have a good handle on steering this process where I want it to go, but I think that's probably part of how it works. It does seem that you can at least point things in a general direction by the sort of raw materials you put in at the beginning.

If nothing else, it's an interesting exercise in composition and visual metaphor and a good way to hone your skills in identifying both.