Tuesday, February 5, 2008

Color Wheel Masking Update

On the two last Color Sunday posts here and here, we’ve been looking at the concept of color wheel masking, the idea that we can map out a color scheme on top of a color wheel.

Blog reader David Briggs has made a brilliant contribution to this idea. He discovered a website called couleur.org by Philippe Colantoni that allows us to actually visualize a color scheme in three dimensions.

What you’re looking at here is a Dinotopia painting called “Palace in the Clouds” paired with a representation of the color scheme expressed in three dimensions.


The little floating blobs represent the amounts of each component color of the composition. On the left you see the hexagonal color wheel in pastel tones. Floating above that, white and blue blobs indicate the large areas of white and blue areas of the picture.

The diagram on the right shows the same diagram as viewed from the side. The vertical dimension indicates the value—or the level of lightness or darkness. The yellow and brown colors show up as little beads floating to the right of the red center line.


Here’s another painting with its color scheme chart. You’ll notice that most of the colors are fairly grayed down, which places them near the center of the axes. The light warm colors show up as a sprinkling of dots in the yellow-orange region at the upper right. The diagram on the right shows dark dots clustered at the bottom, greenish on the left of the red line, and brownish to the right.

My thanks to Mr. Colantoni and and Mr. Briggs for your generosity in sharing your discoveries. Don’t miss this coming Sunday, where we’ll continue our exploration of color wheel masking.

Tomorrow: Two Values

11 comments:

ZD said...

I really don't understand this or why it is useful. Hopefully if I read it enough times it will start to sink in.

Anonymous said...

"The little floating blobs represent the amounts of each component color of the composition." i think the point is that you can see if, say, the painting is strongly biased toward certain colors or areas of the color wheel. For example if there was a lot of yellow in the painting, there would be large blobs in the yellow portion of the color wheel.

Anonymous said...

Sure, but what do the 3 dimensions add to it?

James Gurney said...

Sorry, ZD, I may be overstating the value of this, but I'm kind of a closet geek when it comes to any kind of computer visualization.

In a way the graph shows what we already know about the picture's color scheme. I agree with Anonymous that the dimension of value doesn't add that much. Perhaps the readout would be more useful if it gave us the readout as a polygonal shape over a normal color wheel. This way we could test to see if we were really using the color ranges we had intended.

Michael Dooney said...

Uh, you lost me here, just when I was thinking the color masking technique was straight forward ;)

James Gurney said...

Don't worry about it, Michael. The more I look at these readouts the more confusing they are to me, too. The simple paper mask over the color wheel that I've been presenting on Sundays makes the whole idea much more understandable and useful. This doesn't have to be complicated at all.

briggsy said...

What are you finding confusing about them James?

James Gurney said...

Briggsy--I think what's going on is the usual trade-off when graphing data. The more accurate the representation and the more factors you include the more "busy" the result becomes.

This system accurately represents two additional dimensions beyond my more rudimentary paper masks, namely the dimension of value and also the quantity of pixels of a given type. I'm very enthusiastic about this system because of that additional insight. But to my eye, what's lost is a simple graphic sense of the overall boundaries of the color scheme, showing what parts of the color wheel are in or out. Unless people are accustomed to looking at 3-D graphing of color, it can be harder to grasp.

briggsy said...

This is the other output option, showing the colours without the histogram effect. Any better?

http://djcbriggs.googlepages.com/windmill2.jpg
http://djcbriggs.googlepages.com/palace2.jpg

James Gurney said...

Wow! Thank you, Briggsy. I haven't had time to explore the output options of Mr. Colantoni's software, but this expression is much clearer to me, and I recommend that everyone who is interested in this stuff should follow your links.

On Sunday I'll present how I generate a color scheme using my analog version of color wheel masking, and it would be interesting to run the finished painting through this filter after the fact to see how close it actually matches the generating color mask.

All this stuff may sound somewhat clinical and technical to some people, but I think these methods serve to inform our thought process so that it all becomes intuitive.

briggsy said...

Here here to that!

If anyone could use an introduction to the concept of colour having three dimensions, and why that's important, or would just like to see some more examples of Dr Colantoni's program at work, you might like to start here:

http://www.huevaluechroma.com/011.php