Tuesday, March 18, 2008

Dappled Light

Light coming through trees results in the spotted light we know as dappled light. The painting below is by Ivan Shishkin.

The circular spots of light shining on the ground vary in size depending on how high the canopy is above the ground. A high tree canopy leads to larger circles with softer edges. Below, in this early Albert Bierstadt painting called “Sunlight and Shadow” the effects of dappled light are worked out extremely convincingly.

When bundles of light pass through small spaces between the leaves, each of those spaces act like a pinhole camera. The parcels of light are essentially like conical shapes of illumination radiating from each pinhole in the foliage.

The circles of light touching the ground are actually projections of the disk of the sun. In fact on days with a partial eclipse of the sun, the circles of light will appear as half circles. In the 8x10 inch oil study above, the circles of light on the roof of the shed are about a foot in diameter.

When each cone of light intersects a sloping surface like a wall, it results in an elliptical shape. On a vertical surface parallel to the picture plane, the long axis of that ellipse will always angle back toward the source of the light.

In the detail above from Dinotopia: Journey to Chandara, the light is coming from the right. The circles become ellipses as the form of the ship's red hull curves away on the left side.

Therefore, ellipses of light projected on a wall in front of us will slope upward to the right only if the sun is also coming from the right, as it is in this photo also.

Apparently N.C. Wyeth was unaware of this principle when he designed this otherwise fine illustration of Ben Franklin’s arrival in Philadelphia. The result is an error in lighting. According to the cast shadows from Franklin’s leg and from the tree branch at the bottom of the picture, the light is coming from above and to the right. But the ellipses of light on the wall point impossibly to a light source above and to the left.

Further discussion of dappled light, with photographic examples, on Edward Tufte's website, link.

Tomorrow: Business Cards, Illustrated


Erik Bongers said...

I'm not sure about the pinhole camera effect with the sun's light.

As far as I know, the sun is so far away that it's light rays are virtually parrallel. That's the reason the shadows (or light spots) are about equally big as the shadow's source or hole. And that's also the reason the stretched out light spots on a wall that is at an angle to the sun, are parrallel and not spread out like they would if the light source was a flashlight.

Of course you can get pinhole images of a landscape with the sun, so the effect does exists, but only visible with really small holes.

As an artist, I prefer the idea that the sun's light travels through holes in parrallel cylinders, like spaghetti dough pushed through the leaves.

I think CG folks will concur that they have different light sources for bulbs and for the sun because latter's rays need to go parrallel.

Don't shoot me...shoot my optics professor !

James Gurney said...

Erik, you're right that the rays of sunlight are virtually parallel, but remember the sun is not a point source, but rather essentially a circle of a fixed diameter with rays heading out in all directions from all points on the circle. So the rays from the entire disk of the sun would project through a pinhole and cast a perfectly circlar image of the sun itself on any darkened surface.

I've made pinhole projectors that show very clear images of the sun, clear enough to see the transit of Venus and partial eclipses. The size of the spots of dappled light are not related to the size of the apertures, but rather they're a minimum size related to the distance from the projecting foliage. If you follow the link at the end of the post, there's a good diagram of this geometry, which is a bit tricky to explain without it.

Marc Hudgins said...

Yet another excellent topic. Being an artist by trade, but the son of an engineer, I really enjoy discussions about the nuts and bolts of picture making and analysis of optical effects. I also love the Tufte reference! Have you ever gone to one of his seminars? They are a real treat.

Erik Bongers said...

Ah, yes indeed. The pictures of the link clearly demonstrate it.
There's indeed a minimum size of spots related to the distance of the foliage.
I'm adjusting my rule-of-thumb now :
Larger holes follow the 'parrallel light' rule, smaller holes will show as circles, ALL OF EQUAL SIZE.

The brightness will vary depending on the aperture. The minumum size depends on the distance surface->foliage.

Of course the 'equal size' is only the case if the pinholes are exactly at equal distance. The size is however NOT dependant on the size of the pinhole.

In addition, the blur at the edges of the really big holes will be comparable to the minimum size of these circular spots.
Also, brightness of the really big holes will be constant.

Yes, yes, I see it now.
Thanks for the clarification !

Rob Hummer said...

There are a couple of great photos I've found that clearly show the pinhole camera effect of sunlight passing through foliage. Both were taken during partial eclipses and show multiple images of a crescent sun projected onto a wall and a parked car. They can be found at-



As always, thanks for the great info James.

Darla said...

The Wyeth picture is a really good example! Light and shadows are so interesting. I have been trying to pay more attention to that lately, since I have to create so many things out of my head. Even though I work from client's photos, I often have to change clothes or alter their surroundings, and that effects the light and shadows on occasion.

Charles Alexander said...

Thank you for posting this-- very thought provoking!

James Gurney said...

Marc, how cool that you got to go to one of Professor Tufte's seminars. I haven't heard him speak, but I've read a couple of his books, and I recommend them. Has anyone read the Minnaert text that he refers to in the link at the bottom of the post?

Erik, thanks for that last group of observations, which are very sensitive, and bring the whole understanding a step further.

a. fortis said...

Not to be contrary (because I love this discussion about the pinhole cameras, etc.), but there could be a tree off to the left, out of the picture plane, casting the shadows...but I guess it's still a problem because of the perceived ambiguity of the shadows' source.

limbolo said...

Wyeth's 'error' brings a subtle, uncanny mood to his painting. It is fascinating to know 'why' the picture is uncanny. Thanks for pointing this out Mr. Gurney