Musings on Photography

Depth cues and the landscape

Posted in Uncategorized by Paul Butzi on November 1, 2006

When a person stands in the landscape and looks around, their visual experience is three-dimensional. They ‘see’ things with depth; they experience some objects as proximate and some as remote. This happens even though we can’t really ‘see’ depth; we see lots of little cues (including stereo vision information) and our brain helpfully creates a 3-d model, and part of our visual experience is this 3-d model, not just the direct experience of the light streaming through our pupils and onto our retinas.

One of the problems that landscape artists face, then, is that we’re trying to cram this 3-d experience onto a flat surface. To do this, we include some of the depth cues that our brains recognize, like overlapping. For a flat print, it’s hard to recreate the stereo information. For photographers, cues like distant objects appearing smaller, and convergence of parallel lines, and overlapping – the camera includes them without our intervention. (Yes, you in the back, I know we can jigger with convergence by using movements on our view camera. Sit down.)

But one of the ways we can enhance the apparent ‘depth’ is by either amplifying the depth cues that are there, or adding depth cues that seem to be missing.

Here’s a iconic landscape.

We’ve got sky up top, and we’ve got the ground below. But this looks ‘flat’, like a kindergartener’s depiction of the landscape.

First, let’s get the sky out of the way. When we look at the sky, it’s never a uniform brightness. Typically, it’s brighter at the horizon, especially when it’s early or late. So we can add a gradient to the ‘sky’, and we get

This looks a bit more like sky, doesn’t it? It’s not extreme as gradients go, but it’s enough to give that blue rectangle a sense of sky-ness. Notice the strange sensation of the blue reading as ‘sky’ but the flat, uniform green still reading as if it’s just a green rectangle. It’s as if someone is occluding a view of the landscape with a flat, evenly illuminated green card.

Next, we tackle the ground. When we see a tonal gradient on a flat surface, we tend to interpret that as a plane that’s tilted and thus has different angles to the light source depending on where we look. So we can add some depth cue to the ground by adding a gradient, too.

There’s a bunch of stuff to notice, now. First, we now have a depiction that our brain interprets as a landscape – green foreground running off into the distance, blue sky above. It’s just like the world, only lacking in detail. Notice, too, how the gradient seems to wrap up the sides of the image a bit; it doesn’t, the gradient is perfectly uniform but your visual system is noting the contrast with the edge of the frame and ‘post processing’ that into a change in contrast. Note, also, how your gaze is drawn to the horizon. It’s as if that tonal gradient is pushing your gaze up away from the dark foreground and drawing it up to the light horizon.

Interestingly, you also get a sense of depth in the ground if you reverse the gradient:

Now our gaze is drawn into the foreground, pushed away from the horizon.

Using these depth cues on your photographs is left as a trivial exercise for the reader.

Comments Off on Depth cues and the landscape

%d bloggers like this: