Why I Paint From Reference Photos (And What I Always Change)

There's a faction of oil painters who won't admit they use reference photos. They talk about painting from life only, about the corruption of the photographic eye, about how a camera flattens everything it touches. And they're not wrong about the flattening.

But most of them are working from photos. They just don't say so.

I use reference photos on almost every painting I make. I'm not embarrassed about it. Here's how I use them, and more importantly, here's what I always throw out.

What the reference is actually for

The photo gives me two things I can't manufacture: the composition and the value structure.

Composition is a decision problem, not an execution problem. When I'm building a figure painting, a western landscape, any scene with a subject in it, I need to know where things sit in the frame before I pick up a brush. Reference helps me find that. I'll take a dozen shots from different distances and angles. Most of them are wrong. A few of them have something.

Value structure is simpler. In oil painting, the relationship between light and dark is the whole game. A reference photo doesn't lie about values the way memory does. If the shadows fall hard and fast, they're hard and fast in the photo. I can see it. I can trust it as a starting point.

Everything else is up for revision.

What I always change

Color temperature is the first thing I push. A photo shot in natural light tends toward the neutral. The warm-cool shifts that make oil painting feel alive get compressed into the middle of a sensor's range. So I take whatever temperature the reference suggests and I exaggerate it. Warm light gets warmer. Cool shadows go cooler. Not to the point of cartoonishness, but to the point where the painting feels like it has actual weather in it.

Backgrounds almost always get simplified. A photo records everything in the frame. The tack on the fence post. The truck in the distance. The power line crossing the sky. Oil painting is about what you choose to see, not what the lens collected. So backgrounds get flattened, neutralized, or dropped entirely.

The other thing I always change: saturation in the darks. Photos tend to render shadows close to black, but actual shadow is full of color, especially in California light. Blue-purple in the shade. Green buried in dark grass. Red-brown in a cliff face below the sun line. That's real. A reference photo rarely shows it well. I put it back.

What this looks like in practice

Cowboy Study is a clear example of where the reference gives you one thing and the painting needs to be something else entirely. The composition in the source is solid: figure in profile, lasso mid-arc, horizon placed well. But the sky was flat. The background had too much information in it.

I kept the figure. I kept the horizon placement. I rebuilt the sky from scratch: deeper reds, a warmer gold in the lower band, more contrast at the line where sky meets land. The background got flattened into something close to abstract. The figure went darker against it.

The painting ended up looking better than the reference. That's the only benchmark that matters.

Cowboy Study is available as a reproduction print here.

The photograph is a floor, not a ceiling

People sometimes ask if painting from reference photos is cheating. The question comes from a place of thinking the point of painting is accuracy. It isn't. The point is translation.

A photograph records light. A painting recreates feeling. If the reference gets me to the feeling faster, that's not a shortcut. That's the process.

The reference is where I start. What I change is the painting.