#1265. A few thoughts on contrast

Much of what we do in photography to create a visual style comes down to contrast management. Yet we rarely give it the prominence it deserves in our thought process.

 

Contrast needs to be defined between two areas. The contrast between two pixels refers to their difference in luminance or in colour. The term Contrast in a whole photograph is usually used to refer to the depth of the dominant shadow areas and the lightness of the dominant highlight areas.

It is easy to recognise in a histogram. When all the tones are bunched up together (whether in the shadows, or in the midtones, or in the highlights, it doesn’t matter) you have a low contrast image. When most of the tones are in the highlights and in the shadows, with little in between, you have a high contrast photograph.

The photo above would generally be considered to be low-ish contrast, while the one below is high-ish contrast.

 

The important distinction between the two ideas (pixels vs whole frame) is the scale on which contrast is being considered. In one case, we can measure the luminance difference between two neighbouring pixels (for instance to measure the MTF curve of a lens). In the other we are judging the tonal range of a whole image.

There are many stages in between, and most of the tools in our editing software are based on those.

In Adobe Lightroom, for instance, you have the Constrast slider, the Tone Curve editor, and the Texture, Clarity and Dehaze sliders. All of these alter contrast at various scales ranging from a few pixels (Texture) to the whole frame (Contrast slider, Tone Curve editor). Oh, and the sharpening tool, at the pixel level!

 

Sadly, none of those tools give us complete control over the process. They apply recipes that include local contrast enhancement (of various scale) along with other proprietary tone manipulations (at least, I imagine so).

Other software actually lets you define the amount of contrast you want to add or eliminate from an area of the photograph, as well as the radius of the brush you are applying, and whether to apply to the shadows midtones or highlights. That is nice. You can of course replicate that through the layers of Lightroom. I’m not dissing Adobe, only highlighting different approaches to what constitutes the topic of this post: contrast can be applied more or less locally or globally.

While this might seem like stating the obvious, the feature is not being used nearly enough. And my guess is that the obscure labelling of the various tools that offer the option is partly responsible for that.

 

Before continuing to how to use that contrast, a quick word about colour.

A colour photograph contains luminance information and colour information. You can therefore apply contrast to either or both and it is easy to mix up the two concepts. Colour contrast refers to the nature of the hues present in the image and where they fall (on the colour wheel, for example) relative to one another.

In the photograph above, you see moderate (global) luminance contrast and low colour contrast, because all the hues are blue or green. In the photograph below, made in the same forest but with different colour balance, you again see moderate luminance contrast but much stronger colour contrast. The colours are far more different from one another below than above, in other words.

 

Below is a low luminance contrast (all the tones in the photograph are bunched up in the mid-high zone) but high(ish) colour contrast photograph (green and orange & pink pastels).

 

And below is a high luminance contrast (lots of highlights, lots of shadows) and high colour contrast image (orange, green and blue dominate) photograph. Particularly in the top-right corner.

 

Altering the luminance will also alter the colour, because saturation of colours is highest in the midtones. So lifting the midtones to lighter tones will make the colours more pastel, for example. But we won’t cover this here. Let’s get back to contrast.

Contrast matters because it influences composition.

All humans scan a frame based on a scanpath that is both determined by our physiology and cultural background (Asian people apparently have a more circular scanpath, where Westerners have a more linear one, for example). This scanpath is interrupted by signals of various nature. Bright zones catch the eye more than dark zones. High contrast features catch the eye more than low contrast features. Faces and other emotional or survival-related patters catch the eye as well. The interplay of all those eye-catching things is what creates the composition of a photograph.

 

Ever seen Sylvain Tesson’s photograph of a snow leopard hiding in plain sight? It’s hard to spot because of its “low contrast” (a.k.a. camouflage) compared to the rest of the scenery. But once you’ve seen it, that’s all you see, the face is mesmerizing. You can no longer unsee it and revert to the initial state of discovery.

That’s just one example of how composition works. And the photograph above is almost the opposite. There is nothing noteworthy there, no leopard, not even a bunny wabbit, but the use of contrast has increased the visual prominence of the natural features of that dried grass and created a sort of rhythm, with “holes” (dark negative space) in it.

Now, equally hidden in plain sight is composition’s little secret: a photograph, however elaborate, must be easy to read. That’s why so many of us just plonk a subject in the middle (or rule-of-thirdly) of the frame with a background that’s blurred out of existence by a fast lens. That’s an easy photograph to read, often a beautiful portrait for example.

 

The trick for readability is to know how many layers of different contrasts to apply to a photograph.

In other words, you need to know where you wish to direct the attention. You must make the decision. What do you want to show, and what do you want to hide. Add luminosity and contrast to the former, remove from the latter.

And this can be done with any of the local contrast tools mentioned above, with different visual results.

In the image above, a lot is hidden in the shadow.

In the image below, I have reduced the highlights and clarity of the light areas and done the opposite in the shadows.

A lot more information is present in the second, but is that what you want? What is your eye drawn to in both? Which do you prefer? There is no right or wrong answer, only good and bad combinations of intention and execution.

 

In this example, I kinda like both. But if I added much more clarity everywhere in the frame, it would become grating. Hey, what is a grate, if not a rectangular frame with scores of small high-contrast (for cheese or skin) areas? 😉

You need to be deliberate with where you apply contrast and where you don’t.

The reason why landscape photographers love to shoot when the sun is below the horizon is not really because of the blue, pink or golden light. You can dial in a lot of that in post. But landscape photography is a high-detail style. And those times of day provide uniform, low contrast, light. What you get is a low global contrast image, with local contrast that guides the eye to the various features of the landscape and provides info on things like texture. The eye can spend a lot of time roaming the image, lapping up the detail. And if composition is good, it does so in the order inspired by the author. Add bright sunlight to this and, even if your camera can deal with the dynamic range, the pockets of huge contrast will dominate the scene so much that they will completely mess up the order of more subtle contrast arrangements.

 

So how do you decide?

Well, there’s no formal method that I know of, except that it’s best to start by fixing global contrast for the general look, then alter specific areas of the frame (dodging and burning). And I would also start with large radius contrast and move on to smaller radius edit later. Properly trained visual arts students might have more to add to my purely empirical explanations. If that’s your case, please chime in 🙂

The best solution to decide what your style calls for, in the absence of solid fact, is to look at photographs you like and others you don’t like, to see how contrast is applied to them. Are they high contrast, or low contrast? Do they appear ultra-sharp or softer? Grungy (typically high low-radius contrast), as above? Is the contrast the same everywhere or do some areas seem to have more than others? And so on …

 

I personally like high-contrast scenes with large depth of field. So it’s my job to make sure this doesn’t get visually too busy, that detail in the shadow isn’t lost to a huge puddle of black, that highlights do not clip abruptly, that colours don’t turn all garish …

You may prefer very soft images in which hues can express themselves unincumbered by violent luminance trauma, but very sharp locally, for instance. Or low contrast, low sharpness, monochrome images reminiscent of the old film days.

 

Of course, and as a conclusion, gear plays a major role here.

Low dynamic-range sensors and filmstocks appear to have more global contrast. And vice versa. Remember when Leica got a spanking for creating a Monochrom with dull images? It just used a high dynamic-range sensor and preserved all possible detail for post-processing. Brilliant, but shocking to most people.

As for lenses, what they do is essentially transfer contrast from the scene to the image surface. MTF curves display the contrast transfer levels of the lens for details of increasingly small size. Lens drawing is partly defined by the relative levels of those contrast curves. Yes, there are geometric factors (distortion, swirly bokeh, compression, depth of field) and chromatic factors (mainly the scale and appearance of aberrations) at play, but some lenses are designed to resolve fine detail yet look soft, while others are nowhere near as resolved but appear very sharp when the image isn’t enlarged. And shallow depth of field translates to blurred backgrounds, which are nothing more than locally reduced contrast levels. 3D pop is just more contrast here and less there.

I know there are hundreds of more actionable how-to videos online that describe an A to Z procedure to obtain this or that look. Some I have seen involve plugins that handle dozens of layers that let you fine tune to your taste.

My approach is the exact opposite to this spray-and-pray, everything-is-possible smorgasbord of complexity. First because those tools hide the deeper understanding that sets you free, and secondly because the look that works for the guy probably won’t work for me.

I prefer to try to understand and communicate a very limited set of funamentals and encourage a far simpler, yet more empowering, experimental method. Let me know if this one works for you 🙂