Dynamic range

Learning goal

You will be able to explain dynamic range as the relationship between the difference separating a scene's darkest and brightest points and how much of that difference the sensor can record in detail. You will show that when the scene's range goes beyond the sensor's, one end will certainly be lost; you will explain why the eye sees the same scene without trouble; you will learn to narrow the range by changing the light, the time or the framing, and to widen it by merging several frames or shooting at base ISO; and in backlight, you will find ways to protect faces of different skin tones in the same frame as a bright sky.

The question

On a summer noon, when you photograph a friend sitting in the shade of a tree, you end up having to choose one of two photos. In one, your friend's face is correctly exposed, but the sky behind is a pure white void. In the other, the sky is blue and the clouds can be made out; but your friend's face is a dark shadow. Yet while you were there, you could see both at the same time, easily.

Why does the sky always come out pure white? Because the difference between the bright and the dark in that scene is bigger than what the sensor can record in a single frame. Your eye closes this gap with its own methods; the sensor can't. In this lesson we will talk about the name of this gap, dynamic range, and the paths you can choose when you hit the sensor's limit.

Context

In the previous lesson you saw that the bright end stops at a sharp wall, while the dark end ends in a slope that grows grainier and grainier. The distance between these two ends is the tonal range the sensor can record in detail. If the tones in the scene fit into this range, everything is recorded; if they don't, you decide with the exposure which end will be lost.

Dynamic range is also measured in stops. However many times more light the scene's brightest point reflects than its darkest point, the difference between them is that many stops. On a sunny noon there can be many stops between the sky and the shadow under a tree; on a cloudy day the difference is much smaller. The range sensors can record changes from camera to camera, and the numbers manufacturers give are measured by different methods; that is why, rather than an exact number, what matters is seeing on your own camera what fits into one frame and what doesn't.

Simple explanation

What is dynamic range?

The word dynamic range is used both for the scene and for the sensor. The scene's range is the difference between the darkest shadow and the brightest point in the scene. This is also called scene contrast. The sensor's range, on the other hand, is the tonal width in which the sensor can record detail between jet black and pure white; this is the sensor dynamic range. If the scene's range is narrower than the sensor's, as on a misty morning or a cloudy day, all the tones fit and the histogram touches neither edge.

If the scene's range is wider than the sensor's, one end will certainly be left out. If you set the exposure for the bright end, the shadows drop to jet black; if you set it for the dark end, the bright areas blow out. If you set it somewhere in the middle, you lose a piece from each end. On the histogram, this shows up as the graph pressing against both edges at once, or sticking to one while rescuing the other. In this case, exposure isn't a calculation but a matter of preference: you decide which end matters more.

How does the eye manage this?

The range the eye can record in a single moment is limited too; but the eye doesn't take a single frame. When you look at a scene, your gaze keeps moving: when you look at the sky your pupil shrinks and your retina adapts to the brightness, and when you look at the shadow it does the opposite. Your brain combines these piece-by-piece looks and offers you a scene detailed in both the bright and the dark at the same time. This is called adaptation, and it happens constantly, without you noticing.

The camera, on the other hand, records the scene in a single moment, with a single setting. Whether to set it for the sky or for the shadow is decided once, and this decision is applied to the whole frame. That is why a scene your eye sees easily loses one of its ends in the photo. One habit of experienced photographers is to try to consciously stop the eye's adaptation when looking at a scene: squinting at it shows the difference between bright and dark more roughly, but more honestly.

How do you narrow the range?

If the scene's range is very wide, the first solution is to narrow it; this is reducing scene contrast. You can change the light: when you bounce sunlight onto a face in shadow with a white reflector, the face brightens and the difference between face and sky shrinks. The reflector you used with window light in Lesson 5.6 does the same job here. A small flash can serve the same purpose too. You can change the subject: keeping the person in the shade and putting a shaded background behind them as well takes the sky out of the frame.

You can change the time: in the noon sun, the difference between sky and shadow is at its biggest; early in the morning or late in the evening, in the golden hour, the light softens and the difference shrinks. You can change the framing: taking only a small piece of the sky, or none at all, takes the end that doesn't fit into the sensor out of the scene. What these paths have in common is that they solve the problem before the photo is taken. When you bring the scene's range down to the sensor's range, a single frame can carry everything.

How do you widen the range?

If you can't narrow the range, you can widen the recorded range. Shooting at base ISO is the first step: as ISO rises, the range the sensor can record usually narrows too, because the grain in the shadows grows. Shooting RAW leaves more room for pulling the highlights and shadows together in editing. If a wider range is needed, you can take several frames of the same scene at different exposures and merge them; frames exposed one for the sky and one for the shadows meet in a single photo. This is called HDR merge.

Phones often do this on their own: the computational photography you saw in Lesson 4.5 merges several frames when you press the shutter button and tries to protect both the sky and the shadow. This merging sometimes flattens faces: it can open up the shadow of dark skin too much and turn it gray, or press down the brightness of light skin too much and turn it into a pale surface. After photographing a face in backlight, whichever path the phone or camera has chosen, zoom in on the face and look. If the skin keeps its own depth and brightness, the merge has done its job; if it doesn't, changing the light, for example opening up the face with a reflector, leaves the skin closer to reality.

Visual evidence

Imagine photographing an elderly woman selling tomatoes in the shade of a giant plane tree in a village square at noon. The woman's skin is a deep wheat color; behind her are sunlit whitewashed houses and a cloudy sky. When you set the exposure for the woman's face, her face, the lines around her eyes and the pattern of her headscarf can be made out; but the houses and the sky are a pure white void. When you set it for the sky, the clouds and the windows of the houses can be made out; the woman, on the other hand, is a shape darkened in the shade.

You ask a friend to stand in front of the woman and bounce sunlight onto her face with a white foam board held close to the ground. The woman's face brightens by a stop or two; the difference between face and sky narrows. When you set the exposure between the two, in a single frame both the woman's face, the warm color of her skin and the bright red of the tomatoes, and the whitewash of the houses and the clouds can be made out. The same scene, the same sun; you couldn't change the sensor's range, but you narrowed the scene's range.

Common mistakes

Summary

Dynamic range is whether the difference between a scene's darkest and brightest points fits into the tonal width the sensor can record in detail; if it doesn't, exposure only chooses which end will be lost, while the eye doesn't feel this limit because it moves its gaze around and adapts constantly. It is possible to narrow the range by opening up the shadow with a reflector or flash, or by changing the time or the framing, and to widen it by shooting at base ISO and in RAW, or by merging several frames. In frames merged by a phone or camera in backlight, you need to zoom in on the face and look at whether the skin keeps its own depth and brightness, and change the light if needed.

Mini assignment

In this assignment you will see your own camera's dynamic range limit.

1. At noon, photograph a friend sitting in the shade of a tree, with the sky visible behind them, once exposing for their face and once for the sky, and compare the two histograms.
2. In the same scene, bounce sunlight onto their face with a white card or foam board, and take a single frame with the exposure set between the two.
3. Shoot the same framing with your phone's own camera app, without touching any setting, and look at how the face and the sky appear in the frame the phone merged.
4. Shoot the same place again in the late afternoon, in the golden hour, and note how much the range has narrowed.

Check yourself

• In the noon scene, did you see the histogram press against both edges at once?
• With the reflector, could the face and the sky be recorded in detail in the same frame?
• In the frame the phone merged, did your friend's skin keep its own depth and brightness?

Glossary

Mini quiz

1. What happens if the scene's range is wider than the sensor's?

• A) The photo is always correctly exposed
• B) One end will certainly be lost; the exposure chooses which one
• C) The sensor widens by itself
• D) Only the colors go wrong

2. Why does the eye see without trouble a scene the sensor can't record?

• A) Because the eye's range is unlimited
• B) Because the eye sees the scene in black and white
• C) As the gaze moves around the scene, the eye adapts constantly and the brain combines the pieces
• D) Because the eye sees only the shadows

3. What is one way to narrow the difference between a face in shadow and a bright sky?

• A) Bouncing light onto the face with a white reflector
• B) Cleaning the sensor
• C) Converting the photo to black and white
• D) Stopping the aperture down to its narrowest

4. If dark skin has slid toward gray in a frame the phone merged in backlight, what do you do?

• A) Accept the frame as it is
• B) Set compensation all the way to plus
• C) Make the sky bigger in the frame
• D) Zoom in on the face to check it, and narrow the range by giving the face light with a reflector

Answers

1. B — The sensor can't record all the tones; whichever end you set the exposure for, the other end either drops to jet black or blows out.
2. C — The eye doesn't take a single frame; it adapts constantly to where it looks, and the brain combines these looks to offer a wide scene.
3. A — The reflector adds light to the face in shadow; the face brightens, and the stop difference between it and the sky shrinks.
4. D — The merge can open up the shadow too much and flatten the skin; giving the face real light narrows the range and gives a more natural skin.

Summary card

• Dynamic range: does the difference between the scene's dark and bright fit into the sensor?
• If it doesn't fit, exposure only chooses which end will be lost.
• The eye adapts constantly; the sensor takes a single frame.
• Narrow: reflector, flash, time, framing. Widen: base ISO, RAW, HDR merge.
• In merged backlit frames, zoom in on the face; if the skin isn't natural, change the light.

Bridge to the next lesson

You now know dynamic range and how metering goes wrong. The scenes that push this knowledge hardest are places that systematically fool the camera's metering, such as snow, beaches, concerts and night lights. The next lesson takes these difficult lights scene by scene and shows how to fool your camera, going beyond plus compensation in bright scenes and minus in dark ones.