Those new to SLRs are often confused about why their expensive and sophisticated camera will sometimes under- or over-expose.

This guide aims to explain how exposure works, and three things you can do when the camera doesn't get the exposure right.

If you're already familiar with the basics of aperture and shutter speed, click here to skip past the explanation of that.

Let's start at the beginning. Exposure determines how much light reaches your film or digital sensor.

For natural-light photography, there are only two variables that affect exposure: aperture and shutter speed. (ISO might be thought of as a third, but it doesn't actually affect the exposure, even in the case of a digital camera: higher ISO simply tells the camera to amplify the signal from the sensor.)

Aperture is an iris mechanism that creates a larger or smaller opening inside the lens. A bigger opening lets in more light than a smaller opening. A bigger opening is known as a wider aperture and (slightly confusingly) has a lower aperture number. So f/2.8 is a wide aperture that lets in lots of light, and f/22 is a narrow aperture that lets in very little.

Shutter speed is a plate (or set of plates) that cover the film or sensor. The plates open and close when you press the shutter release. The longer the shutter is open for, the more light reaches the film/sensor. So 1/30th of a second lets in lots of light, 1/2000th lets in very little.

In fully automatic mode, the camera will determine the amount of light (explained below) and then set a mid-range aperture and mid-range shutter speed. This may or may not give you the result you want, which is why it's a good idea to take control of the exposure yourself. Programmed Auto mode (P) is exactly the same as Auto, except you can adjust aperture and shutter speed together to gain a certain amount of control. So it's one step forward, but you'll usually want more control than that.

In aperture-priority mode (A), you choose the aperture and the camera chooses the right shutter speed for that aperture. The main reason you want to control the aperture is that aperture determines depth of field. This is best illustrated by example.

In the first scenario, we want everything to be in focus. This will require a narrow aperture (=high number). In this case, we want the snowballer in focus (a few feet away from us) but also the mountain several miles away:

Photo: JW Hubbers of www.ringbunny.nl

Now let's take the opposite scenario. The time, we want to focus the viewer's attention on one very specific part of the frame, so we want that in focus and everything behind it out of focus. Here's a good example, where I wanted to focus attention on the hand:

Her face is just 20cm away, so we need a very shallow depth of field to throw the face out of focus. In this case, the aperture I used was f/2.8.

Usually, you will want either the foreground or the distant view in focus, but occasionally you will want a blurred foreground and background, with the subject in focus in the middle ground:

Photo: JW Hubbers of www.ringbunny.nl

There will be other types of photo where shutter speed is the more important variable. Here's a very boring but classic example:

If we shot that at a high shutter speed, the car would appear stationary. By choosing a slow shutter speed (in this case, 1/50th) and panning with the car, we have a blurred track and blurred wheels, giving some impression of speed. Here's another example, shot from a train window at 1/30th second to create a very abstract shot:

Conversely, there may be times when we want a very fast shutter speed, to freeze all movement. Photographing a splash of water is the classic example. Here's a variation on that theme:

A modern camera like the D70 offers three forms of automatic metering: spot, centre-weighted and matrix.

Spot-metering is very simple: it measures the amount of light reflected from a tiny area in the centre of the lens. Usually 1% of the area of the scene.

Centre-weighted bases the exposure mostly on the centre of the scene, but also takes into account readings from the rest of the frame. The size of the centre-weighted areas varies by camera, and some cameras allow you to vary it (on the D70, you can choose between a 6mm, 8mm, 10mm and 12mm circle).

Matrix metering is a very clever thing. It not only takes readings from across the frame, but it then compares the pattern of light and dark to a database stored on a chip in the camera, and tries to find a match. For example, if it gets a pattern of light that seems to be person-shaped, it assumes you want the person correctly exposed so adjusts the exposure for them, rather than the background. The D70 database contains over 1000 patterns!

Matrix-metering is the usual default, coping with the widest range of scenes. Centre-weighted is usually used for portraits, and spot-metering is dealt with below.

Given that modern metering is so sophisticated, why does the camera still sometimes get it wrong?

To answer that, we need to understand how cameras measure light. When we take a photo, we don't have the camera pointed at the source of the light, we have the camera pointed at the thing we are photographing. What the camera is measuring, therefore, is the amount of light reflected from the thing we are photographing.

Now, the world contains a great mix of colours. Very obviously, a black object reflects much less light than a white object. Every colour reflects a different amount of light. If we average out all the colours in the world, the amount of light they reflect is the same amount as would be reflected from a mid-grey object. What is technically known as an '18% grey' (so called because it is a shade that reflects back 18% of the light that falls on it).

So, the camera knows that the world as a whole is 18% grey in reflectivity, and adjusts the exposure to let in the right amount of light for a scene which averages out at that.

Which is fine for a general scene like this:

This has a good mix of colours. Overall, the scene averages out pretty close to 18% grey so the camera sets the correct exposure.

The problem comes when we shoot a scene that doesn't average out at 18% grey. Let's take the simple example of a photo of a black camera backpack.

Let's look at what happens if we ask the camera to measure the exposure. The camera measures the amount of light being reflected from the bag:

Remember, the camera thinks the scene is 18% grey. It has no way of knowing that we're photographing a black object, so it does the only thing it can do: it sets the exposure to be correct for a scene that is 18% grey. In other words, the exposure set by the camera will make the bag look lighter than it is. The bag will be over-exposed, and our black bag will appear grey.

The same problem in reverse happens with white objects. The camera assumes they are 18% grey, so it under-exposes them. That beautiful close-up of the white bride's dress will appear grey in our photo.

So, how do we solve this? Well, there are five methods. One of them would require an article of its own to explain, so if you want to know about that one, google on 'the zone system'. We'll deal here with the other four. All four involve putting the camera into manual mode or using exposure compensation so that we can set the exposure we want rather than the exposure the camera thinks is right.

Grey card

First, we can carry an actual 18% grey item around with us and measure the light reflected from that. We then set the correct exposure for that item and use that exposure to take our photo.

The most convenient 18% grey item to carry is a piece of card known, unsurprisingly, as a 'grey card'. You can get this from any photography shop for about a fiver. To use it, you put it in the same light as the object you want to photograph (usually holding it just in front of the item you want to photograph) and measure the light from that.

Start by setting the aperture or speed you want, then take a reading from the grey card and see what the camera suggests for the other variable.

When using matrix metering, you need to make sure that you fill the frame with the grey card, so that the entire scene 'seen' by the camera is 18% grey. Your auto-focus won't be able to cope with that, but that's fine: you don't need to take a photo, only half-press the shutter to see what reading you get. Be careful not to get your own shadow on the card!

If it is not practical to get close enough for the grey card to fill the frame, you can use spot-metering and ensure that the grey card is in the centre of the frame. If you can't get anywhere close to the subject (eg. at a sports event), you can put the grey card anywhere in the same light - ie. in the sun or in the shade - with your camera pointed in the direction you'll be shooting.

Handheld lightmeter

The second approach is to use a handheld lightmeter. You need a meter capable of taking what is known as an incident light reading. What this means is that, instead of measuring the light reflected from the item you're photographing, you measure the amount of light falling onto the subject. It doesn't matter how reflective the object is because we're not measuring reflected light. So, in the case of our black bag:

We then use the reading we get from the lightmeter and manually set the camera accordingly.

To take an incident light reading, you hold the lightmeter in front of the object you want to photograph, with the white dome/cone pointed towards the light, and press a button:

As with the grey card, you can take a reading in the same light and in the same direction when you can't get close to your subject.

Then we check the reading. I use an elderly (but very accurate) Weston Master V lightmeter. You read the needle, set that value on the inner dial and then read off the settings from the middle and outer dials. In this particular case, I wanted to use f/1.8, so I looked at f/1.8 (a big 1 and a smaller 8 above it) to get my shutter speed reading of 1/250th:

In contrast, when I just point my D70 at the bag, the D70 suggests a shutter speed of 1/160th. That's 2/3rds of a stop difference. Let's see what difference this makes to the photo (I wasn't using a tripod, so the two halves don't match up exactly):

The D70 reading (reflected) gives us an over-exposed image which makes the bag look (surprise, surprise) 18% grey, The Weston reading (incident) gives us a much more accurate exposure. The bag still looks grey, because it has a lot of light falling on it, but that is how it really looks in that light.

Spot-metering on a midtone

If you can recognise an 18% grey object (aka a midtone) in your scene, then you can spot-meter on that, note the exposure reading, manually set that exposure and then reframe.

You can also use exposure creatively by choosing something that you want to be exposed as a midtone, even if it isn't.

Judgement!

I said I was going to cover four methods of getting the exposure right. The fourth is simply to use your experience and judgement. Given that we know the camera expects the world to be 18% grey, we can look at what we see through the viewfinder and make a guess as to how much lighter or darker the scene actually is. We can then use exposure compensation to adjust as required.

Helpfully for portrait photographers, caucasian skin is pretty close to 18% grey in terms of light reflectivity, so spot-metering on skin will generally give good results. Indeed, some photographers will use their own hand as a grey card! For black skin, you will tend to need to adjust the exposure downward, the amount depending on the shade of the skin.

To develop this judgement, make a note of any photos you take that are under- or over-exposed. Study the balance of light and dark in the photo and see if you can make sense of why the camera got the exposure wrong. This will give you a feel for how you might need to adjust the exposure for similar scenes in future.

Cautions

With any form of manual metering, be aware that light can change very rapidly! Especially outside with partial cloud cover. If you are taking more than a single shot, it is usually safer to use exposure compensation rather than work in manual mode. In that way, the exposure will be automatically adjusted as the light changes, but you will still have the same degree of compensation in play.

If you are shooting digitally, shoot RAW. This is not a substitute for getting the exposure right at the time - adjustments can only take you so far - but is a good safeguard.

Note that underexposed images are easier to correct than over-exposed ones. In a digital camera, all colours are translated into three numbers (the amount of red, green and blue light). Something that is totally black will be 0,0,0 and something that is totally white will be 255,255,255.

In an underexposed shot, some dark areas may have lower numbers than they should. We can then adjust the RAW image to boost those numbers and recover the hidden detail.

In an over-exposed shot, the brightest areas will go totally white: 255,255,255. No adjustment is possible because there is no data there to recover.

It is thus safest on a DSLR to err on the side of under-exposure. Nikon DSLRs do this by default, which is why new owners tend to complain about dark photos with muddy colours. Nikon assumes you will post-process your photos, so errs on the side of caution. If you don't want to post-process, set exposure compensation to +0.3EV on a sunny day or +0.7EV on a cloudy day.