Sections in this page
- Camera exposure parameters
- - A = aperture, as a ratio
- - T = exposure time, in seconds
- - S = sensitivity, in ISO rating units
- - W = white balance
- Geometrical optics
- Aperture = f/x
- Objects far and near
- Circle of confusion
- Depth of field
- Exposure time
- Parameter values
- Effects of particular settings
- - Aperture, A:
- - Exposure time, T:
- - Sensitivity (ISO rating), S:
- Settings to meet requirements
- A possible way of working
- Reciprocity failure
- EV = stop
- For digital cameras - shoot RAW or JPEG?
- - JPEG
- - RAW
- - So how does that affect exposure?
Making effective use of aperture, exposure and ISO
For anyone wishing to go beyond the fully automated capabilities of a camera, to take control and achieve more interesting effects, it can be daunting to work out which settings to use and when. Most cameras apart from the very cheapest, whether film or digital, have dials for setting lens aperture and exposure times in manual mode but how do we use these effectively? It is always good advice to read the user guide but beyond such basics as where the dials are placed, that will tend to give few clues as to how everything really works together to make a good photo.
This page aims to bridge the gap by providing a basic introduction to the techniques. Most of this is applicable to both film and digital cameras but we will see how digital cameras have changed certain things.
The following four factors can be adjusted to affect exposure in a camera. They each have different effects on the photograph so it is important to learn how to adjust them together to achieve good results. The principles are the same for both film and digital photography but the latter avoids the need to change film to control two of the factors.
Aperture is the diameter of the approximately circular adjustable opening in the iris diaphragm inside the lens. It is measured as a fraction of the focal length of the lens because what really matters is the angle subtended for capturing light. For example, f/8 means the aperture (diameter) is equal to one eighth of the focal length of the lens. A bigger opening in a longer lens may still capture the same amount of light as a small aperture in a short lens because the angle is the same. In that case both lenses have the same A value.
NB: Do not be confused by the diameter stated on the front of the lens. That is simply the diameter of the threaded filter ring, not the optical aperture.
Exposure time is the length of time for which the shutter in the camera is held open. Fractions of a second are shown as reciprocals on the dial. Eg, 1/100 second is shown as 100. Times longer than 1 second use " to indicate units of seconds instead of taking reciprocals. For example, 2" means 2 seconds but 2 by itself means 1/2 second. Most cameras also have a B (for Bulb) setting, for which the shutter remains open as long as the shutter button is held down.
Sensitivity is adjustable in digital cameras but in film cameras it is necessary to load a different film to change it. The more sensitive the film or digital detector, the darker the environment in which pictures can be taken. ISO stands for International Standards Organisation. They publish huge numbers of standards so we should really state which one we mean but photographers are too busy for that. For film the standard is ISO 5800:1987. A revised version, ISO 12232:2006, is for digital photography.
White balance measures the relative amounts of red, green and blue for the type of light falling on the subject. It is noticeable that tungsten light bulbs produce a yellower light than fluorescent strip lighting. A white sheet of paper reflects a slightly different colour in the two different kinds of illumination. Our eyes compensate without us taking much notice but the camera isn't naturally so clever. In digital cameras W is often set automatically, but can be overridden manually. It is often also possible to set it by calibrating on a white sheet of paper. In film cameras a different type of film (daylight, tungsten, etc) must be loaded to alter W; it is a fixed property of the film.
Consider a simple convex lens:
In free space light travels in straight lines, called rays (we cannot see round corners without optical aid). Rays of light coming from objects far enough away (said to be at infinity) are parallel. A lens brings parallel rays to a point of focus at distance f behind the lens. That defines f as the focal length of the lens.
We have to move the lens away from the film/detector to focus objects that are closer than infinity (which most objects are of course):
Notice the construction: a ray through the centre of the lens is undeviated; a ray parallel to the axis is like the parallel rays we considered before and therefore will pass through the focal point at distance f from the lens. The intersection of those two kinds of ray determines where the focussed image of the object will be, and it is further away from the lens than f.
Camera lenses are always more complicated than this, having many glass surfaces, but the principles we are describing still apply.
A = aperture
It is the lens diameter, expressed as fraction of f. Eg, f/2.8
The amount of light collected is proportional to the square of A because it depends on the area of the open aperture receiving light, not on its diameter.
Using a 100mm lens (that means f = 100mm) and setting A to f/8 implies that the aperture diameter really is 100 / 8 = 12.5mm. However, we rarely need to do that calculation because f/8 means a certain light gathering capability regardless of the size of the lens, and light gathering is what matters for exposing a photo.
The lens aperture A can be stopped down or opened up again by an iris diaphragm inside the lens assembly. The diaphragm is designed so that
1 stop up or down = double or half the amount of light collected.
Lens or camera scale (stops): f/2 f/2.8 f/4 f/5.6 f/8 f/11 f/16 f/22
Successive values in the scale are in the ratio of the square root of 2 because remember that the amount of light is proportional to the square of the aperture (the area of the opening rather than its diameter).
When the red object is in focus the blue object nearer to the lens is out of focus on the film/detector. Generally only objects at one particular distance will be sharply focussed and all others will be blurred to some degree.
Using the full aperture of the lens: the nearer object (shown by the blue rays) is well out of focus, a blurred disc on the film/detector.
In this real photographic example it is the further objects (street lights) which are out of focus. In this case they should be small points of light but their out-of-focus discs of confusion can be seen clearly here. Sometimes this makes a good creative effect.
When lens aperture A is stopped down by an iris diaphragm the disc of confusion is smaller, so the nearer object appears to be in focus too.
Depth of field = range of distances for objects in focus.
This photo of Dresden across the River Elbe has been taken deliberately at ground level to demonstrate a large depth of field. Objects from a few centimetres to hundreds of metres away from the camera are all in focus. This was done by stopping the lens down as far as it would go. Because the camera was sitting firmly on the ground a long exposure could be used, to compensate for the small amount of light coming through the small aperture.
T = exposure time, measured as fractions of a second.
Camera scale: 1000 500 250 125 60 30 15 8 4 2 1
Eg, 100 means the shutter is open for only 1/100th of a second.
Times longer than 1 second use special notation: 2" 4" 8"
For both aperture A and exposure time T, 1 stop = double / half the amount of light captured.
In practice, camera scales for both A and T go in half stops. In some cameras even smaller steps are possible.
Reciprocity: decreasing A by n stops and increasing T by n stops results in the same amount of light on the film/detector. The two factors have a reciprocal relationship to each other. Any combination of settings which lie on the red line in this graph will produce the same amount of exposure in the photo.
Film only: the relationship breaks down for long exposures, an effect called "reciprocity failure".
Notice how the overall brightness of the two images here is about the same. A and T have been adjusted by an opposite number of stops to keep it so. You can see that changing A has affected the depth of field, as expected from the earlier discussion.
3rd factor: ISO speed = sensitivity of film/detector.
Have to change film to change speed but can set like A and T for digital cameras.
Double / half speed = 1 stop: reciprocity again.
A, T and S values can all be changed in steps which double the amount of light used in the camera. The numbers for A may not seem like that because doubling the aperture lets in four times the amount of light - a square relationship. The steps are called "stops" in photographic circles, referring to the act of stopping down a lens to reduce its aperture.
Standardised whole-stop values which are likely to be available for your lens, camera and film are shown in the next table, each row of which may extend further in each direction.
|A||f32, f/22, f/16, f/11, f/8, f/5.6, f/3.5, f/2.8, f1.8|
|T||1000, 500, 250, 125, 60, 30, 15, 8, 4, 2, 1, 2", 4", 8"|
|S||50, 100, 200, 400, 800, 1600, 3200|
|← reduceincrease →|
Because the variables all have steps which double the amount of light, altering one of them by a certain number of steps can be exactly counteracted by adjusting another one by the same number of steps in the opposite direction. This very useful relationship is known as "reciprocity". In certain circumstances it fails, as we shall see later, but for the moment assume it is perfect because in common shooting conditions it is.
Most cameras and lenses allow finer control for A and T by having steps in half-stop intervals, as shown by the inserted red values in this table:
|A||f/32, f/27, f/22, f/19, f/16, f/13, f/11, f/9.5, f/8, f/6.7, f/5.6, f4.5, f/3.5|
|T||1000, 750, 500, 350, 250, 180, 125, 90, 60, 45, 30, 20, 15, 10, 8, 6, 4, 3, 2, 0"7, 1, 1"5, 2", 3", 4"|
|← reduceincrease →|
Sensitivity (S) settings on digital cameras are sometimes in whole stop intervals, in which case one step change in S is equal to 2 step changes in the opposite direction for A or T.
Some cameras allow even finer control, in one-third stop intervals. (See the discussion of RAW mode below for reasons why this is probably unnecessary for digital cameras.)
The following tables summarise the direct consequences of varying each of the parameters by itself.
|Value||Primary effect||When to use||Consequence|
|Large A||More light captured||May have to use in low light||Small depth of field|
|Small A||Less light captured||Can use in bright light||Large depth of field|
Small depth of field means that only objects in a narrow range of distances from the camera will be in focus at once. This is great for making particular objects stand out against a blurred background but sometimes we want to show everything clearly, in which case a large depth of field is often required. Aperture settings control the depth of field.
|Value||Primary effect||When to use||Consequence|
|Long T||More light captured||May have to use in low light||Blurred motion|
|Short T||Less light captured||Can use in bright light||Frozen motion|
Sometimes we deliberately want moving objects to be blurred, to portray their motion. At other times we want to freeze moving objects. T is the main factor for controlling the sharpness of moving objects.
There is another consideration related to T. Long exposure times mean that objects will blur due to shake if the camera is held in the hand. So the T setting determines whether a more stable base, such as a tripod, is necessary. We will come back to this later.
|Value||Primary effect||When to use||Consequence|
|Large S||More sensitive||May have to use in low light||More grain/noise|
|Small S||Less sensitive||Can use in bright light||No grain/noise|
Grain is visible in more sensitive film after development because the silver grains which result from exposure to light really are larger. The digital equivalent is random electronic noise which becomes a problem when the detector signal is amplified more, to boost sensitivity. Detector chips can be cooled to reduce this noise, and this has to be done in astronomical cameras using very long exposure times to photograph faint objects. Cooling is not an option in consumer cameras, where it is not considered to be necessary.
The preceding tables summarised useful information but when we want to take a photo we are really thinking about things the other way round. To achieve a good picture with certain requirements, how should we set A, T, S and W? The next table provides some general answers.
|Type of light (daylight, tungsten, etc)||Set appropriate W (use appropriate film)|
|Bright light||Reduce A and/or T and/or S|
|Low light||Increase A and/or T and/or S|
|Want to focus on very narrow range, background blurred||Increase A, so reduce T and/or S|
|Need to get back and front of subject in focus at once||Reduce A, so increase T and/or S|
|Subject moving towards/away from camera||Reduce A, so increase T and/or S|
|Subject moving side to side or up/down||Reduce T, so increase A and/or S|
|Want to minimise grain (noise)||Reduce S, so increase A and/or T|
Cameras have automatic settings which do an excellent job in many circumstances but a camera has no way of knowing your order of priority of the requirements in the table above for a particular situation.
Often in real life there is some conflict between requirements so the skill lies in making the best compromise. For example, in trying to photograph a bird in flight on a cloudy day there is a conflict between using the shortest possible exposure time to freeze the motion and needing to capture as much light as possible which will therefore have to be done by increasing aperture or sensitivity. Increasing aperture reduces depth of field. That implies focussing will have to be more accurate but that could be tricky for a subject which might suddenly change direction. Therefore higher sensitivity is probably the thing to choose rather than larger aperture. It is too late to make such decisions when the bird is in the view-finder, so some forethought is necessary. I would probably set S and T beforehand and let the automatic metering adjust A when variable cloud cover alters the amount of light available. But then again, it also depends on the type of bird! Is it a gull, enjoying a slow ride on a thermal, or a swallow darting around to catch insects? It is much easier to keep the gull in focus as it moves slowly, so setting a large-ish value for A first and allowing T to vary with the lighting might be better in that case.
Practice is evidently necessary to enable quick assessment of the factors when the need arises.
If there is time (not for birds in flight) everything can be set manually, for complete control. This would be a good way for beginners to learn the effects of the different settings.
First set white balance, W, either by loading the appropriate film type for the lighting (daylight, tungsten, etc) or by setting the digital camera to automatic white balance. Also decide on a suitable setting for sensitivity, S, based mainly on how much light is available and whether some graininess can be tolerated. Remember that a film is chosen by both its W and S properties but in a digital camera these are two independent settings.
Then use automatic metering in the camera to find a combination of A and T which give the right exposure. Make a note of the two values. Switch the camera to manual mode and set the measured values of both A and T. Then adjust A in one direction for desired depth of field, noting by how many steps it was changed, and alter T by the same number of steps in the opposite direction. Or adjust T first because of motion, then change A in the opposite direction.
For film exposures of several seconds the simple opposite relationship between A and T starts to break down. This is called "reciprocity failure" - the film response is not linear for long exposure times. Digital cameras do not suffer in this way.
People seem increasingly to be using the term "EV" instead of the photographer's traditional "stop". They are the same thing. EV stands for Exposure Value.
Both terms are used for a relative scale in which 1 unit means a factor of 2 change in the amount of light recorded by the camera. From any given camera setting in terms of the 3 variables, exposure time T, aperture area A2 and ISO sensitivity, doubling any one of them is an increase of 1EV, halving any 2 of them is a change of -2EV, etc.
Surprisingly, the decision as to whether to shoot in RAW mode or JPEG mode can also affect the choice of settings for A, T and S. We will first consider the pros and cons of the two modes before seeing why it can affect how we expose.
The modes are named after the type of file created to store the image data on a memory card or disc.
JPEG (Joint Photographic Experts Group) really refers to an agreed standard algorithm (computational procedure) for storing image data in a very compact way. It is extremely useful because it enables multi-megapixel images to be held in only a few tens of kilobytes rather than several megabytes. However, this comes at a cost. The compression done by the algorithm is lossy. In other words, details are thrown away and can never be recovered. The displayed image from the file may look fine when seen on a small screen or across a room but may not bear close inspection. Edges are softened and areas which should have uniform colour become rippled. There is an adjustable parameter which determines how much detail to throw away, in a trade-off between quality and file size. At the highest quality setting the files are still measured in megabytes and the loss of detail may be hard to see but nevertheless some information has still been lost.
JPEG files have a further restriction which is less obvious. They can only store 16.8 million distinct colours in an image. That may seem a lot but look at it another way. That number is the cube of 256. It means that for each of the 3 primary colours (red, green and blue) which are used in both film and digital detectors (and indeed in the human eye) a JPEG file can only distinguish 256 different levels. That may be fine if the image as photographed is suitable to be printed without any adjustment but as soon as any enhancement is required the limitations become apparent.
256 is 2 raised to the power of 8, so technically we say that each of the 3 colour channels is recorded with at most 8-bit resolution in a JPEG file.
Each major camera manufacturer has its own algorithm for compressing image data into files without any loss of information. Typically the file size is about one sixth of the size of the original data - a few megabytes rather than tens of megabytes. Processing software from the manufacturer can read the files again and some standard applications, such as the full version of Adobe Photoshop, can also read them. My application, GRIP, can read them too. This lossless compression enables the original data captured by the detector to be stored compactly; not as compactly as JPEG, but the original data are still able to be fully recovered.
Most consumer digital cameras capture 12 bits, or 4096 distinct levels, per colour channel. So for each colour we have 16 times as many levels to choose from, compared to 8-bit images. That makes it possible to adjust brightness or contrast in different parts of the image in different ways to bring out details (rather like the old dark-room technique of shading areas under the enlarger). Attempting to do this with only 8 bits per channel easily gives rise to a contouring artefact: supposedly smooth variations of colour across an area become obviously stepped.
To make JPEG files a camera has to first select from the full raw data the 256 levels for each channel (out of 4096) which best represent the appearance of the image. To do that it may apply some colour saturation or contrast enhancement processing and sometimes parameters for such processing are adjustable on the camera's menu. The stored result after JPEG compression is a (usually pretty good) guess at what ought to be suitable for direct printing. It will look brighter or punchier than a RAW version of the same scene but it will not be capable of adjustment in the way that the RAW one would. The RAW file can be processed in an application such as Photoshop to make a far superior final result. The results of such processing would then be stored in a standard file format which does not lose any data - typically either Adobe's PSD format or the less company-specific TIFF format.
So I always shoot in RAW mode. I don't throw data away that I have carefully captured. I retain maximum flexibility for getting the most out of the image when I want to present it.
Some further considerations:
- Increasingly consumer cameras are providing 14 bits (16,384 levels) per channel in RAW mode, giving even greater capabilities when processed.
- When it comes to astrophotography, particularly for measuring star brightnesses, we need the greatest range of magnitudes to be correctly represented, without saturating, and yet to be visible above the background. This is discussed in some detail on another page: click here.
There is a danger with many kinds of scene that the brightest features may be overexposed. Meters give an average reading over an area, which may be small or large depending on settings but nevertheless it is not practical always to find the very brightest and darkest points in the scene and expose perfectly between those extremes. So quite often some points in the image will exceed the maximum brightness which can be recorded at the current A, T and S settings. Those bright points will therefore all be recorded at that one maximum level, making a flat area from which no detail can subsequently be recovered.
RAW mode, by enabling local enhancement of areas of the image, makes it feasible to underexpose so that saturated areas are less likely to occur. Therefore, I always set my camera to deliberately underexpose by a certain amount. I leave it to the reader to determine a suitable amount. However, there is yet another thing to be aware of when doing this: the darkest areas are more susceptible to noise (grain) so underexposing by too much will make the recovered detail in the shadows noisy. This is another factor to throw into the balancing act we have been discussing.
Magnification is not usually specified in absolute units because there is no way of knowing at the time when a photograph is taken, at what size it will later be printed or displayed. Instead, when discussing a particular format of camera, such as 35mm, everyone has got used to describing magnification in terms of the focal lengths of lenses. For 35mm film cameras, 50mm is often called a standard lens (unit magnification). Lenses with shorter focal lengths are "wide angle", cramming more into the view-finder and therefore magnifying less. Lenses with focal lengths greater than 50mm are "telephoto", magnifying more.
Many consumer digital cameras have sensors smaller than the 36 x 24 mm film frame of a true 35mm camera but lens focal lengths are nevertheless quoted in 35mm terms. The manual for the camera will state a multiplying factor by which the lens focal length should be multiplied to obtain the equivalent magnification to a true 35mm camera. For example, the low to medium range Canon EOS digital SLR's have a factor of 1.6. For these cameras the standard lens is really 50 x 1.6 = 80mm. This means they cannot go as wide as real 35mm cameras but for telephoto work they have a big advantage: a 300mm lens is equivalent to 300 x 1.6 = 480mm.
In macro photography (photography of small, but not microscopic, objects) absolute magnification, in terms of the size of the image on the detector compared to the original object, is sometimes stated. It is given as a multiplication factor, from the object size to the image size. If you think about it though, your usual lenses are greatly reducing the scene onto the detector, so achieving an absolute magnification even as great as x1 is not easy. For a start, you have to move the lens even further away from the detector in order to focus: hence the availability of extension rings. Another consideration is that camera lenses are rarely designed to focus on objects much less than about 30cm away, so we end up reversing the lens with special adapters to use it optically truer to the way it was designed (the object being closer to the lens than its image).
We mentioned earlier a fairly self-evident fact that longer exposure times result in blurred images when the camera is hand-held. In such circumstances a tripod or other stable base is required. It is really the combination of exposure time and lens magnification that matters. The more the scene is magnified, the more likely that blurring due to shake will be visible in the photograph.
A common rule of thumb is that the reciprocal of the focal length in mm is about the longest exposure time in seconds that can be used if the camera is hand held. So a 50mm lens is safe to hold in the hand for exposure times shorter than 1/50th second. For a 200mm lens it is necessary to use less than 1/200th second or use a tripod.
Image-stabilised lenses make it possible to hand-hold the camera for exposure settings of 2 or perhaps 3 whole stops longer than normal, for a given magnification.