Exposure and image control
CAMERA EXPOSURE
The term exposure in photography describes the total quantity of light energy incident on a sensitive material, which in general terms is the photographic exposure. It may alternatively describe the process of controlling the light energy reaching a sensitive material in a camera, which is more specifically the camera exposure.
As described in Chapters 1, 8 and 11, two parameters are used to control photographic exposure: the amount of light incident on the imaging sensor and the duration of the sensor’s exposure. This is described by Eqn 8.1, the reciprocity equation:
where H is exposure, E is the subject illuminance and t is exposure duration.
As described in Chapter 11, in a camera, the amount of light, i.e. the subject illuminance (E), is controlled via the lens iris diaphragm. The iris diaphragm is calibrated in units of relative aperture (N) – see Chapter 2. The duration of exposure (t) is controlled by the shutter speed. Because of the reciprocal nature of Eqn 12.1, a range of different combinations of aperture and shutter speed (termed ‘equivalent exposures’) will provide the same quantity of incident radiation on the sensitive material.
A number of factors influence the optimum camera exposure for any imaging situation and the resulting quality of the image. Fundamentally, the results from any exposure are dependent upon the relationship between four factors: the radiation coming from the scene, the transmission of that radiation through the optical system, the exposure on the sensor, and the sensitivity and characteristic response of the sensor to that radiation. We introduce the parameters influencing these aspects of exposure below, to be further elaborated later in the chapter.
In any given scene there will be a range of areas of different luminance which, as introduced in Chapter 8, may be characterized by the subject luminance ratio or the subject luminance range. These will provide a range of exposures to the sensor. Various aspects of the scene and the lighting conditions will influence the subject luminance range. Any scattered light results in a reduction in scene contrast. The luminance range is reduced by flare and, as a result, the tonal increments in the shadows are compressed. The colour of the subject and its surface reflection properties will affect the amount of light directed at the sensor. Additional factors include the type of light source(s), the direction and intensity of the light, and the distance between the light source and the subject, which follows the inverse square law (see Chapter 3). The greater the distance between subject and light source, the lower the intensity of the light that reaches the subject. Subjective preferences also influence the photographer’s choices for optimum exposure.
As already described, the amount of light from the scene reaching the sensor is controlled by the aperture and shutter speed combination, but is also affected by other variables which are summarized by Eqn 6.61 in Chapter 6. This defines image illuminance as a result of the combined effects of subject luminance, transmittance (which includes lens transmittance, the effects of filters in front of the lens, and in digital systems, the transmittance of anti-aliasing filters, infrared filters and sensor microlenses), lens aperture, magnification and natural vignetting (see Chapter 6).
The sensitivity of the sensor (the ISO speed) is determined by the sensor type and characteristics, as well as subsequent processing: development conditions for film and subsequent analogue and digital processing of the signal from image sensors. Chapter 24 discusses speed and sensitivity of sensors in detail. The characteristics of the sensor, in terms of contrast reproduction and dynamic range, described by the characteristic curves of films and the transfer functions of image sensors, will ultimately define the relationship between the original scene luminances and their reproduction in the image. This is discussed later in the chapter and tone reproduction is also the subject of Chapter 21.
Figure 12.1 Exposure relationships. The diagram shows the relationships between ISO speed, exposure duration and aperture setting to give equivalent exposures (shown here as the areas of rectangles). A long exposure duration implies a reduced aperture and a low ISO speed setting, a larger rectangle, for a subject of given luminance.
Subject luminance range, ISO speed, lens aperture and shutter speed are therefore the four essential quantities of camera exposure; they are also the ones which are, to a greater or lesser extent, the easiest to change (the other factors detailed above are often inherent properties of the system). The selection of a suitable combination of aperture and shutter speed at a given ISO speed is a primary creative control in photography and may be dictated by the subject itself or the treatment required (see Figure 12.1). Together they influence a number of important image attributes, such as depth of field, motion blur, image sharpness, contrast and noise. The ability to vary the illumination on the subject, where possible, provides more options for varying the other factors. Camera exposure may be determined by practical trial, e.g. by an estimate based on previous experience, followed by assessment of the resulting image, but measurement by some form of light meter is quicker and usually preferable.
EXPOSURE RELATIONSHIPS AND LOGARITHMS
As described in Chapter 4, the human visual system (HVS) has an approximately logarithmic response to light. This means that light intensity must change by an approximate factor of 2 to be perceived by the HVS as a just-noticeable difference (JND) in brightness. This has been shown by both physiological and psychophysical studies on contrast perception. Psychophysical experiments led the German psychologist Ernst Heinrich Weber to the conclusion that just-noticeable differences in the perception of stimuli are constant fractions of the magnitudes of original stimuli. The German psychophysicist Gustav Theodor Fechner extended this work and described the mathematical relationship between the intensity of subjective sensation and the stimulus intensity. This relationship was shown to be logarithmic. Fechner named this as ‘Weber’s Law’ but it was later renamed as the ‘Weber–Fechner Law’.
Based on the Weber–Fechner Law, the following equation describes the HVS response:
where S is the magnitude of sensation, C is a constant and R is the magnitude of the stimulus.
Camera exposure is also controlled according to a logarithmic scale. Changes on the camera exposure scale are described in terms of stops. A change of one stop results in the doubling or the halving of the amount of light reaching the image plane. As a result, when the aperture is changed by one stop the aperture area is doubled or halved (see below), while when the shutter speed is changed by one stop the exposure time is doubled or halved. As described in Chapter 24, ISO speeds are also based upon a logarithmic (base 2) scale, with changes between consecutive ISOs approximately corresponding to an increase or decrease in sensor sensitivity by a factor of 2. This means that, for the same scene, the overall exposure required at 100 ISO will be twice that required at 200 ISO, a change of one stop.
Relative aperture
The aperture f-stop values represent relative apertures which are calculated from the focal length of the lens divided by the diameter of the entrance pupil for a lens focused on infinity:
where N is the relative aperture, f is the focal length of the lens and d is the diameter of the entrance pupil (see also Eqn 2.23 in Chapter 2 for further explanations). The f-number, written as f/, is the value of N, and expresses the ratio of the focal length to the value of the diameter of the entrance pupil: an f-number of 5.6, for example, indicates that the focal length is 5.6 times the pupil diameter.
It can be seen from Eqn 12.3 that when the diameter of the entrance pupil increases for constant focal length, the f-number decreases. As previously mentioned, a change of one stop results in the doubling or halving of the amount of light that passes through the lens, because the area of the aperture is doubled or halved. To double the aperture area, the diameter must be increased by a factor of 
(approximately 1.4). This results in a series of f/numbers with a one stop difference: f/1, f/1.4, f/2, f/2.8, f/4, f5.6, f/8, f/11, f/16, f/22, f/32, f/45, f/64, etc.
Shutter speed
The shutter speed is measured in fractions of a second (see Chapter 6). As described earlier, there is a range of combinations of f-number and shutter speeds that result in equivalent exposure at a given ISO speed (see Figure 12.1). Selection of a suitable combination depends on the specific scene and lighting conditions. For example, larger apertures allow faster shutter speeds, but result in reduced depth of field (see Chapter 6). Subject movement requires a minimum shutter speed, dictated by the speed of the subject, to produce a sharp image. A slow shutter speed, however, may be used in this case, if the effect of motion blur is desired in the final image. Slow shutter speeds may introduce blurring in the image due to camera shake, especially when using lenses with long focal length (see Chapter 11). In many cases, when a specific combination of f-number and shutter speed is desired, changes in the illumination may be necessary.
The choice of shutter speed is limited when using electronic flash with cameras which incorporate a focal-plane shutter. As described in Chapter 11, the shutter needs to be open for long enough for the entire image-sensing area to be exposed simultaneously by the flash. This limitation does not apply to cameras with leaf shutters, because the flash is fired when the shutter is fully open. The flash synchronization speed (the fastest shutter speed that can be used with flash) depends on the camera model. At the time of writing, high-quality SLR cameras provide flash synchronization speeds as fast as 1/250th of a second, while older cameras synchronize flash at lower shutter speeds, for example 1/60th of a second. It should be noted that when using slow shutter speeds the ambient light affects the exposure in addition to the illumination from the flash.
SUBJECT LUMINANCE RATIO
As introduced in Chapter 8, a primary classification of a scene is by its subject luminance ratio, sometimes called the subject luminance range. This is the ratio between the maximum and minimum luminances within the scene and is numerically equal to the product of the subject reflectance ratio (the ratio between maximum and minimum reflectances, dependent upon surface characteristics in the scene) and the lighting ratio (the ratio between the maximum and minimum illumination levels, dependent upon the characteristics and position of the light sources). Together, these define variations in subject luminance and therefore subject contrast.
For example, a subject whose lightest and darkest zones have reflectances of 0.9 and 0.09 respectively, i.e. a reflectance ratio of 10:1, when illuminated such that there is a lighting ratio of 5:1 between maximum and minimum illumination levels, has a subject luminance range of (10 × 5):1, i.e. 50:1. A high luminance ratio denotes a subject of high contrast. Luminance ratios from as low as 2:1 to 1,000,000:1 (106:1) may be encountered.
In practice, the luminance range is the difference in stops between the brightest highlights and the deepest shadows. The average luminance ratio based upon measurements from a large number of indoor and outdoor scenes is 160:1. The majority of normal scenes will have subject luminance ratios between 27:1 and 760:1. As each increment of one stop in exposure results in an increase in luminance by a factor of 2, this maximum corresponds to an exposure range from shadow to highlight of between 9 and 10 stops (29 = 512 and 210 = 1024).
Based upon appropriate scene measurements, it is usually assumed that an average outdoor subject has a luminance ratio of 128:1 or 27:1, corresponding to a seven-stop (or 7EV, exposure values – see later) range. Such a subject is also assumed to have an average reflectance of 18–20%.
Integration of individual luminance levels gives a value for the average reflectance of a scene, which depends on the subject luminance ratio and is used in reflected light measurements. As detailed later, there are a number of different reflected metering methods, using different areas of the scene. When using reflected light readings with hand-held exposure meters, however, whole scene integration to produce an average reflectance is a more common method. Most meters are calibrated to a mean reflectance of 18% grey, which can cause exposure problems for scenes with very high or very low luminance ratios, as the mean reflectance of the scene will be significantly different from this value (illustrated later in Table 12.1). In such cases in-camera metering using one of the various specialized matrix metering modes, or spot metering from highlights and shadows and taking an average, both discussed later in the chapter, can prove more successful. Exposure correction factors, also described later, may be applied for particular types of subjects.
A more fundamental problem arises when the subject luminance range exceeds the dynamic range of the sensor (described in the next section). In this case, detail will be lost from the highlight or shadow regions of the image, depending on where the exposure is placed. The effect of this on image quality is dependent on various factors, such as how much of the image area is affected (detail lost from small specular highlights, for example, may not be a problem) and on the type of sensor (highlight clipping in digital images is generally much more noticeable than shadow clipping, or the loss of highlights in film). Strategies to alter the image contrast may be preferable, such as the reduction of the subject luminance ratio by the use of additional illumination.
DYNAMIC RANGE OF SENSORS
Several definitions of dynamic range exist. Most often dynamic range describes the ratio between the largest and smallest possible signal values of light intensity. Dynamic range can be either input-referred (ratio of scene light intensities, the scene or subject luminance range described above, with luminances measured in cd m−2) or output-referred (ratio of the response signals delivered by an imaging sensor). The dynamic range of both silver halide and electronic imaging sensors (output-referred dynamic range) refers to the ratio of the highest to lowest level of radiation (in visible imaging light) sensed by the sensor. It generally relates to an imager’s ability to record both bright and dark objects in a scene with the same exposure settings.
The dynamic range of still digital cameras incorporating CCD or CMOS sensors is defined as the ratio between the maximum charge that the sensor can collect and the minimum charge that is just above the sensor’s noise level (see Chapters 9 and 14). Detection of the smallest possible light intensity is limited by the noise floor, as any signal below it is undetectable. The largest possible captured intensity is limited by saturation. Dynamic range is usually expressed in logarithmic units:
where Imax and Imin are the maximum and minimum intensities respectively; FWC and NF represent the full-well capacity and the noise floor respectively (see Chapter 9).
If the image sensor has a linear response curve, then only the saturation and the noise floor levels limit dynamic range. Extending the dynamic range requires the shifting of saturation to higher input signal levels by giving the sensor a non-linear response curve. In such a case the local slope of the response curve (called incremental gain) decreases so that saturation is reached at much higher luminance levels than for a linear imager (see Figure 12.2). However, lowering incremental gain will in turn decrease incremental signal-to-noise ratio (iSNR; defined as the ratio of the input luminance signal to the luminance-equivalent noise that is equal or greater to 1.0, also known as noise equivalent contrast, used as a criterion for detection), meaning that there is an incremental signal-to-noise trade-off when increasing dynamic range. A more general definition of dynamic range that is also valid for non-linear sensors is expressed as the luminance range between Imax and Imin in which incremental signal-to-noise satisfies the minimum threshold criterion of registerable signal intensity. In such cases dynamic range is given by Eqn 9.1, which expresses dynamic range in decibels (dB) and uses a scale factor of 20. The threshold criterion is application specific: 1.0 is the theoretical detection limit and 10 the requirement for ‘acceptable’ photographic image quality.
The output-referred definition of dynamic range of electronic sensors can lead to dynamic range numbers built merely on technical specifications, such as the bit depth of an analogue-to-digital (A/D) converter and has limited value in predicting dynamic range. The relationship between input and output signals is established by the imager’s response curve (Chapters 9 and 21). Limited dynamic range and quantization in digital capturing devices thus lead to inadequate data representation and artefacts such as contouring and tone compression.
Scene luminance ranges are often greater than the dynamic range of imaging sensors and thus a single exposure often results in losses in the shadows or in the highlights. The part of the scene luminance range where the losses occur depends on the choice of exposure settings. Lower exposure (i.e. underexposure) allows the capture of highlights by avoiding sensor saturation but results in clipping the shadows and thus losing any shadow detail. On the other hand, higher exposure (i.e. overexposure) allows a good representation of shadows, but the highlights are clipped (‘burnt’). Figure 12.3 shows slightly underexposed (a) and overexposed (b) images of a high (relative to the sensor) luminance range scene and the corresponding histograms, showing clipping of the shadows (a) and the highlights (b).
Figure 12.2 Extending dynamic range with a non-linear response curve.
Adapted from Hertel (2009).
Digital images captured using a series of different exposure settings allow coverage of a wider scene luminance range, revealing more detail than would have been possible by a single shot. The process is the most common way to achieve high-dynamic-range imaging and is usually achieved by the following steps: (i) capture of multiple exposures of the same subject, (ii) camera response function estimation, (iii) high-dynamic-range construction by superimposing all subject exposures, and (iv) tone mapping to the display or the print media. High dynamic range is briefly discussed later in this chapter and in Chapter 21.
The relationship between subject luminance range and film density, as well as different factors influencing the exposure latitude of silver-based media (defined as the minimum camera exposure required to produce adequate shadow detail without losses of highlight detail), is described in detail in Chapter 8.
OPTIMUM EXPOSURE AND THE TRANSFER FUNCTION
Figure 12.4 illustrates the response curves of a conventional black-and-white negative film (a) and that of a typical CCD, before encoding (b). In both media, ‘correct’ exposure of the subject luminance range should record most of the scene luminances in the linear part of the response curve. However, the characteristic curve of negative material possesses a long toe. The part of the curve used by a ‘correctly exposed’ negative includes part of this toe and the lower part of the straight-line portion (see Chapter 8 for details). Further, because of the advantageous shape of the photographic film characteristic curve, the use of the shoulder in scene exposure on film is not as detrimental as with electronic sensors, since the slope of the curve at this point is relatively low, allowing a smoother gradation in the recording of highlights. Digital imaging sensor responses curves are often mapped (usually during encoding) to responses similar to those of film media for this very reason (see also Chapter 21).
Figure 12.3 The effects of overexposure and underexposure on an image with a high luminance range. (a) Underexposed image and histogram. (b) Overexposed image and histogram.
Original image © Elizabeth Allen
Figure 12.5 shows the positioning of a typical scene luminance range on different parts of the CCD response curve, depending on exposure. Here again, the correctly exposed scene luminance range is fully falling on the linear part of the curve and thus it is reproduced without compression. Underexposure leads to clipping due to exposure falling at the noise floor, whereas overexposure leads to clipping due to sensor saturation. In both cases the reproduced luminance range is reduced, resulting in a reduction in image contrast – illustrated by the compacted output range in the figure.
EXPOSURE METERS
Exposure meters (or light meters) are devices which measure the reflected light from a subject, or incident light to a subject, and express its relationship to the sensitivity of the film or ISO sensor setting used, in terms of f-numbers and shutter speed settings.
Hand-held exposure meters allow the use of both incident and reflected light measurements and are still used to some extent in studio settings. Incident light measurements in particular can counteract the exposure problems caused by scenes containing large areas of bright or dark tones, which can produce anomalous reflected light measurements. However, partly because of the far greater range of digital sensor characteristics compared to those of film, and the many developments in complex in-camera metering systems, alongside the ability to immediately review images and adjust exposure, the use of external metering with digital systems is less widespread than it was when film predominated.
In-camera metering systems, although only able to measure reflected light, offer many sophisticated metering modes and directly relate the light reaching a sensor to its response; they have therefore found more widespread adoption in many types of digital imaging. It should also be noted that the use of the image histogram to establish correct exposure with digital cameras is common practice (see later in this chapter). This has led to a slight difference in the function of the exposure meter with a digital camera. Exposure metering often provides a starting point from which exposure is adjusted to produce a desired distribution of tones, rather than a final decision for shutter speed and aperture.
The structure and operation of hand-held exposure meters are described below. Chapter 11 covers in-camera metering systems in some detail, although the various metering modes are discussed later in this chapter.
Hand-held exposure meters
An exposure meter consists of a light-sensitive cell, some means of limiting its acceptance angle, an on–off switch and a range-change switch, a power source (as necessary), a scale in light values, a set of calculator dials or readout of exposure data, which can be in matching pairs of f-numbers and shutters speeds or EVs, and a diffuser for incident light readings. The diffuser is domed and is placed over the photocell (see Figure 12.6).
Figure 12.4 Correct exposure of scene luminances results in the use of the linear part of the response curve in both (a) analogue and (b) electronic sensors.
Figure 12.5 The effects of underexposure and overexposure on the positioning of the subject luminance range on the response curve of a CCD sensor.
A variety of types of photocell are in use, including the early selenium ‘barrier-layer’ photovoltaic cells, cadmium sulphide (CdS) photoresistors, silicon photodiodes (SPDs) and gallium arsenide phosphide photodiodes (GPDs). The relevant properties of CdS and SPD types are listed in Table 12.1.
Figure 12.6 Block diagram of the principal components and features of an exposure meter. Ambient light, flash and colour temperature may all be measured. B, battery; BC, battery check; BF, baffle; CT, colour temperature; D, display; DF, diffusers; Eλ, spectral illumination; F, filter; OA, optical attachments; P1, P2, P3, photocells.
Selenium cells were the earliest cells to be used in exposure meters and are obsolete today. They do not require batteries but have a large cell area. The response of the selenium cell to light is dependent upon cell area. A plug-in booster cell can increase sensitivity by some 2EV. The battery-operated CdS was a newer type of cell, with higher sensitivity than the selenium cell. Higher sensitivity means that the size of the cell is reduced, an important factor for built-in camera exposure meters. A CdS photocell meter may have a resistor network to change its response range; alternatively it may use a neutral density filter or a small aperture in front of the cell. Its response may be affected by its previous exposure to bright light and it may take a few minutes for this ‘memory’ effect to disappear. Today the most common cells are SPDs. Some cameras incorporate GPD cells in their exposure metering system. Both SPDs and GPDs have a very fast response to light changes and for this reason they are also used in flash measuring devices.
The spectral sensitivity of the various types differs markedly (see Figure 12.7). The CdS and SPD types have high infrared (IR) sensitivity. The SPD is usually encapsulated with a blue filter to absorb IR to reduce this sensitivity and is referred to as ‘silicon blue’ cell. A selenium cell is also usually filtered to match its spectral response to that of the standard CIE photopic visual response curve (Chapters 4 and 5). GPD cells are not sensitive to IR radiation.
Figure 12.7 Spectral sensitivity of photocells. The graphs of relative spectral sensitivity against wavelength show the approximate spectral response or sensitivity of different photo-cells used in exposure metering systems. (A) Visual spectral sensitivity and a filtered selenium cell. (B) CdS photoresistor. (C) Silicon photodiode (SPD) fitted with a blue filter (‘silicon blue’ cell). (D) Gallium arsenide phosphide photodiode (GPD).
Acceptance angle
Both CdS and SpD photocells are small devices with a response that is independent of cell area, but require special optical arrangements when used in exposure meters. All photocells are non-imaging devices, which integrate the incident flux. The response is restricted to the subject area covered by the camera lens, or a smaller area. Alternative techniques of integrated, small area and spot readings are used.
Traditionally, the selenium cell meter was limited to have an acceptance angle of some 50°, corresponding to the field of view of the standard lens of a camera format, but with a much more rectangular aspect ratio, to reduce the effect of a bright sky area. The removal of the baffle increases the acceptance angle to some 75°, to assist sensitivity. The much smaller CdS and SPD cells may simply be located at the bottom of a cylindrical black-lined well or have a positive lens bonded to the receptor surface. These arrangements give acceptance angles of some 25–30°, allowing more selective readings of zones of the subject area. A small viewfinder may help to aim the meter cell at the desired area. Accessory devices may reduce this angle even further to 15° or even 5°. The spot meter is a specialist meter of acceptance angle 1° or less. Hand-held spot meters include a viewing system. Spot metering is often available in SLR cameras as one of two or three alternative means of selective image measurement (see later).
The measurement produced by an exposure meter is based on the assumption that the scene being measured integrates to a mid-grey. The grey value for which exposure meters are calibrated is the middle grey on a geometric scale from white to black (18% reflectance). The meter cannot therefore assess whether a subject is bright or dark, which can lead to errors in exposure readings for scenes containing a substantial percentage of light or dark areas. Measurements returned by exposure meters are also based on the assumption that the distance between the lens and the focal plane is equal to the focal length of the lens. Correction of the exposure is therefore necessary when using extension bellows (see the later section on exposure factors).
There are two types of exposure meters depending on the method used to measure light: incident light and reflected light exposure meters. Hand-held exposure meters can measure both incident and reflected light. Built-in exposure meters in SLR cameras measure the intensity of the light reflected from the subject and pass through the lens (TTL). Non-SLR cameras have a photocell in front of the camera body or lens (see Chapter 11).
Exposure values
The exposure value (EV) system originated in Germany in the 1950s and was intended as an easily mastered substitute for the shutter speed/aperture combination, giving a single small number instead of two (one of which was fractional). The EV system is based on a geometric progression of common ratio 2, and is thus related to the doubling sequence of shutter and aperture scales. The relationship used is:
where N is the relative aperture and t is the shutter speed.
Change of the base of the logarithm gives:
or
For example, a combination of 1/125 second at f/11 from Eqn 12.6 gives an EV of 14. Similarly, all exposures equivalent to this pair will give an EV of 14.
Although no longer used for such purposes, the idea of the EV scale is useful in exposure measurement practice, remembering that an increment of 1 in the EV scale indicates a factor of 2 in exposure, or a change of one stop. A primary use is to give a quantitative measure of the sensitivity of a metering system by quoting the minimum EV that can be measured for an ISO speed of 100. For example, a hand-held exposure meter using an SPD cell may have a measuring sensitivity of between EV − 2 and approximately EV 23. A sensitivity of EV − 2 can allow metering in moonlight. The use of various attachments such as spot meter arrangements or fibre-optic probes reduces the sensitivity considerably; a high initial value is therefore needed for meters intended for such accessories.
In-camera meters are limited in their sensitivity by the light losses involved in their optical systems, and by the maximum aperture of the lens for full aperture TTL metering. For comparisons this aperture must be quoted: for an f/1.4 lens a sensitivity of EV 1 at ISO 100 is typical. The operating limit of an in-camera autofocus system is quoted similarly. Another use of the EV system is the EV graph (see Figure 12.8), where the scales on the rectangular axes are shutter speed and aperture. A series of diagonal lines connecting pairs of values of equivalent exposures represent the EV numbers. These EV loci may be extended to an adjacent graph of ISO speed and subject luminance. On such a graph it is possible to show the behaviour of a particular camera, for example in terms of the envelope of performance with the range of settings available. Additionally, in the case of an automatic camera the different exposure modes and forms of programmed exposure can be compared.
Finally, EV ‘plus’ and ‘minus’ scales are commonly used on the exposure compensation and autobracketing controls of automatic cameras, both film and digital SLR cameras, and to indicate filter factors and exposure compensation for interchangeable focusing screens.
Exposure factors
Exposure factors are the multiplying numbers used to correct exposure when using filters on the lens (filter factor), or bellows to extend the distance between the lens and the focal plane (bellows extension factor). They are used with non-TTL exposure meters. Exposure correction using the exposure factor is achieved by either multiplying the exposure time by the exposure factor, or by converting it into stops and correcting the f-number on the lens accordingly. Conversion of the exposure factor into stops is based on the fact that a change of one stop doubles or halves the exposure. So a factor of 1 means no change in exposure, a factor of 2 means a one stop change in exposure (as each stop doubles or halves the previous stop), a factor of 4 a change of two stops, and as a general rule:
where CF is the correction factor and n is the number of stops.
The filter factor is recommended by the manufacturer and is used to compensate for any light losses which occur as it passes through the filter, before reaching the sensor. There are practical ways of determining the filter factor (see Chapter 3). The bellows extension factor compensates for any light losses occurring due to the increased distance between the lens and the focal plane. It is calculated as follows:
Figure 12.8 Exposure value (EV). Relationship between scene luminance and ISO speed can be expressed in terms of an exposure value, which is a combination of aperture and shutter speed. The rectangle indicates the typical operating range of a camera fitted with an f/2 lens.
RECIPROCITY LAW FAILURE
The reciprocity law refers to photographic exposure and takes into account the amount of light that reaches the sensor and the duration of exposure (Eqn 12.1). With silver halide materials, there are cases, however, where the exposure relationship breaks down and the law no longer holds true. This is known as reciprocity law failure (RLF) and is described in detail in Chapter 8. Reciprocity law failure depends on the duration and intensity of the exposure. Two types of RLF therefore exist: low-intensity RLF and high-intensity RLF. Low-intensity RLF, which is the case most commonly encountered, occurs with exposures where the light intensity is low, requiring very long exposure times. The effective speed of the film is reduced at very low light levels, with a greater loss of speed in the shadow areas of the image compared to the highlights. This may be explained because the low intensity of light results in photons reaching the sensitive material in much fewer numbers over a given period of time. According to the Gurney–Mott theory (see Chapter 13, Figure 13.4), below a certain number of absorbed photons, the latent image is unstable and there is a high probability of recombination of electrons into the silver halide lattice. Therefore, in the shadow areas of the image this probability of recombination is increased. The result is a decrease in the density in low-density areas of the film compared to the high-density areas. This results in higher contrast and a possible change in colour balance. The high contrast produced can be reduced by decreasing the development time.
High-intensity RLF occurs when the light intensity is very high, requiring very fast shutter speeds for correct exposure and lowering the effective film speed. A decrease in the high-density values results in a loss of contrast, which may be compensated for by increasing the development time. Changes in colour balance may also appear because the layers of the colour film are affected differently. In both cases of RLF, these colour changes may be corrected with the use of filters.
Due to the different response of the film when used under very high or low light intensities, manufacturers design emulsions so that they have optimum performance at the light intensity normally used for that film. Astronomical plates are an example, with an emulsion specially designed for very long exposures at low light intensities.
TYPES OF LIGHT MEASUREMENT
Incident light measurement
As previously mentioned, hand-held exposure meters can provide both incident and reflected light measurements. In incident light measurement, the exposure meter measures the amount of light that illuminates the subject. The measurement is taken from the subject position, with the exposure meter’s sensor pointing towards the camera. It is suitable for cases where the lighting is uniform or when controlled lighting is used. It is also useful when the background is dominant. When the subject is backlit correction may be needed on the exposure. With this type of measurement, however, the reflective properties of the subject are not taken into account. Another limitation is that, if it is to be used to determine exposure of a distant scene, the exposure meter should be suitably positioned so that the light falling on it is the same as the light falling on the scene. In the majority of cases this may not be practical.
Flash meters are incident light meters which are designed to have a very fast response time to changing light conditions. Incident light measurements are easy to make with a hand-held light meter but are not suitable for an in-camera metering system.
Reflected light measurement
Reflected light exposure meters measure the light that is reflected from the various different surfaces in the scene. When using hand-held meters, depending on the exposure meter, light measurement can be conducted using a wide angle (approximately 30°) or a narrow angle, also known as spot (approximately 1–2°). With reflected light exposure meters, the reading is from the camera position, at the lens axis, with the sensor towards the subject. Care must be taken to avoid any incident light on the exposure meter’s cell because light scattering will affect the reading. Hand-held spot exposure meters include a viewer which is used to select a specific area of the subject or scene with precision. Reflected light measurement with wide-angle exposure meters is suitable for scenes with balanced highlights, middle tones and shadows because they average the luminances in the whole scene that is measured (‘integrate to grey’). For this reason, scenes with dominant highlight or shadow areas may result in underexposure or overexposure respectively. The reflective properties of the subject may also have an effect on the measurement of the exposure meter. The measured area also plays a significant role in determining correct exposure, as described in the next section.
IN-CAMERA METERING MODES
Built-in camera exposure meters measure reflected light using a range of different metering modes. Their use is described below and the mechanisms of some of these modes are also described in Chapter 11.
Spot metering
As for a hand-held spot meter, an in-camera spot meter measures a narrow angle, corresponding to approximately 4% of the area of the scene. The measurement may be from the centre of the frame or from user-selected off-centre spots, with some models supporting multi-spot modes from different areas within the scene. Spot metering is useful in high-contrast and backlit scenes, where the subject of interest is significantly darker than the background, and allows the required area to be metered correctly, without the measurement being significantly influenced by the large brighter areas surrounding it, although highlight detail will be sacrificed. Spot metering is also useful when taking an average of highlights and shadows, described later.
Partial area metering
Partial area metering takes a measurement from a larger area than spot metering, of around 10–15% of the frame, which is usually at the centre of the frame, although as with in-camera spot metering, some models allow user selection of alternative off-centre AF points in the scene. As with spot metering, partial area metering can be useful for backlit subjects and can provide a better balance of exposure between foreground and background, which is particularly important when trying to avoid clipped highlights in exposure with digital cameras.
Centre-weighted average metering
Often a default method of in-camera exposure metering, centre-weighted average metering measures the entire scene, but weights a larger percentage of the measurement towards the centre of the scene, with the periphery having much less influence on the reading. Its widespread use is partly because it produces reasonable results when dealing with scenes with a high luminance ratio such as a brightly lit sky above a darker landscape, without requiring significant exposure correction. This is the reason, in cameras, that centre-weighted average readings have entirely replaced whole-scene average metering, which is now obsolete in all but classic film cameras (see the later section on measurement of total scene luminance).
Matrix or multi-zone metering
Also called segmented, honeycomb or evaluative metering, as described in Chapter 11, these types of metering systems (of which there are many different types – see Figure 11.20) divide the scene into segments and take individual exposure readings from each of the different areas. The results are analysed in terms of average luminances and luminance differences between zones, as well as overall scene luminance, and compared to data from many different types of scenes pre-stored in the camera to decide on what type of scene is being photographed and establish optimum exposure. The zones are often weighted according to the selected autofocus point, to ensure that the region of interest is correctly exposed, but other factors may also be considered, such as subject distance, depth of field and back or peripheral lighting. Although the aim of such systems is to reduce the need for exposure compensation, and in general they are extremely successful at producing good results for the majority of scenes, results can be inconsistent for certain types of scenes, and are dependent on both the sophistication of the metering method from a particular manufacturer and on the autofocus point.
ELECTRONIC FLASH EXPOSURE
Exposure with an electronic flash can be a challenging task, especially due to the fact that the electronic flash does not incorporate modelling lights and therefore the previsualization of the final image is difficult. An advantage when using a digital camera with an electronic flash is that the image is almost immediately available for viewing. This enables the photographer to make suitable changes in the lighting if necessary. Still, due to the difficulties in previsualization, the photographer may have to take several photographs before achieving the desired result. This may be possible with still life, but with action scenes or scenes with moving subjects, it may not be possible to repeat the shot with different flash settings.
Electronic flash can be used in either manual or automatic mode. In manual mode the photographer sets the suitable f-number on the camera lens according to the distance between the flash and the subject and the guide number of the flash, which is described below in more detail. These settings also depend on the ISO setting on the camera. Note that the shutter speed is set to the flash synchronization speed for cameras with a focal-plane shutter. In automatic mode, the flash may incorporate a light sensor which measures the light reflected from the scene and interrupts the flash when the light reflected from the subject is adequate for correct exposure. Alternatively, the flash unit may use the information on scene luminance from the TTL measurement.
Lighting control with the use of additional light sources is often applied to alter the tonal range of a scene. If, for example, the contrast of the scene is high, introduction of fill-in lighting will make the shadows appear brighter and increase shadow detail. Fill-in lighting can also be introduced with the use of suitable reflectors opposite the main light source. High-dynamic-range scenes, for example when a subject is photographed indoors standing in front of a window, need the use of additional lighting to balance the foreground to the background and reduce the dynamic range of the scene, so that it more closely matches that of the sensor. A reading is taken from the background, the camera is set to correct exposure for the background and the flash is set to expose correctly for the subject.
Guide numbers
The light output and the required camera exposure for a given flash unit can conveniently be expressed by an exposure guide number, or flash factor. For ISO speed 100 and a subject distance d (in metres), the necessary f-number N is given from the guide number G by the relationship G = Nd.
This guide number is often included in the alphanumeric code designation of a flash unit. For example, inclusion of the number 45 would usually indicate a basic guide number of 45 in metres for ISO speed 100. The guide numbers are only guides, and practical tests are needed to establish their accuracy, which depends to some extent on the effects of the surroundings, ambient light level and the shutter speed used. The guide number also changes with fractional power output selection as offered by many units, some of which are able to select full, half, one-quarter, one-eighth and one-sixteenth power or less, or, when a camera motor-drive setting is selected, giving reduced power and rapid recycling. Likewise, an alteration in flash covering power to suit different lenses will also affect the guide number.
Guide numbers are devised on the assumption of a near-point source and are based on the inverse square law. A doubling of flash output or doubling of ISO speed used increases the guide number by a factor of 1.4. A quadrupling of output or ISO speed doubles the guide number. Doubling the distance between the flash source and the subject distance needs a change of +2EV in exposure (factor ×4).
While such calculations are reasonably simple to make for one flash source, it becomes increasingly difficult to do so for two or more sources, especially if they are at different distances and of differing powers. The positioning of flash-heads without the assistance of modelling lights is a task requiring experience and skill. But as a guide a simple twin-head flash, using one flash off-camera as the main light and the other on-camera as fill-in light, the fill-in light should provide approximately one-quarter to one-half the output of the main light.
The guide number system has been replaced in automatic flash units by the use of variable flash output and hence duration as a means of controlling exposure, but it is still applicable in situations such as the use of ring flash units or in lenses where a guide may be set and linked to the focusing control to give a readout of the f-number necessary for a given subject distance. Similarly it is used in autofocus cameras where the flash output is fixed, and the appropriate f-number set by the distance setting from the autofocus system.
REFLECTED LIGHT MEASUREMENT TECHNIQUES
As described above, optimum exposure may be determined by measurement of the luminance of the subject (or zones of it) or, alternatively, by measuring the illumination on the subject (incident light), followed by location of the log luminance ratio in a suitable position on the characteristic curve of the material. Ideally, the luminance ratio itself should be determined, but this requires experience in locating suitable zones and is by no means easy to measure accurately. This method is usually too complicated for everyday use. Simpler methods are preferred in practice, even though they may, in theory, be less accurate.
Measurement of total scene luminance
This ‘integration’ method accepts all the light flux from the chosen area of the subject to be photographed. Such a reflected light measurement is simple to carry out and requires little selective judgement. It is used in hand-held meters and in flash meters, but is now obsolete in in-camera metering systems, having been replaced by centre-weighted average metering. As described earlier, such a method will be accurate only for scenes with balanced highlights, mid-tones and shadows.
A correction is usually necessary to the indicated values of exposure settings for subjects of very high or very low contrast, as shown in Table 12.2, or for a subject of non-typical tone distribution, such as a high-key or low-key subject. Most cameras with an automatic exposure mode have an associated exposure compensation control which offers manually set increases or decreases of up to ±3EV or more in increments of 1/2 or 1/3 EV to cope with atypical subjects or luminance distributions.
Measuring a middle grey surface (key tone)
As mentioned previously, the exposure meter is calibrated to give a combination of aperture and shutter speed settings for a given ISO speed so that the surface measured will be reproduced in the print as middle grey. A light measurement technique, suitable for scenes with uniform lighting, is to select and measure a suitable area of the subject with middle luminance (key tone). A Caucasian flesh tone (about 25% reflectance) may be used as a key tone. This technique is similar to the incident light measurement. It is important to note, however, that it does not provide any information on the tonal range of the scene luminances.
In cases where the highlights and the shadows are out of the dynamic range of the image sensor or film, control of the scene illumination may be necessary to ensure that both highlights and shadows are suitably located on the characteristic curve (film) or the transfer function (image sensor). The middle luminance area of the subject can be substituted by a middle grey card (18% reflectance) such as the Kodak 18% reflectance neutral card, which is calibrated to the same middle grey value as the exposure meter. This card is used to conduct light measurements in scenes with variable luminances. The card is placed within the scene under the same incident light, so that it receives the same illumination as the subject. A measurement is taken with the exposure meter close enough so that the card fills the acceptance angle of the photocell. This ensures that the reading is not influenced by any variations in the scene luminances.
Table 12.2 Luminance range and exposure correction
Many cameras offer an ‘exposure memory lock’ feature to allow the metering system to measure and set an exposure for a key or substitute tone, after which the picture is recomposed and the stored exposure given. This memory may be self-cancelling after exposure, or may require manual cancellation.
Measurement of luminance of the darkest shadow
This method locates this tone at a fixed point on the toe of the characteristic curve of the film, giving the correct density to the deepest shadow that is to be recorded. With digital photography, however, this technique may result in clipped highlights, so in most cases it is not used. Practical difficulties include the necessary metering sensitivity, the effects of flare light on measurements and the effects of highlights in the field of view of the meter.
Measurement of luminance of the lightest highlight
This method is similar to the artificial highlight method, which works well for colour slide material. With the arti-ficial highlight method, an exposure measurement is taken from a matt white card, with approximately 90% reflectance. The indicated exposure on the exposure meter is then divided by 5 (90%/18%) to correctly reproduce the scene’s tonal range. The method of measurement of luminance of the lightest highlight locates the highlight at a fixed point of the characteristic curve with the highlight detail and other tones as differing densities. The exposure on the exposure meter is then divided by 5. Shadow detail is usually adequately recorded, but a subject of average contrast may receive more exposure than necessary. Note that when using this method, measurements must not be made on specular highlights.
Averaging values
This method determines exposure by measuring the lightest and darkest areas in the scene that need to be recorded. The output combination of lens aperture and shutter speed for both measurements is then averaged. For this method a spot exposure meter is more suitable than a wide-angle exposure meter, which requires a wide area of approximately uniform luminance. It should be noted that an area that appears to have uniform luminance may appear to have variation in luminances when it is observed or measured at a closer distance. Although this method gives information on the brightest and darkest areas in the scene and the difference in stops between them, it does not give information on how the intermediate tonal values are distributed.
EXPOSURE TECHNIQUES AND DIGITAL CAMERAS
As described in the previous section, some of the traditional exposure measurement techniques are equally applicable to digital capture; however, the characteristics of image sensors and the limitations imposed as a result of the process of analogue-to-digital (A/D) conversion mean that certain aspects require careful consideration and emphasis when establishing correct exposure. A fundamental concern is the combined dynamic range limitation of the sensor and subsequent A/D conversion. Additionally, clipping of highlights and the characteristics of digital noise in shadow regions of the image have a significantly detrimental effect on image quality. Further, if images are captured to the JPEG file format, then the dynamic range limitations are more pronounced, as the images are gamma corrected and quantized to 8 bits. Fortunately, image processing and the option to capture RAW images, along with the ability to view the image histogram at capture, provide opportunities to optimize the exposure. In-camera processing and RAW capture are covered in some detail in Chapter 14. Chapter 17 discusses digital file formats, and the characteristics and mechanisms of the JPEG compression algorithm are described in Chapters 23 and 29.
Using the image histogram
Digital capture has a fundamental advantage over capture on film: the ability to immediately review the image. This means that the exposure can be checked and, if incorrect, adjustments can be made for quick re-exposure. As with film, bracketing of exposures increases the chance of obtaining an optimum exposure.
However, one of the issues with reviewing images is that unless the camera is tethered to a computer and images are remotely captured to be displayed on a large and calibrated monitor, then assessment of the exposure must be achieved on a small LCD screen on the back of the camera. Although the size, resolution and colour reproduction of camera LCD displays continue to improve, image quality, lack of calibration and widely variable viewing conditions mean that judgement of optimum exposure is unreliable from the screen image alone.
Many digital cameras (almost all SLRs and many compacts) offer the option to display the image histogram alongside the image and this provides a much more accurate assessment of exposure. Clipping of either highlights or shadows is immediately apparent (see Figure 12.3), and the position and spread of values in the histogram provide an assessment of overall exposure and contrast. A distribution of tones across most of the histogram indicates an image of high contrast. Over- or underexposure, as described earlier, will shift all values in the histogram to the right or left respectively and will reduce the spread of values, signifying decreased image contrast. Other options include the ability to display clipped areas of the image by overlaying the pixels with a saturated colour, which can help to determine how visually problematic the clipping will be and the display of separate red, green and blue channel histograms.
‘Exposing to the right’
A fairly recent concept, the idea of exposing to the right is a technique used with RAW capture, to better distribute the levels available after quantization, while aiming to avoid clipped highlights and to reduce noise and contouring in the shadow areas of the image.
As described earlier, the image sensor has a linear response, meaning that there is a directly proportional relationship between the numbers of photons recorded at a photosite and the value produced. Tone mapping or gamma correction must be applied when the data is converted to an output-referred state for viewing. As described in Chapter 14, in most digital SLR cameras, the sensor data is commonly quantized to 10, 12 or 14 bits. Assuming 12-bit quantization, this equates to 4096 (212) discrete tonal levels. However, the usable dynamic range of the camera is commonly between five and nine stops of exposure, due to limitations from noise, signal amplification and various non-linear tonal mappings.
Because each change in exposure of one stop represents a log (base 2) change in recorded luminance, and the sensor is recording luminance increments linearly, this means that half of the available levels (2048) will be recorded for the brightest zone of the image (the highest exposure stop). For each stop down the scale of the recorded luminance range, the number of levels will be halved (1024 levels will be allocated to the next stop, 512 to the one after that, etc.). The shadow areas of the image will be allocated very few levels, which can result in visible contouring in these regions.
The term expose to the right refers to the selection of exposure settings such that the image histogram is skewed to the right (as a result of slight overexposure) but without blowing out the highlights; the scene is effectively exposed for the highlights. This shifts each image zone up the scale in terms of the number of levels each one is allocated (albeit compressing the range allocated to the brightest zones) and generally represents a more efficient distribution of the levels available.
The exposure is then adjusted during RAW conversion, to move the mid-tones and shadows to the left of the histogram. Therefore, these image regions are initially allocated more levels than they would be at the normal exposure, reducing the possibility of contouring and better utilizing the range of levels of available. Additionally, the over-exposure increases the signal-to-noise ratio in the darker areas of the image (see Chapter 24). The amount of noise in the shadows is therefore lower using this method than it would be at the ‘correct’ exposure.
Care must be taken at exposure, remembering that the meter is usually calibrated to a mid grey; therefore, achieving the correct amount of overexposure will be a process of trial and error with careful assessment of the histogram. Further, the success of the method is very much dependent on the subject luminance range. A very-high-contrast scene may not allow much ‘pushing’ to the right of the histogram, or subsequent ‘pulling out’ of the shadows.
THE ZONE SYSTEM
The zone system was devised by the American landscape photographer Ansel Adams for black-and-white photography, although its principles are also applied for the exposure of colour films and digital sensors. This system allows the photographer to control the results accurately, according to personal working conditions and the needs of subject and interpretation. Tones can be expanded and contracted to result in specific controlled contrast.
In the zone system the tonal range of the image is controlled with exposure and film development, rather than by altering the contrast in the printing process. It was created for black-and-white photography using large-format sheet-film camera, where the negatives could be processed individually.
When using the zone system, it is essential to observe the subject and ‘pre-visualize’ how the tone values of colour subjects might be recorded on the final black-and-white print. The whole process, film exposure and development, is then implemented to achieve these results. The desired result depends on the imaging application and context. For example, it may be essential to reproduce a close representation of the tonal range of the subject, or to control the contrast to create a specific mood.
The zones
In the zone system, the series of pre-visualized subject brightnesses extend from the deepest shadows to the brightest highlights, and are then related to a set of zones which correspond to luminances on the print. These zones start from zone 0 through to zone X. Zone 0 represents total black (maximum black reproduced from the paper) and zone X represents total white (white of the paper). In these extreme zones of the scale, 0 and X, no detail is recorded. Detail appears from zone II and up to zone IIX. Middle grey, the value equivalent to 18% reflectance card, is represented by zone V. A change of one zone corresponds to a change of one stop. For example, a part of the scene with illuminance half that of the illuminance corresponding to zone V would fall into zone IV. If the illuminance is twice as high it will fall in zone VI. When the exposure is changed by one stop all the subject illuminances are reallocated one zone up or down the scale. Pre-visualization of the final print allows the photographer to decide which areas of the subject are desired to be exposed as zone IV, zone V, zone VI, etc. The photographer, for example, may want a specific area of the subject to correspond to zone IV. After measuring the area with the exposure meter, the subject would then be underexposed by one stop.
The stop difference between the shadows and the highlights should also be taken into account. The subject luminance range covered by the 10 zones from I to IX is over 1:500. This, however, occurs in cases such as backlit subjects with a high-intensity light source which produces hard lighting. A typical luminance range is around 1:128 (seven stops difference) which covers a range of eight zones of the zone system. In addition, to record detail in both shadows and highlights the subject tonal range should lie between zones II and VIII. With black-and-white films, the tone is controlled with development so when using the zone system these films are exposed for the shadows and developed for the highlights. Use of the zone system with colour films is more limited, as the processing time cannot be altered to control contrast. In this case, only exposure is used to control the zones that will represent each tone on the print. Colour negative films are exposed for the shadows while colour transparencies are exposed for the highlights.
The zone system and digital cameras
As mentioned before, there are limitations in using the zone system with digital cameras. One of the reasons is the lower dynamic range of the digital image sensor compared to the dynamic range of film. In addition, overexposed highlights lose detail due to clipping. In the zone system this would correspond to highlights exceeding zone VIII. When clipping takes place, detail cannot be recovered and is permanently lost. To retain detail in the highlights, measurements are taken from the highlights and the exposure is set so that they fall into zone VII or VIII.
HIGH-DYNAMIC-RANGE (HDR) IMAGING
As mentioned earlier in the chapter, the subject luminance range often exceeds the dynamic range of the image sensors. In such cases a single capture results in clipping of either highlight or shadow regions in the scene. Solutions for HDR image capture have been introduced and have become very popular in the last decade. These include: (i) multiple exposures of the same scene using a digital camera with a low dynamic range capability and (ii) sensors with extended (higher than usual) dynamic range capture capability. The latter will not be discussed in this chapter.
Capture of multiple bracketed exposures of the same scene involves a number of different exposures of the subject, for appropriate capture of the shadows, mid-tones and highlights. Multiple exposures are then combined into a single image by a number of steps:
1. Derivation of the opto-electronic conversion function (OECF) is used to estimate how the camera sensor reacts to changes in exposure (see Chapter 21 for measuring OECFs of digital cameras).
2. HDR image construction. The HDR image is comprised of floating point values (stored in 32 bits per channel), which have higher tonal resolution than the usual low-dynamic-range data. The principle in this step is that each pixel in each separate image provides an estimate of the radiance value of the specific point in the scene. Different estimates are assembled by means of a weighted average and taking into account the reliability of the pixel itself. For example, very low pixel values coming from low exposures are often noisy and will be weighted less, whilst the same pixels are well exposed in images acquired with higher exposure settings and will be weighted more.
3. Tone mapping to display or print media (see Chapter 21 for more details). Output devices have a dynamic range which is lower than that of the HDR image, with the exception of the recent HDR displays. As a result, the image appears dark when displayed on typical low-dynamic-range displays. Reduction of the bit depth of the final image (to 8 or 16 bits per colour channel) is therefore carried out to map the dynamic range of the image to that of common output devices. Rendering to a lower bit depth is achieved by tonal mapping. Kuang et al. (2007) have given an overview of different HDR image-rendering algorithms. These include: a sigmoidal contrast enhancement function; histogram adjustment using local luminance adaptation levels; Retinex (an HVS model that simulates the eye local adaptation); a Retinex-based adaptive filter; iCAM (image appearance model which combines colour appearance with spatial vision metrics); modified iCAM (includes the Reinhard and Devlin tone mapping operator); a bilateral filtering technique (the image is decomposed into a base layer and a detail layer and the base layer’s contrast is compressed); and local eye adaptation (compression of the luminance channel dynamic range by imitating the retina’s photoreceptor responses). An HDR tonal mapping applied directly on the colour filter array (CFA) of the sensor instead of the image after demosaicing has also been proposed by Alleysson et al. (2006).
Figure 12.9 An HDR image created with three exposures.
Multiple-frame capture is achieved by varying the shutter speed rather than the aperture to ensure that depth of field remains constant in all exposures. With this method, it is important that the subject is static and that camera shake is prevented to avoid blurring of the subject; thus the use of a tripod is essential. Imaging software used for HDR image construction and tonal mapping to an output often provide features to minimize possible blurring due to camera shake by aligning the different exposures. Software used for HDR imaging provides options for implementation of tonal mapping using various different algorithms.
BIBLIOGRAPHY
Adams, A., 1981. The Negative. Little, Brown and Co., Boston, MA, USA.
Alleysson, D., Meylan, L., Süsstrunk, S., 2006. HDR CFA image rendering. Proc. EURASIP 14th European Signal Processing Conference, Florence, Italy, 4–8 September.
Hertel, D., 2009. Extended use of incremental signal-to-noise ratio as reliability criterion for multiple-slope wide dynamic range image capture. Journal of Electronic Imaging, Volume 19, Issue 1.
Jacobson, R., Ray, S., Attridge, G.G., Axford, N.R., 2000. The Manual of Photography. Focal Press, Oxford, UK.
Johnson, G.M., Fairchild, M.D., 2003. Rendering HDR images. Proc. IS&T/SID 11th Color Imaging Conference, Scottsdale, pp. 36–41.
Kuang, J., Yamaguchi, H., Liu, C., Johnson, G.M., Fairchild, M.D., 2007. Evaluating HDR rendering algorithms. ACM Transactions on Applied Perception, 4, Article 9.
Lukac, R., 2009. Single Sensor Imaging: Methods and Applications for Digital Cameras. CRC Press, Taylor & Francis Group, Boca Raton, FL, USA.
Meadhra, M., Lowrie, C.K., 2006. Exposure and Lighting for Digital Photographers Only. John Wiley, Indianapolis, Indiana, USA.
Peres, M.R., 2007. The Focal Encyclopedia of Photography, fourth ed. Focal Press, Oxford, UK.
Rand, G., 2008. Film and Digital Techniques for Zone System Photography. Amherst Media.
Reinhard, E., Ward, G., Pattanaik, S., Debevec, P., 2005. High Dynamic Range Imaging: Acquisition, Display, and Image-Based Lighting. Morgan Kaufmann, San Francisco, CA, USA.
Salvaggio, N.L., 2008. Basic Photographic Materials and Processes, third ed. Focal Press, Oxford, UK.
Saxby, G., 2001. The Science of Imaging: An Introduction. Taylor & Francis, Institute of Physics Publishing Bristol and Philadelphia.
Schaefer, J.P., 1999. The Ansel Adams Guide, Book 1: Basic Techniques of Photography. Little, Brown and Co., Boston, MA, USA.
Stone, H., Sidel, J.L., 2004. Sensory Evaluation Practices. Academic Press, Oxford, UK.
Stroebel, L.D., Current, I., Compton, J., Zakia, R.D., 2000. Basic Photographic Materials and Processes. Focal Press, Oxford, UK.