Exposure in Digital Photography – Basics
This article reviews several basic aspects of exposure in digital photography for helping beginners and mid-level enthusiasts to get better pictures. More details on this topic are available on Internet and also in very large number of books.
Exposure is the total light energy incident either on film or on image sensor. Physically, this is the luminous flux collected by the lens during the exposure time or shutter speed. You can control the exposure either by changing the luminous flux, or by changing the shutter speed, or by changing both the luminous flux and the shutter speed. You should be aware that from the exposure standpoint there are significant differences between the film photography and digital photography.
In film photography, the exposure changes the film transparency. In digital photography, the exposure creates a voltage across each photo element of the image sensor, proportional to its exposure.
Digital photography should be back compatible with film photography in several aspects. One aspect is film speed, which is a measure of film sensitivity to light, expressed in ISO_number or ISO_speed, such as ISO100, ISO200. Light sensitivity of all digital cameras is also expressed and set in ISO_speed, the same as film sensitivity. The smallest ISO_number has the lowest light sensitivity, but also reveals the finest details in the picture and has the lowest noise level. When increasing ISO_number, light sensitivity increases, but also increases the noise level and the picture has less details.
We explain briefly film sensitivity related to Film Exposure diagram showing the optical density or transparency of the photographic film versus the incident light energy. The optical density graph of the film is nonlinear, as any transmission function. The bottom level of optical density graph is the noise region known as Density of base + fog, inherent in any photographic film and also in any image sensor.
The total amount of light energy incident either on film or on image sensor depends on the light collected by the lens across its entrance pupil expressed by the f/number, and on the shutter speed or exposure time. f/number and shutter speed concepts have their origins in film photography, but they migrated naturally to digital photography with the same meaning. However, exposure sensitivity, f/number and shutter speed are related to light control in photography, including digital photography.
Always the lens specs give the minimum f/number as a mean to specify maximum entrance pupil. A lower number such as f/1.2 means a faster lens, collecting twice more light than f/1.4 lens. In the schematic Price Difference Between f/1.2 and f/1.4 Lenses below, the lens with f=50mm, f/1.2 has 41.67mm entrance pupil diameter, but the lens with f=50mm, f/1.4 has 35.71mm entrance pupil diameter. Notice the price difference between these two Canon lenses, only for 0.2 difference in f/number. High price tag of f/1.2 lens comes from special optical elements used for compensating the geometric and chromatic distortions towards the lens boundary.
For the same amount of light coming from the scene, film illumination is controlled by the combination [f/number; shutter_speed] usually known as Exposure Value or EV. In other words, a certain EV can be obtained by any [f/number; shutter_speed] combination which gives the same amount of light energy on film or on image sensor. The schematic Constant Exposure Value – EV below shows several combinations [f/number; shutter_speed] giving the same EV and two pictures shot with this EV. Notice that EV-1 means two times more light than EV, and EV+1 means half the light than EV. The difference in two consecutive EV is known as stop. There is no conflict between stop used as difference between two successive EV values, or between two successive aperture values, or between two successive shutter speed values. Any unit change either in aperture or in shutter speed makes a unit change in EV.
Shooting a picture is always a mix between the camera capability and photographer’s skills, with emphasis on photographer’s side. Light metering is very important on shooting, but the photographer decides either to use EV as suggested by camera, or to do Exposure Compensation – EC. Most of digital SLR cameras such as Canon EOS Rebel T3i 18MP CMOS APS-C Sensor DIGIC 4 Image Processor and Nikon D7000 16.2MP DX-Format CMOS Digital SLR to name just two very popular models have EC feature. EC is done totally according to photographer feeling for altering the camera decision on f/number and shutter speed based on factory provided algorithms. Several hints about EC are provided later in this article. When EC increases with one stop, EV decreases with one stop, allowing more light to illuminate the image sensor. All digital SLR cameras have EC features. The popular models Canon EOS Rebel T3i 18MP CMOS APS-C Sensor and Nikon D7000 16.2MP DX-Format allow EC span from -5EV to +5EV in 1/3 EV or 1/2 EV steps.
The lens aperture is the circular opening into lens diaphragm located in lens optical center which controls the exposure by controlling the luminous flux. Having an adjustable size like human eye iris, lens diaphragm is sometimes called iris diaphragm, or simply iris. The f/number usually known as stop, is part of a sequence such as f/1.2, f/1.4, f/2, f/2.8, f/4, f/5.6, f/8, f/11, f/16, f/22, f/32 where denominator values approximate powers of square root of two. Each stop corresponds to an aperture size. When changing from one stop such as f/2.8 upwards to f/4, the aperture area halves, which reduces to half the luminous flux through the lens. The math is simple: if the aperture follows the rule of powers of two, its diameter follows the rule of square root of two. When the stop changes from f/2.8 downwards to f/2, the aperture area doubles and two times more light goes through the lens.
Besides controlling the light flux through the lens, the aperture is also an optical element of the lens, controlling its depth of field – DOF. In the schematic Constant Exposure Value – EV, both pictures aim the same subject, illumination conditions and EV are identical.
The left-hand side picture was shot with slow shutter such as 1/15 and with small aperture such as f/16. The scene has large DOF, with details in the background and with water drops looking like a bunch of thin strings. You should be careful with too small-size apertures beyond f/8. The picture might loose sharpness or details versus a picture shot with high aperture such as f/1.8, because of diffraction.
The right-hand side picture was shot with high speed shutter such as 1/1000, but with large aperture such as f/2. You see a shallow DOF with well focused “frozen” water drops in the front, but with a blurred background.
The light going through the lens reaches the image sensor during the time frame of opened shutter, or shutter speed. Typical numbers for shutter speed in seconds, are: 1, 1/2, 1/4, 1/8, 1/15, 1/30, 1/60, 1/125, 1/250, 1/500, 1/1000. For high-end digital SLR cameras such as Canon EOS Rebel T3i 18 MP and Nikon D5100 16.2MP CMOS Digital SLR, shutter speed sequence spans from 30s to 1/4000.
The schematic Shutter Speed and f/number Examples has pictures of the same scene, with the same illumination, but shot with different exposure conditions, each picture having its EV. Across each row, f/number is constant, but the shutter speed goes from fast to slow as specified in schematic. Obviously, within the same row, the pictures are lighter toward slower shutter. Across each column, the shutter speed is constant, but f/number increases upwards. Accordingly, the pictures are going from lighter on the bottom side to darker toward the top side. In conclusion, we notice two extreme situations: underexposure on top left side of the schematic and overexposure on bottom right side. In-between these extreme cases, there are more or less acceptable exposures of the scene.
The regions of underexposure and overexposure of photographic film are marked with their respective names on Film Exposure graph.
The schematic Image Sensor Response to Illumination shows the structure of Foveon X3 CMOS color image sensor and the graph of voltage generated across a single image sensor versus the luminous flux incident on it, expressed in number of photons. The image sensor has several major differences versus the photographic paper:
(IS1) The image sensor consists of large number of discrete individual sensors or pixels. Across its surface, each pixel samples part of the image incident on it.
(IS2) The incident luminous flux on each sensor generates a voltage proportional with the illuminance, following almost linear dependence.
The human vision is able to see from very low light level such as starlight when there is no color perception, up to bright sunlight. This is a very high dynamic range of 10,000,000,000 in power, or over 100dB in ratio, or about 33.2 stops.
Some high-end negative films such as Kodak Vision3 have a dynamic range of 11,500 or 13 stops.
The dynamic range of LCD displays widely used as rear digital camera displays, computer monitors and TV sets, commercially referred to as contrast ratio is about 1000:1, or 9.5 stops.
We recommend several advanced digital cameras such as Canon EOS Rebel T3i 18MP CMOS Digital SLR Camera and Nikon D5100 16.2MP CMOS Digital SLR Camera span light sensitivity across 14-bit range, or 16,384:1, or 14 stops. However, tonal gradation of 14 stops is satisfactory for vast majority of situations.
The schematic Image Sensor Response to Illumination shows also the dynamic range of human eye of 33.2 stops, which is larger than the dynamic range of photographic paper, of the image sensor and also much larger than the dynamic range of LCD displays. The purpose of photography is to catch on picture the reality as faithful as possible. The biggest problem in photography is to compress or squeeze the dynamic range of eye perception in the dynamic range of the image sensor and also in the dynamic range of picture displayed on LCD screen and printed. This problem appeared first in film photography where the compressing technique is known as tone mapping. The Film Exposure schematic shows the optical density zones used for matching the picture on photographic paper with the real world seen by human eye. The interested reader can follow the links to find more about tone mapping and zones system.
The pictures above span the entire tonal range of respective cameras as marked on each schematic. However, tonal range of the real life is larger than shown by pictures: we cannot see details in bright areas and also in very dark areas. The schematic below shows two examples of digital pictures containing dark areas and also bright areas. Most digital cameras today show at request the picture histogram. The histogram shows on horizontal axis tone levels or simply tones and on vertical axis shows the number of pixels for each tone value. Maximum number of tones depends on camera model. Digital cameras such as Canon EOS digital SLR series and Nikon D5100 16.2MP CMOS Digital SLR Camera use 14-bit or 16,384 levels analog to digital conversion of signals from image sensor. Accordingly, the maximum tonal value on histogram follows the analog to digital conversion accuracy of the camera. The bottom picture was underexposed on purpose. In its histogram you see more pixels in dark areas and practically no pixels in bright areas.
Some high-end digital SLR cameras such as Nikon D5000 12.3MP DX Digital SLR Camera
provide at request either tonal histogram as in schematic below:
or color histograms across a select area of the image, as in schematic below:
What is a histogram good for? Usually, people ignore the image histogram, unless there are situations when should make a decision on exposure, and how to change it. The schematic Exposure Cases below has three pictures and their histograms, in three situations: normal exposure, when the histogram maximum is roughly at mid tonal range, underexposure when the histogram maximum is shifted toward dark zone, and overexposure when the histogram maximum shifts toward light zone.
The picture expected by photographer depends on light metering and on the exposure modes.
Normal exposure is computed by digital camera based on light metering selection of the photographer. The real life scenes could be very different from one shot to another. The illuminance can be from very faint up to very bright across the entire image projected on image sensor, spanning more than the entire tonal range supported by the image sensor. Finding the optimum exposure across the entire image is a real challenge even for most sophisticated systems. The schematic Light Metering Zones shows light measuring points grouped in areas of light measuring interest and symbols of their use when selecting light metering modes. As we mentioned before, the right exposure is always a combination between the photographer’s skills and the camera capability. According to the scene content and to the subject, the photographer selects one of light metering options as shown in Light Metering Selection schematic.
In Evaluative Metering mode, the camera measures the light with a large number of dots located within a grid across the entire image. In some high-end digital SLR cameras such as Canon EOS-1D X 18.1MP Full Frame CMOS Digital SLR, the matrix measuring sensor can have up to 252 distinct elements for general metering, with 35 elements used for low-light metering. With illumination information collected from all the elements spread across the image and based also on ISO_speed, the camera computes f/number and shutter speed according to the exposure mode selected by photographer. Evaluative metering is recommended for most shooting situations.
In Partial Metering mode, the camera measures the light within the circular region around the center. This mode is recommended for scenes containing bright and also dark areas. Based on the light measurement across aimed area and on selected ISO_speed, the camera sets the f/number and shutter speed.
Center-Weighted Metering mode is recommended for subjects positioned in center of the image.
In Spot Metering mode, the camera measures the light with a specific sensor selected by the photographer from the entire sensor matrix. The selected sensor is highlighted by camera in the spot metering pattern visible in the view finder. For better visibility in Light Metering Selections schematic, the sensor is surrounded by a red circle. In this case, the photographer is particularly interested in the exposure within the area centered on the selected light measuring point.
After selection of light metering modes, the photographer selects one of the exposure modes common to most of digital cameras, such as:
Auto and Scene modes, when the camera selects automatically the ISO_speed, f/number, shutter speed and focus on the closest part of the scene.
P – Programmed Auto, when the camera adjusts automatically f/number and shutter speed for optimal exposure. The photographer has the freedom of manual selection of the combination f/number and shutter speed which keeps the same EV.
S or Tv – shutter mode, when the photographer selects the shutter speed and the camera finds the f/number for optimum exposure.
A – aperture mode, when the photographer selects the f/number and the camera finds the shutter speed for optimum exposure.
M – manual mode, when the photographer selects both the f/number and the shutter speed according to personal preferences.
Normal Exposure uses factory defined algorithms for selecting f/number and shutter speed. In many situations, normal exposure gives correct exposures. There are still situations when normal exposure is not the right answer for f/number and shutter speed combination. The large diversity of scenes and of subjects within the scenes require also photographer’s correction of Normal Exposure outcome.
In white dominated scenes, Normal Exposure goes toward darkening the picture, as seen in Exposure Compensation Examples schematic. The photographer must add some EC steps to shoot a good picture. There are available either 0.3EV or 0.5EV steps for EC. Referring to Image Sensor Response to Illumination schematic, with Exposure Compensation the photographer shifts EV span upwards, starting from above the DARK level and going beyond the BRIGHT level. Now the details hidden initially below DARK level become visible, but of course the details above BRIGHT level become invisible. The photographer can see EC outcome either looking at the picture on rear camera LCD, or looking at the pictures histograms before and after EC. If not satisfied, another EC is required. The picture is brighter when increasing EC and darker when decreasing EC.
In dark dominated scenes, Normal Exposure goes toward lightening the picture. The photographer must subtract some EC steps for having a nice picture, as seen in Exposure Compensation Examples schematic. Referring to Image Sensor Response to Illumination schematic, using EC the photographer shifts downwards EV span, starting from below the BRIGHT level and going beyond the DARK level. After EC, the details hidden above BRIGHT level become visible, but of course the details below DARK level become invisible.
WARNING: Picture quality must be improved before shooting, for revealing the details in the picture within extreme exposure areas, either dark or bright. There is no image editor able to reveal details from a badly exposed picture. That picture does not contain the expected details!
The EC situations of images explained above should work fine until the subject losses its patience! Exposure Bracketing is required for shooting a scene with areas illuminated beyond the tonal range of the image sensor. Referring to Image Sensor Response to Illumination schematic, the Normal Exposure with EC=0 covers a window of only 14 stops, which is the tonal span of the image sensor. For catching tones beyond full tonal span of Normal Exposure upwards and also downwards, you must shoot the same scene with different EV, usually with one or two stops higher and lower. In this way, you get the pictures with Normal Exposure and also underexposed and overexposed pictures of the same scene, as you can see in Exposure Bracketing schematic. Using one of the image editors available in the Photo Editing Software section of our Image Lab department, you can build stunning High Dynamic Range – HDR pictures.
Exposure in digital photography is very extensive subject. This article highlighted only the main aspects such as light sensitivity, light metering and exposure compensation. There are also other aspects, such as selection of aperture, shutter speed, HDR photography, to name just a few which will be covered in future articles. We invite the readers to express their opinions about this article and also to suggest us other subjects for future articles.