Underwater photography exposure control
Version September 22, 2023
Bart Hazes
© 2023. This work is licensed under a CC BY NC SA 4.0 license.
Index
Aiming for true subject luminosity
Adjusting the assumed gray-scale
Aiming for maximum dynamic range
Underwater ambient color spectrum
Ambient + strobe white balance
‘Strobe-dominated’ versus ‘balanced’ strobe:ambient ratio
Strobe:Ambient ratio and water background color
Strobe:Ambient ratio and distance
Controlling ambient light exposure
Maximum exposure time - motion blur
Minimum exposure time - strobe synchronization speed
What exposure time should you use?
Aperture and optical resolution (diffraction limit)
Aperture and depth of field (DOF)
Aperture and lens imperfections
Aperture and strobe:ambient ratio
What aperture setting should you use?
Does increasing ISO increase noise?
What ISO setting should you use?
Manual strobe with manual camera control
Manual strobe with automated camera control
TTL strobe with manual camera control
TTL strobe with automated camera control
TTL strobe with exposure-time restricted aperture priority mode
Numerical exposure visualization
Before writing this guide I had not fully appreciated what is involved in obtaining a correctly exposed image, especially when using a strobe. This is in part a testament to the amazing technology in cameras and strobes that often give surprisingly good results using automated exposure settings. So why write a guide if automatic modes work? There are several reasons:
Camera exposure concepts are very well covered in existing websites and books. The website https://www.cambridgeincolour.com/ has very well-written tutorials for land-based photography that I found very helpful and would recommend. Here I will discuss exposure in the context of underwater photography based on 15 years of experience and studying the subject.
We see the world around us based on light reflecting off of subjects. The reflected light intensity (luminosity) depends on both the properties of the subject (reflectance) as well as the intensity of the incident light. Only if you know the intensity of the incident light can you establish the true luminosity of the subject. But cameras only measure reflected light and have no knowledge of the incident light. So for a given amount of detected light the camera does not know if this is a dark subject in bright conditions or a bright subject in dark conditions.
As an imperfect solution, the camera has to make an assumption about the luminosity of the scene in your frame. A typical assumption is that it reflects 18% of light, ‘18% gray-scale’ (aka ‘middle gray’), but camera manufacturers can pick different values in that ballpark area. This means that if the true luminosity of the scene is indeed equivalent to 18% gray-scale, then the recorded image reflects the true subject luminosity.
Real-world scenes often have a wide range of luminosity values across the frame. Apparently, when you take the average luminosity in such cases that value often ends up being close to 18% gray-scale, explaining why it is used. But obviously there are scenes with notably lower (low-key images) or higher (high-key images) luminosity than 18% gray-scale. For low-key images the camera measures less light than expected and will assume the incident light level is low. In response it adjusts aperture, shutter speed, ISO, and or strobe power to boost exposure. This results in brightening images that should be dark. The opposite happens in high-key images where the camera assumes incident light intensity is high leading to undesirable darkening of the scene.
The typical method to deal with low- and high-key scenes is to recognize them as such and dial in exposure compensation. You use positive compensation for high-key scenes to keep them bright and negative compensation for low-key scenes to keep them dark. Trial and error, plus experience, is needed to develop a sense of how much compensation is needed.
A different method is to change the camera's assumption about the subject luminosity. In spot metering mode, my camera can use regular spot metering (assuming 18% gray-scale) as well as ‘spot-low’ or ‘spot-high’ modes. The latter two tell the camera that the spot meters a subject that is dark (spot-low) or bright (spot-high). Spot-low and spot-high assume gray percentages of about 2% and 72%, respectively, on my Olympus OM-D E-M1mkii. See the Exposure Metering section for more on this topic.
A different approach is to just see the sensor as a light capturing device where each pixel can measure light intensity between zero and some maximum value. Weak signals have poor signal-to-noise ratio so you want to maximize exposure. However, all pixels with a signal exceeding the maximum value end up with the same value, thereby losing any image detail in such regions (‘clipped’ or ‘overexposed’ pixels).
Maximizing dynamic range usage means using the strongest exposure that has no, or a tolerable amount of, clipped pixels. Maximizing the ‘dynamic range’ gives the cleanest image without overexposing the bright areas. It will render low-key images too bright but you can adjust that later in software without quality loss. See the intensity histogram section for more information on how to maximize exposure without clipping.
The discussion above looked at overall light intensity and explained the need to assume ‘18% gray-scale’ because the camera does not know the intensity of the incident light. For color photography this reasoning has to be extended to the red, green, and blue pixels on the sensor. In this context, the camera needs to know the color spectrum of the incident light to render the scene appropriately. This is referred to as the white balance.
Cameras have presets for the spectrum of common light sources, including sun light and different types or artificial light. You can select the appropriate preset or trust that ‘auto white balance’ will make a suitable choice.
A more general solution is based on the fact that the spectrum of many light sources behaves like the radiation coming from a hot ‘black-body’, with the actual spectrum depending on the temperature of the ‘black-body’. The higher the temperature the more blue and less red there will be in the spectrum. As a user you can select an appropriate temperature, expressed in Kelvin. Note that to boost red you need to tell the camera that the incident light has less red, so a higher Kelvin temperature.
Underwater photography has to deal with a non-standard ambient light spectrum because red light is rapidly lost from the spectrum as you dive deeper. This situation cannot be represented by any of the white balance presets or a Kelvin setting. If not dealt with properly, you end up with an unattractive blue cast on your images. One solution is to use an underwater strobe (or strong video light). If exposure is dominated by the artificial light source then you can use the strobe-light white balance preset (or auto white balance). This will be discussed further in the Ambient + strobe white balance section below.
One way to let the camera determine the actual ambient light spectrum is to take an image of a scene known to have equal amounts of red, green and blue (a gray card or a white dive slate). Using the ‘one-shot white balance’ function you tell the camera to image the known reference and calculate how to adjust the red, green and blue light intensities to make them all equal.
It is important to understand that the white balance does not change how the sensor measures light. It simply defines how to multiply the red, green and blue pixel intensities to obtain natural looking colors. The problem for underwater photography is that the red signal gets so weak that you need a very strong multiplier, which unfortunately also multiplies the noise. The better solution is to boost the red signal with a longer exposure, but that would overexpose the blue and green. A workaround is to use a red filter that attenuates blue and green, but not red, light. If you attenuate blue and green by 2 stops, you can increase the exposure by two stops. This gives a 4-fold boost in red without overexposing blue and green.
The white balance of a strobe-lit scene depends on the strobe:ambient ratio and on the color spectrum of each light source. On land the ratio has not much impact on white balance because typically the strobe and ambient light have similar spectra. In addition, exposure is often dominated by either the strobe (in dark conditions) or ambient light (when using a strobe as ‘fill flash’, see the Fill Flash section).
This all changes underwater because now the spectra are very different, with red virtually missing from ambient light. This has two consequences: 1) the white balance becomes sensitive to the strobe:ambient ratio and 2) to restore red colors, you want the fill flash to use a higher strobe:ambient ratio than the land-optimized default settings.
The easiest way to use strobes underwater is to significantly underexpose the ambient light and let the strobes provide the bulk of the light. With TTL strobes the camera and strobe can do this automatically. With strobe-dominated exposure you are guaranteed to get strong reds and a stable white balance corresponding to that of the strobe.
With a balanced exposure, both the strobe and ambient light contribute significantly to the exposure and this renders colors closer to the way they are experienced during a dive. This is particularly relevant for wide angle images where part of the scene is ‘blue water’ as discussed next.
Exposure of ‘open water’ in the background of images needs special consideration because water does not reflect strobe light. Accordingly, its exposure depends just on ambient light. Strobe-dominated exposures, which cut out most ambient light, result in dark blue, or virtually black, water color. This can make your subject ‘pop’ in a desirable manner. But if you want to capture the characteristic blue color of ocean water you need to use a balanced exposure that lets in more ambient light. There is a good description of this by Alex Mustard. With a strobe:ambient ratio of 1 you will be underexposing red by up to 1 stop. This may be a small enough difference to ignore or you can boost red colors later using software, especially if you save your images in RAW format.
The ambient light intensity does not depend significantly on the subject distance. In contrast, the intensity of strobe light decreases by the square of the distance to the subject. That means that the strobe:ambient ratio, and thus the white balance, changes across the scene depending on distance from the strobe. The result is a very colorful foreground grading increasingly towards blue for more distant areas. This can be desirable as it draws the eye to the foreground while the grading to blue gives a sense of distance and the ‘atmosphere of the dive environment’. However, if you want red colors to reach further you need to increase the strobe:ambient ratio.
In a digital camera you expose a sensor to photons derived from the scene of interest. For a given lens, the number of photons that hit a sensor pixel depends on only two things: exposure time (shutter speed) and aperture. Each photon is converted to an electric charge that accumulates in the pixel so that the charge is proportional to the number of captured photons, up to the maximum charge capacity of the pixel.
Exposure time is conceptually the easiest parameter to understand, if you double the exposure time you double the exposure. However, only in the simplest situation, a stationary subject in ambient light, are there no other side effects.
If the subject, or camera, moves significantly within the duration of the exposure, motion blur will occur. Image stabilization reduces the impact of camera shake but there remains a limit to how long an exposure can be without risking motion blur. Of course, the exposure time also needs to be short enough to prevent overexposure and, if using a strobe, underexpose ambient light sufficiently to accommodate the desired amount of strobe light.
Cameras have a minimum exposure time that, without strobe, is so short that it never becomes an issue. However, underwater we typically use a strobe and then the exposure time is limited to the ‘strobe synchronization speed’ (1/250th of a second for my camera). Up to the synchronization speed, one of the two camera shutters (‘first curtain’) opens completely before the other shutter (‘second curtain’) starts to close to end the exposure. The strobe uses this brief period where the entire sensor is exposed to fire a very short but intense light pulse. It is very important to understand that the strobe can deliver its full power even at the maximum synchronization speed. It means that reducing exposure time up to the max. synchronization speed reduces ambient light without affecting strobe light.
If you are aiming for strobe-dominated exposure you can simply set exposure time to the max. synchronization speed for your camera. This minimizes risk of motion blur and maximizes the strobe:ambient ratio, giving bright reds and a stable strobe-defined white balance. For a more natural water background color you increase exposure time as needed to get a nice blue color. You can also use a longer exposure time to artistically induce blurring, for instance to convey the notion of motion.
The aperture reflects the diameter of the diaphragm that lets light into the lens. If you double the diaphragm diameter you quadruple the surface area of the opening (aka entrance pupil) and thus you quadruple the exposure. The reason we don’t talk about the diaphragm diameter is that a lens with double the focal length (f) needs double the diameter (D) to achieve the same photon density on the sensor. The aperture ‘F-number’, defined as f/D, is more convenient because it corrects for the focal length effect and thereby allows apples to apples comparison between lenses of different focal lengths. Apart from these technical details, it is important to know that aperture affects many things in addition to just exposure.
Because the Aperture F-number is defined as f/D it is inversely proportional to the diaphragm diameter. As a result, a bigger diaphragm opening corresponds to a lower F-number. This can be confusing. Another potential source of confusion is that to double the exposure you have to decrease the F-number by 1.4 (√2), for instance going from F=5.6 to F=4.0. For this reason, cameras and lenses label F-numbers so that each one is 1.4 times bigger than the previous one and 2 times bigger than the one before that (e.g. 2.0, 2.8, 4.0, etc.).
The lens diaphragm opening captures a cone of light from each point in your scene. The wider the diaphragm opening, the wider this cone of light. In addition to increasing the exposure level, the angular width of this cone also determines the theoretical optical resolution. Wider angles result in higher resolution. This includes ‘lateral resolution’ which defines the smallest size of a feature that can still be resolved.
The actual recorded resolution is further determined by the resolving power of the sensor. As long as the optical resolution exceeds the sensor resolution you can keep closing the aperture without loss of resolution in the captured image. Once the aperture is so small that the optical resolution becomes the limiting factor, the image becomes ‘diffraction limited’.
The aperture at which diffraction limitation starts depends on the sensor, with low resolution sensors (or low-resolution video) being less prone to the effect. It also depends on the subject distance because getting closer increases the width of the light cone captured with a given aperture. So in macrophotography you can use smaller F-numbers than at greater subject distance.
In addition to ‘lateral resolution’ the optical image also has ‘longitudinal resolution’, which defines the depth of the ‘in-focus range’ (how far before and behind the focus plane objects are still in focus). This is the same as the Depth of Field (DOF). Anything that increases lateral resolution also increases longitudinal resolution and thus decreases DOF. This is important for macro photography where you often have to trade off lateral resolution for DOF.
Lens imperfections deteriorate image quality and the effect is greater farther away from the image center. Deterioration is also greater when you work at higher optical resolution by opening the aperture. The aperture at which this becomes problematic varies from lens to lens and you must decide what you tolerate. When using a dome port for wide-angle lenses you may need to stop down the aperture even more because dome effects can also become an issue.
Camera autofocus systems work better with ‘fast’ lenses (low maximum F-numbers). This increases the amount of light and decreases the DOF. The first makes focus measurements more accurate, even in darker conditions, and the latter means that small deviations from perfect focus can be picked up more easily. Therefore, autofocus always works with the lens aperture wide open, independent of the exposure setting used to shoot the image. This means that wide-aperture ‘fast’ lenses tend to focus better than less expensive ‘slow’ lenses, especially in dim environments.
Changing the apertures affects ambient and strobe light to the same extent and therefore does not by itself change the strobe:ambient ratio. However, if closing the aperture is accompanied by increasing strobe power then you do increase the strobe:ambient ratio, but at the expense of greater strobe battery drain.
With aperture having so many effects there is not one right answer. If you want to play it safe it is better to close down the aperture a bit more than needed. Your strobe will have to work a bit harder and you may not maximize resolution but you are more likely to end up with an in-focus and properly exposed image (assuming TTL is working well and the strobe is powerful enough). This is what typically happens with strobe-dominated exposures. So, rather than overthinking this, you can just pick a ballpark setting around F8.0 and close down further if you get closer to subjects.
If you want to take your photography up a notch and aim for blue water backgrounds, maximize lateral resolution, or make artistic use of limited DOF then it is time to experiment with wider apertures.
In the special case of macro photography you often have to give up some lateral resolution, by closing the aperture beyond the diffraction limit, in order to get adequate DOF. This is not necessarily bad because suboptimal resolution is, well, not optimal. But important parts of the subject out of focus often relegates the image to the digital dustbin. However, once you are in diffraction-limited territory there is no longer a point to add teleconverters or close-up lenses. Instead, your best hope is to shoot your subject from an angle where less DOF is needed or just focus on part of the subject.
Exposure time and aperture control how many photons reach the sensor, but how that translates to the actual signal depends on sensor properties and sensor readout. Sensors that can detect many photons before their pixel charge saturates have a low ISO value and sensors that saturate faster have a high ISO value. Accordingly each sensor has its own ‘native ISO’ based on its underlying technology. But matters are a bit more complicated because the maximum signal that can be detected per pixel can be limited by the pixel itself, sensor ISO, or by the analog-to-digital readout electronics, readout ISO.
Sensor ISO depends on how much charge is accumulated per photon and how much total charge can be accumulated before the pixel ‘saturates’. This is a static feature of the sensor and cannot be changed. One important source of noise in images is ‘photon noise’ and the signal/noise ratio for photon noise improves the more photons are detected. It is therefore always best to optimally use the ‘pixel dynamic range’ by making the brightest parts of your image nearly fully saturate the sensor.
Before a pixel’s voltage is digitized it is amplified. At the ‘base ISO’ (aka ‘native ISO’), the amplification is such that a fully charged pixel corresponds to the maximum digital value (2**14 -1, for a 14-bit sensor). Increasing the ISO setting on a camera increases the voltage amplification. So at double the base ISO, a half-charged pixel will already give the maximum digital value. This has no effect on what happens on the sensor, including photon noise and other sources of noise that occur prior to readout. Its main benefit is that you can underexpose an image and still use the full dynamic range of the voltage digitizer. This slightly reduces readout noise and results in a brighter image, more suitable for in-camera review. However, a drawback is that even pixels that are not saturated on the sensor can become overexposed at higher ISO because they exceed the maximum digital value.
There is a limit to how much the amplifier can boost the voltage prior to digitization. This corresponds to the ‘native ISO range’. The ‘extended ISO range’ offers even higher ISO values but this is simply achieved by multiplying the signal AFTER digitization. This is exactly equivalent to boosting brightness with post-production software. Its main purpose is to allow in-camera review of images at an appropriate brightness level.
Because post-digitizer ISO boost is simply a digital, and error-free, multiplication it has no effect on the image quality. Whether you use this to boost the image brightness in camera or ‘in post’ makes no difference. Which is why post-digitizer amplification gives image quality that is ‘ISO invariant’. There is an interesting youtube video comparing a regular (Canon R5), ‘dual-ISO’ (Black Magic 4K) and ‘ISO invariant’ (RED Komodo) camera, showing the impact on noise profiles.
No. In theory, pre-digitizer amplification reduces the impact of read noise, but this only matters for greatly underexposed (parts of) images. What it does do is make small variations in pixel values more obvious, including other forms of noise that arose prior to readout. So it does not create noise but makes it more noticeable. To reduce noise you need to increase the actual photon exposure via exposure time, aperture, or strobe power.
Changing ISO affects ambient and strobe light to the same extent and therefore does not by itself change the strobe:ambient ratio. However, if increasing ISO is accompanied by decreasing strobe power to avoid overexposure then you do shift strobe:ambient ratio towards more ambient light and you reduce strobe battery drain.
Whenever you increase ISO it means you are not using the full capacity of pixels to detect photons. Weaker signals always result in more noise. So if you can boost ‘true’ exposure by changing exposure time, aperture, or strobe power, that is preferred over increasing ISO. If other considerations prevent you from doing so then increasing ISO is perfectly fine.
Note: For many underwater photographers, wide angle in particular, manual strobe control is the preferred method. However, I use TTL so if you want to use manual strobe control I advise you to learn from people with actual experience. I’ll just briefly go over this topic for completeness sake.
In manual strobe mode, the camera triggers the strobe but light output is set with a dial on the strobe itself. If ambient light is relatively constant, as it often is, and you keep ambient light underexposure at a similar level, then the required strobe output depends mostly on subject distance. Once you figure out the strobe power for your typical subject distance, you just need to increase or decrease strobe power if the subject is farther or closer than typical, respectively. Practice makes perfect.
If you use an on-camera flash to trigger a strobe in manual mode, you should set the camera’s strobe to manual mode with minimum output. This reduces camera battery drain, minimizes recycling time, and will be plenty strong to trigger the off-camera strobe.
Using both the camera and strobe in manual mode gives maximum user control. Manual camera operation is actually not that hard because, although there are many combinations of exposure time, aperture, and ISO that give the same ambient exposure, you often only use a subset of what is possible. For instance, you mostly keep ISO at its base value, set aperture to a ‘default starting value’ and then adjust exposure time to get the desired underexposure level. Because ambient light typically doesn’t change much during a dive you only need small changes in aperture and exposure time once you have found a good starting combination.
In most cases, the aperture is the most interesting parameter to control because it affects so many aspects of the image. If you find yourself changing aperture and then having to adjust shutter time to keep ambient exposure constant you could consider shooting in ‘Aperture priority mode’. In this mode you set the aperture and the camera selects the appropriate exposure time. You may not like the level of ambient exposure the camera thinks is optimal, but you can use exposure compensation to adjust that as needed.
If you don’t want to deal with camera settings at all you could shoot in ‘Program mode’ where the camera selects both the exposure time and aperture, but I personally wouldn’t trust the camera to make optimal decisions for underwater use.
TTL works by comparing ambient exposure alone to exposure including a small pre-flash. The camera then calculates how strong the true flash has to be to get the right strobe:ambient ratio and overall exposure. After the shutter fully opens, the camera fires the strobe and continues the ambient exposure continues until the shutter closes.
For my fish portraits I go from macro focus distances of less than 10 cm to 1-2 meters and each doubling of distance requires quadrupling of strobe power. The distance to fishes can also change rapidly as they erratically swim towards or away from you. In addition, I need to find interesting subjects, get close without scaring them and time the shot for ideal posture or behavior. The less I need to worry about camera and strobe settings the better, so I rely on TTL. Another reason is that I use the camera for recording and not as an artistic tool. Finally, I work mostly with strobe-dominated exposures so small changes in strobe:ambient ratio are not going to impact color palette much, if at all. But it does beg the question; how does TTL decide what strobe:ambient ratio to use?
If you always want to fire the strobe, even in bright conditions, you use fill flash model. On land, fill flash is meant to ‘fill in’ shadows for back-lit or side-lit subjects. The target strobe:ambient ratio is optimized for this purpose and depends on camera brand, ambient light intensity, and exposure mode. The CambridgeInColour website includes the table below as a typical scenario.
Exposure mode | Action |
Auto | bright ambient: no flash dark ambient: 1:1 or higher ratio |
Program | bright ambient: ‘fill flash’ dark ambient: 1:1 or higher ratio |
Aperture or Shutter priority | bright ambient: ‘fill flash’ dark ambient: ‘fill flash’ |
Manual | Whatever is needed for correct exposure |
Fill flash typically means 1:2 to 1:8 strobe:ambient ratio. On land that is fine because ambient and strobe light have a similar color spectrum and you only need to ‘fill’ dark shadows in the subject. However, underwater the strobe needs to restore the red color missing in ambient light and a higher strobe:ambient ratio is warranted. The obvious way to achieve this is to simply use manual exposure mode, which is what I do. However, it may be possible to boost fill flash power using exposure and flash exposure compensation. I don’t know how well this works but will test this on my next trip and report the result later.
Using manual camera control with TTL is very similar to using it with manual strobe control. You underexpose ambient light to the desired level and let the strobe provide what is needed to achieve proper exposure. However, with TTL you do not need to adjust strobe power based on subject distance or other lighting parameters. TTL technology will work it out for you. In my experience, and for my purposes, this works very well.
As mentioned before, aperture is the most interesting exposure to control and in ‘Aperture priority mode’ you set the aperture and let the camera select the exposure time. However, based on the table shown earlier, the camera uses a land-optimized strobe:ambient ratio that is too low. Nikon cameras can increase the strobe:ambient ratio, without changing the overall exposure, by simultaneously applying a negative exposure compensation and an equal but positive flash compensation. For my Olympus camera I can get the same behavior by changing a menu setting (gear>F>EC+FEC>On). Check for your camera to see if this is an option. If it works as advertised it may take the guesswork out of controlling strobe:ambient ratio. I am keen to try it soon.
Because I prefer strobe-dominated exposures, I like to keep the shutter speed fixed at the max. synchronization speed. On my camera I can achieve this in aperture-priority mode, by setting the maximum exposure time to the max. synchronization speed (gear>F>Flash Slow Limit>1/250th). At first this makes no sense. Why let the camera select the exposure time if the only time it can select is 1/250th. It would be simpler to set the shutter speed to 1/250th in manual mode. However, manual mode ties up both dials for exposure time and aperture adjustment, even though I never change the exposure time. In aperture-priority mode only one dial is tied up and I can reprogram the other to set flash exposure compensation. This makes it easy to increase strobe output as needed.
The first rule of underwater photography is to get close to your subject. If you do, then strobes tend to have plenty of power to light your scene during a night dive. However, under bright conditions you need to close the aperture and/or decrease exposure time to underexpose ambient light. Reducing the exposure time does not diminish strobe power and should be your first choice. Stopping down the aperture eliminates both ambient and strobe light and forces the strobe to work harder. If conditions require you to stop down the aperture so much that the strobe can no longer deliver adequate power then you must either get even closer to your subject (perhaps by using a wider angle lens), buy a stronger strobe, or go diving when the sun is less intense.
There are now strobes that can use shutter speeds beyond the max. synchronization time. They do this by firing a rapid burst of short light pulses instead of one big pulse. That way the strobe behaves like a continuous light source and there is no longer a limit on shutter speed. However, using HSS is not the same as having a camera with a faster max. synchronization speed because the latter can reduce ambient light with no effect on strobe light capture. With HSS, the burst of light pulses will be partly blocked by the shutter. In other words, HSS is no solution for an underpowered strobe in bright conditions.
The main benefit of HSS, especially under bright conditions, is that you no longer have to stop down the aperture further than you want. With HSS you set an aperture based solely on your needs and rely on fast shutter speed to ensure proper exposure. In this scenario, what you lose in light being blocked by the shutter you gain by using the wider aperture. For instance, this helps to get black water backgrounds in combination with shallow DOF.
To correctly expose an image you need some way to evaluate the exposure level for the current exposure time, aperture, and ISO settings. In automated exposure modes the camera uses this information to select what it believes to be the optimal settings. In manual mode the camera gives the user feedback in various forms about how far the current parameters are from what it considers to be a correct exposure. This is discussed below.
In ‘reflex cameras’ a mirror diverts light from the lens to an exposure sensor. In mirrorless cameras exposure is measured by the imaging sensor itself. Here I will just consider the latter option, though many of the concepts are the same.
Different camera brands have different metering modes but typically they have at least one that evaluates exposure of the entire frame, one that ignores the corners and edges of the frame, and one that evaluates only a small spot in the center or at the location of the focus point.
In “evaluative” metering (aka matrix metering) the camera evaluates exposure in many locations over the entire frame. By considering many small areas it has a sense of intensity distribution to help it avoid overexposing parts of the image. Evaluating a large area also makes it more likely that the ‘average brightness distribution’ is closer to the assumed 18% gray-scale.
Cameras from different brands may use other information, such as focus distance, location of the focus point and, possibly, subject detection by AI. Evaluative metering is the most advanced option and it is typically the recommended metering mode for general use.
If the edges/corners of the frame are significantly darker or lighter than your subject of interest in the center you can use center-weighted metering to let the central part of the image determine the metering value. The camera adjusts exposure to render the center at an 18% gray-scale level. For example, think of a dark diver surrounded by a bright blue water background.
This is like center-weighted metering but using a spot of only about 2% of the entire frame. Because the spot is small, metering values change a lot as you move the spot to parts of the scene with different brightness. In the past I have used this and pointed the metering spot at my subject, thinking that it would give optimal metering where I wanted it. But if the subject luminance is far from 18% gray-scale, then it will be rendered too bright or too dark. In addition, a dark subject on a brighter background tends to overexpose the latter. To use this properly you want to point the spot to an area that is neither high- nor low-key. Ideally you would lock exposure on an area with an apparent brightness close to 18% gray-scale with a shutter half-press and then recompose to get the proper framing.
Another situation that could call for spot metering is when you intentionally want to underexpose the background. Think about a dive inside a wreck and you want to image the outside world through a porthole. Spot metering on the bright porthole will properly expose the light coming from outside and underexpose the dark inside of the wreck making it appear virtually black.
Some cameras offer spot metering variants that assume you are pointing the camera to a bright (spot-high) or dark (spot-low) area. As an experiment, I adjusted aperture and shutter speed in manual mode to get correct exposure in spot mode. I then switched to spot-hi and the meter reading changed to -2EV. What happened is that by using spot-hi I told the camera I was pointing at a bright object while I wasn’t. It also indicates that in spot-hi mode the camera assumes the subject corresponds to a 72% gray-scale (4 times 18%). When I did the same for spot-low the metering read +3EV. So in this mode the camera assumes the subject corresponds to a 2% gray-scale (18% divided by 8). In diving terms, you may want to use spot-hi to shoot a white nudi on black volcanic sand and spot-low to shoot a dark nudi on white coralline sand. Note: instead of spot-hi and spot-low, you can also use exposure compensation to deal with subjects that are much darker or lighter than 18% gray-scale.
Automated exposure modes select exposure parameters using the results from one of the metering methods discussed above. However, in manual mode the camera needs to give you feedback to help you make the best choices. Cameras give you multiple options with different strengths and weaknesses.
In manual mode, the camera compares exposure with the current settings to its calculated ‘correct’ exposure. It shows the difference in the viewfinder. For my camera this is shown as an exposure bar along the bottom. Dashes below it indicate if the image will be underexposed (dashes extend to the left), overexposed (dashes extend to the right), or correctly exposed (one dash below zero). This is also shown numerically, with negative and positive numbers reflecting under/over-exposure, respectively. For example: if you underexpose by 2 stops, ¼ of the exposure will come from ambient light and ¾ from the strobe, giving a strobe:ambient ratio of 3.
Instead of a numerical exposure value you can just look at the image on the LCD screen or viewfinder to get an empirical sense of the exposure. This is very intuitive but has one major caveat - it does not work well if you use a strobe and ambient light is significantly underexposed. In this situation the ambient light image is too dark to properly judge exposure, focus and maybe even composition. The technical solution is to boost the viewfinder image by automatically amplifying the sensor readout signal as needed. On Olympus cameras this is called Live View Boost (LVB). With LVB on, the viewfinder is always bright. This gives a nice clear image but renders the viewfinder useless for exposure evaluation. Unfortunately, histograms and highlight&shadow displays, to be discussed next, are calculated from this same boosted signal. So, at least on my camera only the numerical exposure indication is reliable when LVB is used.
A numerical exposure value does not indicate the dynamic range of your image. An image with correct ‘average exposure’ may still have areas that are over- or underexposed or it may not make use of the full dynamic range of the sensor.
Histograms are bar graphs showing pixel intensity from zero on the left to maximal on the right. The height of each bar corresponds to the number of pixels with that intensity. Bars at the far left and right of the graph represent under- and overexposed pixels, respectively. Olympus cameras use blue and red for under- and overexposed pixels and superimposed green bars reflect the histogram for pixels in just the central area of the frame. You should check how it works for your camera brand.
If you don’t use LVB, you can inspect histograms before you take the shot to evaluate ambient light exposure. When using a strobe you want the ambient to occupy the left half or less of the histogram, leaving the rest for the strobe to fill in. More typically, you will use LVB and you want to see the dynamic range including the strobe light. In that case you can look at the histogram in playback mode after the shot is taken.
Neither the numerical value nor histogram give spatial information about where in the image high and low intensity regions occur. The highlight&shadow display does this by showing over- and underexposed pixels in red and blue, respectively. For strobe-lit photography this only makes sense after you take the shot and see the combined effect of both ambient and strobe light. You can do this in playback mode.
If you only use ambient light, ambient plus video light, or a low strobe:ambient ratio, then it may be worth not using LVB so you can visually see exposure effects before taking the shot. If you can do without LVB most of the time but need to turn it on occasionally (entering a cave or wreck for instance) there may be options to do so depending on your camera brand. For my Olympus camera switching LVB on/off has to be done via the menu which is cumbersome. But you can switch LVB off in the menu and have it come on when you magnify the viewfinder image (D2>LV Close Up Settings>Live View Boost>On) or if you request a depth of field preview (D2>DOF Preview Settings>Live View Boost>On). Alternatively, you can assign a camera button to trigger S-OVF (simulated optical viewfinder) which has a similar effect as LVB. Check your manual to see if your camera can do the same.
Correctly exposing images while using a strobe turned out to be a much more complex topic than I had anticipated. In studying it I noticed a number of things I have done wrong and discovered other options that I was not aware of and will experiment with on my next trip. I hope you picked up some new insights by reading this. If you find a serious error or want to discuss a topic, you can read me at biodives@gmail.com.