Manual exposure mode (M) is often perceived as a mode that only most advanced photographers can use. While it is true that manual mode will immediately highlight any exposure errors, which is why it is also often used in photography schools to teach how exposure works, manual mode also provides some advantages when compared to aperture priority (Av) and shutter priority (Tv) modes. Let's take a closer look.
Any exposure metering device, be that a compact camera, an SLR or a light meter, measures the amount of reflected or incident light falling on the photosensitive surface within a given amount of time. For reflected light, the exposure is computed to place the captured light in the middle of the range supported by the as film or a sensor, which renders it gray regardless of the actual subject brightness.
The latter may seem contrary to what the common sense would suggest, but you can see this process in action if you take two shots in any automatic mode (e.g. Av, Tv or P) of two monotone-colored surfaces, one darker than the other. Many would think that the shot of the darker surface would actually look darker than the other one, while what you will see is that both shots will be in the mid-tones.
The following three shots are taken in Av mode at f5.6 and ISO 200, so the camera adjusted the shutter speed in order to set the exposure. The first shot just shows the scene, while the second and third were taken using spot metering pointed to the sunny side and shadow side of the walkway and converted to black and white using luminance. Spot metering was used to make sure that just one concrete tile was metered and other parts of the frame are ignored. Hover the mouse over each title to see the difference and their histograms. Note that the larger spike, which corresponds to the larger area of brightness in each shot, is in the same location on the histogram.
If there are two levels of brightness in the area of the frame covered by active metering zones, the camera computes an average between the two values and places the average in the middle of the resulting histogram. Similarly, in a real-life shot, all brightness levels in the part of the scene covered by active metering zones are averaged out and the average is placed in the middle of the histogram.
The shape of the metering zone may be different from a camera to a camera, but it’s important to realize that it’s some geometric shape, such as a circle, rectangle, hexagon, etc, which means that each shape will cover a part of the subject and will also contain parts of the background, so the computed average will include less important parts of the scene mixed in as well.
For example, you are taking a picture of a bird in flight against the sky, the metering zone will cover a part of the bird, as well as the sky. The sky is usually much brighter than the bird, so the average will have a relatively high value and the camera will compute the exposure to make the shot darker in order to place the computed average to the middle of the resulting histogram, producing a silhouette of the bird against a well-defined sky that has good color and cloud details.
In order to correct automatic exposure, exposure compensation has to be used, which shifts the exposure by the specified amount. For example, adding two stops to the computed exposure will make the entire shot brighter, so bird detail will become visible, but the sky will most likely lose some detail and color if the blue channel is clipped.
At the first glance, it appears that exposure compensation remedies the problem, but in practice it is not quite so. The active metering zone will move between the bird and the sky as the bird is being tracked and a different amount of exposure compensation would be required for each ratio of the subject and background covered by the active metering zone. Similarly, different exposure compensation would be required when a bird moves between light and dark backgrounds.
This is where manual exposure comes in - exposure parameters do not change automatically in manual mode and as long as the subject is illuminated by the same amount of light, it will be exposed as intended.
Manual exposure works well in most cases, except where light changes dramatically and frequently within a short period of time. Shooting on days with variable cloudiness or on stage with a lot of moving search lights does present some challenge in having to adjust the exposure on the fly.
Before jumping to manual metering, let’s review most commonly used metering modes available in modern cameras, which include multiple metering zones (sometimes called matrix or evaluative metering), center-weighted average and spot metering.
Multiple metering zones mode is driven by the AI in the camera and it works well with automatic modes (Av, Tv and P). It is usually linked to one or more active AF point to make sure the exposure is computed for the subject identified by the active AF point. It may not work as well for focus-and-recompose shots. Another downside of this mode is that it’s not clear which metering zones were used in the last shot and whether they will be used again, which is essential for exposure compensation to work reliably.
Center-weighted average meters light across most of the frame, giving more weight to the light falling on a large circle that covers most of the frame height. It works well for evenly-lit scenes or scenes where light and dark areas even each other out.
Spot metering uses a small circular zone in the middle of the frame that covers about 2-3% of the frame. Spot metering is the most useful metering mode for manual mode shooting because it allows the photographer to meter without having to look at the resulting image for additional exposure compensation. However, spot metering does require some skill to use and, unlike automatic metering modes, the camera can no longer be used as a point-and-shoot because most first shots of the scene will require some set-up time.
You can see the difference in how each mode handles the same scene in these shots taken in Av, except the last one. No exposure compensation was applied in automatic modes (otherwise all shots would look exactly the same). Spot metering in the EOS 7D covers about 2.3% of the frame. Canon's partial mode is similar to spot, but covers 9.4% of the frame. The shot labeled spot (M) is taken in M and exposed to preserve the highlights. A slight upward curve was applied in post-processing to brighten the shadows while keeping the highlights more natural.
All automatic modes clipped the sky, which is especially visible in the tree branches in the top left corner, where a bit of blue is visible between the branches, but not in the sky above. You can also see that clipping of the red channel introduced more cyan in the reflection of the sky. Note that the partial metering shot is slightly brighter because there is more dark in the center of the image and the camera shifted the histogram slightly to the right in order to compensate. Interestingly enough, spot metering in Av mode made the shot darker again because the bright gravel and dark reeds in the center of the image balanced each other out.
It is also important to realize that using spot metering in automatic modes and without exposure lock will often produce over- or underexposed images, depending on what happened to be in the metering circle.
Many cameras also implement a mode where shutter speed and aperture can be set manually and the camera then automatically adjusts ISO speed to set the exposure. This is not a manual mode, even though the mode dial may be set to M, and using this mode still requires exposure compensation to account for the background covered by the active metering zone.
A tonal curve is a transformation function applied to RAW image data before any post-processing is done. One of the purposes of the tonal curve is to map a large number of brightness levels in the scene onto the smaller range most devices can reproduce, at the cost of losing some detail in the shadows and highlights.
In order to see how the tonal curve works, let’s take a look at an out-of-focus shot of a grayscale gradient rendered on a computer screen. Pixels in the original histogram are evenly spread throughout the entire range of possible pixel values. Pixel values in the shot of the gradient show less detail in the both ends of the histogram and more details in the middle. This transformation is caused, to a lesser degree, by the gamma correction applied to the original image displayed on the screen and mostly by the tonal curve applied to the shot of the histogram.
Digital sensors are linear devices and generate signal directly proportionally to the amount of light falling onto the photosensitive elements. Without a tonal curve applied, most images would show detail only for specific areas of the image, such as sky. Some image editors provide controls to disable the tonal curve and doing so illustrates quite well how much applying a tonal curve affects the image. In DPP3, click the Linear checkbox in the toolbox to see the effect.
RAW images have so much data that applying the tonal curve and even further adjusting exposure, within reasonable bounds, will not introduce negative side effects, such as banding or loss of detail. These bounds will vary with ISO speed and usually are within about a stop at lower ISO speeds and practically none in the higher ISO range.
Pixels that are close to the either edge of the histogram are not so easy to manipulate - shadows will show excessive noise if brightness is increased too much, which will be more profound as ISO speed increases, and highlights may contain clipped channels, which may introduce abrupt luminance and color transitions if brightness is decreased too much.
For example, this shot taken with a manual flash was overexposed by two stops and brightness was lowered in post-processing. However, clipped highlights are unrecoverable and lowering brightness introduced color banding. The curve had to be adjusted upwards in the highlights to smooth out the difference between channels, which created this washed out look.
The shot below, on the other hand, was underexposed by as much as three-four stops and adjusted in post-processing. You can see significant noise throughout the entire picture, even though the shot was taken at ISO 200. Hover your mouse over the 100% crop label below to see the read-out noise (vertical lines), as well as remnants of chrominance and luminance noise.
In addition to channel clipping and excessive noise, pixels falling on the sloping shoulders of the tonal curve will also have less contrast.
Manual mode can be used most efficiently with spot metering. It doesn't mean that other metering modes cannot be used in manual mode, but without knowing what metering zone is used, a couple of test shots might be needed to figure out the correct exposure. Spot metering makes it possible to shoot without ever having to look at the camera screen.
The basic rule of spot metering is very simple - pick any surface in your scene and decide what relative level of brightness this surface represents when compared with the rest of the scene, then point the metering circle to that surface and move the metering arrow to that brightness level.
For example, if you intend to shoot on a sunny street, point your camera to the concrete pavement in front of you and choose your shutter speed to place the metering arrow on the +1 and 2/3 mark. As long as your subjects are in the same light, you can be sure that your exposure is correct, regardless of what background is behind the subject.
Things get more complicated once you start working with mixed surfaces. For example, if you point the metering circle at a person’s face, you would normally move the metering arrow to about +1. However, if the metering circle also covers person’s dark hair, then you have to move the arrow closer to the zero, depending on how much hair is in the circle and how dark is the hair.
Knowing how the subject brightness changes in different light is always a big help. For example, the difference between a sun-lit surface and the same surface in the shade that is less than a few meters away from the sun/shadow boundary is about 3 stops and about 4-5 stops further away in the shadows. If a person moves from the sun to the shadows, you can just increase the exposure by three-four stops without having to waste time on metering.
The shot below was taken in a harsh sunset light. Keeping ISO speed as low as possible for an acceptable shutter speed increased the dynamic range of this shot and preserved highlights while leaving enough data in the shadows to allow making shadows brighter in post-processing.
The exposure had to be adjusted by 3 and 2/3 stops for the shot in the shadow of a building. Changing the ISO speed immediately by 3 stops and then adjusting the shutter speed as necessary shortens the metering time, which is very important for scenes that don't last very long.
Shooting manual is all about knowing the brightness of your subject relative to the rest of the scene, knowing the dynamic range of your camera and how much you are willing to sacrifice in the highlights or shadows in order to capture your subject with as much detail as possible. It’s also important to realize that any detail falling onto the shoulders of the tonal curve will have less contrast, compared to the image data that falls onto the straight part of the curve.
In general, it’s best to set the exposure to place the subject either into the middle of the available dynamic range or slightly to the right, but not too close to the right end of the histogram. Spot metering works very well for this purpose because you can meter highlights, shadows and the subject separately.
For example, if you are setting up a general landscape shot, point the metering circle to the sky and set the exposure so the sky is about +2 stops. Check darker areas of the shot, which will be at about -3 mark. This does not mean that the scene has only 5 stops - remember that spot metering averages out the area of about 3% of the frame and there still may be pixels brighter than the average within this circle.
On the other hand, if the primary subject of your landscape shot are the clouds, it’s better to set them at +1 to make sure to preserve as much detail in the clouds as possible, and then deal with shadows in post-processing. If shadows are as important, using multiple exposures is a good choice as well (sometimes it‘s the only choice).
If you are pointing the metering circle at a surface that has mixed dark and light areas, the circle will move slightly between these areas due to subject movement and camera shake, so the metering arrow will jump erratically, making it more difficult to set the desired exposure while pointing the circle at the chosen spot. A good technique in this case is to choose a spot to point your camera to before raising the camera and decide where you would like this spot to end up on the histogram, then point the metering circle to this spot and note where the metering arrow is. Now you can move the arrow to the desired location by either changing the shutter speed or the ISO speed and without having to keep the metering circle in the same spot.
For example, if you pointed the metering circle at a person’s face and the metering arrow was at -1, then you can drop the shutter speed by one stop and increase ISO by one stop without having to keep the metering circle on the face, which will take the metering arrow to +1 if you were to point it at the face again.
Shutter Speed, ISO Speed or Aperture?
Metered exposure is just a single number and there are three parameters that can be changed to arrive at this number, which can be intimidating or time consuming at times. Let’s take a look at whether to change shutter speed, aperture, ISO speed or all of the above for the next shot.
Aperture is the most important parameter and has to be chosen very carefully. Small changes in aperture values will have dramatic effect on the how the resulting image will look like, while changes in ISO and shutter speed are less noticeable, within reasonable bounds. That is, besides the obvious changes in depth of field and background blur, wide-open lenses will produce visibly softer images due to the effects of spherical aberrations, compared to a lens stopped down two-three clicks. A very narrow aperture, on the other hand, such as f11 and less, will produce softer images due to the effects of diffraction.
Similar changes in shutter speed and, to some extent, ISO speed, will not be as visible. For example, photographing a running person at 1/500 or 1/1000 won’t make much difference. Along the same lines, switching from ISO 400 to ISO 800 will not change image quality much, but could allow you to shoot at a faster shutter speed.
With this in mind, always choose aperture first to suit your scene and get the best from your lens. For example, for street walk-arounds, you can open the aperture, which will produce softer pictures, but will better separate subjects from the busy background of the street. For bird photography, you may want to close the aperture a couple of clicks to capture more detail, especially if your lens isn’t very good optically. If you plan to fit more than one subject in a shot, you need to close down the aperture even more to make sure all subjects are in focus.
After setting the aperture, pick the shutter speed range that matches your subject movement and the focal length. A walking person requires shutter speed of about 1/250, a running person requires about 1/500, a bird in flight requires minimum 1/1250, a bird splashing in a bird bath requires 1/2000 and so on. You should be able to adjust shutter speed about 2/3 of a stop in both directions without affecting your shot if the light changes. For example, if you set shutter speed to 1/500 and intend to photograph people on the street, then you can adjust your exposure between 1/320 and 1/800 without having to switch ISO speed, which usually takes more button clicks in most cameras.
October 14th, 2013
After shooting almost every day for a couple more years I find that dropping the shutter speed by 2/3 of a stop is often counterproductive for shots with any moving subjects and instead I started to adjust the ISO speed to accommodate any changes in light. For example, if in the past when I needed to adjust the exposure by -2/3 of a stop, I often reduced the shutter speed from 1/500 to 1/320, now I bump up the ISO speed by one stop and increase shutter speed to 1/640, which still gives me the same 2/3 of a stop increase, but without dropping the shutter speed to the point where I would likely end up with a blurry shot.
Note that no metering took place at this point yet - you just set the aperture and picked the shutter speed range sufficient for the scenes you intend to capture. Now point your camera at any surface illuminated by the same light as your subject eventually will be and change the ISO speed to set the exposure, so that this surface is placed at the intended range on the histogram. For example, point it at a light-shaded pavement on the street and set the metering arrow in the range between +1 and +1 2/3, depending on how reflective is the pavement. Experiment a bit with different shades to get a better feel of this process.
After this exercise you will have the camera set to shoot anything in this light without having to change your exposure, which usually is handy if you don’t have the time to adjust it right before the shot. As you move from light to the shadows (e.g. moving through sunny and shady areas of the street), you need to adjust the ISO and the shutter speed by about 3 stops if you want to be able to shoot without having to adjust the exposure before each series of shots.
When you do have a few seconds to adjust the exposure when your subject is in front of you, then pick the part of the subject you want to meter before raising your camera and only then meter off that surface. You can quickly adjust the exposure by changing the shutter speed by about 1 stop or so, but for bigger adjustments you need to change the ISO speed first and then follow up with a shutter speed adjustment. It is often more convenient to set your camera to change ISO speed in full stops, as opposed to the default 1/3 of a stop - takes fewer clicks to adjust the exposure this way.
High ISO Speed
When you have to choose between having a very noisy shot and not having a shot at all, the choice is obvious. Never set limit on ISO speed, but when shooting at high ISO, keep in mind that there is no room for brightness adjustment in post-processing - any attempt to increase brightness will make noise very visible. When choosing exposure for a high-ISO shot, place your subject in the right part of the histogram, but be careful not to clip important highlights.
Is there future for M?
It may seem that with new advancements in metering technology and subject detection techniques, manual metering will become a thing of the past, but it is not likely to be so because manual mode offers very predictable outcome and it can be used very effectively with electronic viewfinders, which become more and more common.
For example, let’s take a look at a compact camera that has the M mode, such as Canon’s S95 or any similar that does not simulate "correct" exposure on the LCD screen or has a way to turn it off (i.e. if your shot is underexposed, the screen should look dark). Just like with an SLR, you need to set the aperture you want and then pick ISO speed to put your shutter speed into the range suitable for your scene and then spin the exposure wheel to set the shutter speed until the picture on the LCD looks good. You don’t even need to meter! Note that the metering arrow is in the middle of the scale in a well-exposed shot below just by coincidence and it won't always be the case.
Once the exposure is set, you can shoot anything in the same light without having to worry about camera metering off the wrong surface or having to use exposure compensation and if you see the scene become too dark or too light, you just turn the wheel and change the shutter speed to make the scene look good again (or ISO speed to keep the shutter speed within the selected range). It’s that simple - no more worrying about whether the camera picked the right subject, trying to figure out exposure compensation if you want to focus on a black suit, but meter off the white dress, etc.
While it is possible to shoot reliably in manual mode with a good electronic viewfinder, one additional disadvantage of electronic viewfinders is that they consume energy to power the internal LCD screen. Having to keep the electronic viewfinder on for every shot will shorten battery life. This may not matter much for occasional shooters, but those who spend a few hours a day with the camera in their hand will feel the pinch.
Spot metering, on the other hand, may not do just as well - many photographers nowadays seem to choose matrix metering when shooting in manual mode. Consequently, they sometimes have to take a shot or two to figure out the exposure, which works out nicely for relatively static scenes, but will result in wasted time for those scenes that don't last very long. Hopefully, camera manufacturers will keep proven technology alive for those photographers who rely on their skill when taking pictures.
Shooting modes do not define one as a photographer - what's in the shot does. Over the years I saw most beautiful photographs taken by people with compact cameras that offer very limited control over shot parameters and far too often I saw photographs captured with SLRs and rangefinders that made me wonder why their owners spent so much money on something they cannot use. However, some shooting modes are more counterproductive than others and only those who shoot in Av and M can fully appreciate what it takes to capture the photograph that comes out just as it was intended and only those who shoot in M and use spot metering, with enough practice, know what they will have captured even before they press the shutter release button.