For a long time DSLRs have been sophisticated enough to cleverly expose the scene in front of them. Essentially all DSLRs start by metering the subject, which means they use the in-built lightmeter to determine the intensity of light reflected off or emitted by the subject. The DSLR next calculates the different possible exposures which would bring the subjects intensity to 18% grey.1 Modern DSLRs use sophisticated algorithms to figure out what the subject is - a group of faces, a close-up portrait, a landscape etc. This information helps it fine tune the exposure. For instance if the camera detects one or more faces in the frame, it tends to open the aperture as wide as possible to give a shallow depth of field. Similarly, if the camera detects a landscape shot, it closes the aperture to give as wide a depth of field as possible. More times than not, this automatic exposure works well. So why should one manually expose? For me there are two motivations - one technical and the other philosophical.
The technical one first. When the photographer doesn't manually expose, he lets the camera choose the subject. He also lets the camera decide the intensity of the subject, which it slavishly does (remember, the sole driving factor for all automatic exposures is that the intensity of subject should be 18% grey). This transfers creativity from photography to post-processing. During post-processing it's just not possible to freeze motion, or add motion blur to a background, or simulate a shallow/deep depth-of-fields well. Adjusting the exposure, contrast, editing curves etc. are all the adjustments that can be done to some degree. So, depending on how close the camera exposed the scene to the photographers vision, this post-processing effort could be anything from an impossible feat to zero effort. The photographer should always decide his exposure.
Now for the philosophical motivation - these days DSLRs don't enjoy the automatic image quality advantage over other camera systems (mirrorless, smartphones etc.) that they used to a few years back. The raison d'etre for a DSLRs existence is the quick manual control it affords to the photographer. On a prosumer DSLR body, all controls are just a fingertip away, allowing the photographer complete and immediate control over all aspects of exposure. If manual exposure is not your cup of tea, why go for a DSLR?
Exposure is determined by the Aperture and the Shutter Speed. Exposure, along with ISO speed, or the sensors sensitivity, contribute towards the overall look of the captured image.2 All DSLRs have (at least) 4 modes of exposure control, each mode giving the photographer varying degrees of flexibility and ease to adjust the exposure.3
Programmed Auto(P): The camera evaluates and meters the scene, and comes up with a list of possible, and usually equivalent, exposure values. For example, these are all equivalent exposures: + 1/60s, f/5.6 @ ISO 100 + 1/120s, f/4 @ ISO 100 + 1/30s, f/8 @ ISO 100 The photographer can step through these exposures and pick the one that gives the best look. For a beginner this mode is useful to get a feel for different exposures and how they affect the final image.
Aperture Priority(A): The photographer chooses an aperture. The camera selects an appropriate shutter speed. This mode is useful when photographing static subjects, where control over the depth of field (deep for landscape vs shallow for portraits) is critical.
Shutter Priority(S): The photographer chooses a shutter speed. The camera selects the right aperture. This mode is useful for photographing sports or other fast moving action where creative control over the look of motion (frozen vs a blur) is important.
Manual(M): This is self-explanatory, the photographer picks both the aperture and the shutter speed.
In all modes, the photographer can either specify the ISO or set it to ‘Auto-ISO’, in which case the camera determines the ISO too.
I started dabbling in photography with a simple point & shoot camera. It gave no direct control over the exposure, one had to play around with different profiles (Overcast/Landscape/Profile/Night etc.) to affect exposure. So, when a friend brought along his DSLR for a vacation and I got to play around with it, I was blinded by power! Its buttons and dials allowed me to change the exposure without taking my eye off the viewfinder, just by using muscle memory.
Pretty soon I had bought my first DSLR, a Nikon D3100. I jumped to manual mode. But after a few outings with it, I realized Aperture priority mode worked better for me. I was getting more and better photographs that way. In retrospect, I can see why that was the case. The D3100 has only one command dial. With it you can change only one parameter - either Aperture or Shutter Speed. To control both quickly meant using one of the external buttons to overrride the command dial. But I was using one already to shift the ISO. So Aperture-priority it was.
A few years later, I added the excellent Fujifilm X100S to my kitty. It is a beautiful small mirror-less camera, with mechanical dials and aperture rings. So none of my motor skills with the DSLR adapted to it. Nevertheless, I got pretty adept at handling it quickly. One of the features I liked most about the X100S is it's auto-ISO mode, which was more sophisticated than that on the Nikon. Basically, you set a range of acceptable ISO's and depending upon the aperture and shutter speed that you have set, the camera picks an appropriate ISO. This approach worked well.
I now use a Nikon D500 almost exclusively, with a Nikon D7200 playing the role of a backup body. Both are in the professional to prosumer end of Nikons' range of DX cameras and have two command dials. This allows me to effectively work in manual mode. Sometimes, depending upon the situation, I use auto-ISO to help me.
For the most part my photography has been restricted to landscapes and events. My method to expose for these underwent considerable change over the years before I settled down on something that works very well for me now.
However, over the past year or so, I’ve been getting more and more interested in bird photography. It is here that my methods are still quite raw and see experimentation.
To quickly switch between these two ‘modes’ I use custom banks on my D500 and the U1/U2 settings on the D7200. The first mode, for static subjects, is fully manual. Both cameras give me quick access to shutter-speed, aperture and ISO without having to take my off the viewfinder. The second mode, for bird photography and other action oriented photography, differs from the first only in that auto-ISO is ON.
Over the years my method to expose has evolved. Initially it used to be a simple feedback loop - 'guess' the appropriate exposure, take a photo, look at the preview, adjust the exposure, take another photo, check the preview and on and on till the preview looks good. I ignored the lightmeter.
Since this was an iterative process, I ended up taking several photographs with slightly different exposure settings. Besides being time consuming during photography, it also meant greater effort during post-processing - to import and triage. Nevertheless, it was a good learning experience - there is no substitute for experimenting with settings. Playing around with different settings and seeing for myself how they affected the image in the back of my camera helped me get a good handle on the basics of exposure.
The preview on the cameras LCD screen doesn't really represent the actual exposure, it depends on the quality of the screen, its brightness and the preset that the camera applies to generate the preview. Since I prefer the look of dark, moody images more than bright ones, several times I’d get an image more underexposed than it should have been. I handled this during post-processing, by boosting shadows, but this added noise to the overall image, worsening it’s quality. In retrospect, relying on the camera's built in LCD screen for judging the exposure just wasn't a great way to work.
The next stage in my method was to start using the lightmeter in conjunction with Spot or Center-weighted metering. I moved the central focus point (because it is the most sensitive) onto my subject and used the lightmeter to see if it was under or overexposed. Accordingly I made changes to the aperture or shutter-speed as the situation warranted. If light-levels were too low, then I’d step up the ISO. Once I had the exposure and ISO locked, I recomposed and took the photograph. This was a big step forward in my photographic process - I was was using a deterministic approach to arrive at the exposure instead of by trial and error. As I look back at the images I took during this period, the general quality of exposure shows an improvement and the number of edits done in post a reduction from before.
I deviate from this approach sometimes, when the scene has a high dynamic range despite being well lit. My usual approach in such situations would result in a base ISO exposure but with too dark shadows. Pulling them up during post-processing adds noise.
For instance, this was photographed late on an evening around sunset. The sun is low, near the horizon and therefore in the frame, which gives the scene fairly high dynamic range. The camera ends up underexposing the image, since its exposure calculation is dominated by the bright sun.
Here, I’ve brought the overall exposure up a bit and also increased the shadow level in Lightroom.
A 100% crop of this post-processed image shows the amount of noise that is now apparent.
This is where Expose-to-the-Right (ETTR) helps.4At base ISO of a camera, to get the least amount of noise in the final image, it is best to maximize the amount of light captured.5 But this is a fine line - let in too much light, and you lose highlights in the image. So the trick is to over-expose as much as possible before clipping the highlights. After having done some tests, I know that on my D7200 and D500 an overexposure by 3 stops can still be recovered in post-processing. Similarly, at the other end, an underexposure of about -3 to -3.5 stops can be recovered before details disappear amongst pure noise. With this knowledge, my aim was to overexpose the brightest parts of my subject by 3 stops. This maximized light capture and allowed me to bring the exposure down in post, reducing noise levels in the overall image.
For instance, this is a scene which suited itself well for ETTR. I overexposed the image by about 2 stops. The captured image looks completely overblown.
In Lightroom, I brought the exposure down by 2 stops and recovered highlights. The image (after a few more adjustments) looks a lot better now.
A 100% crop of this image shows how little noise there is.
In bird photography, the subject and conditions typically don’t allow me to work with the same method as before. Birds keep hopping around, move behind bushes, fly into or out of the shade which keeps changing the settings needed for a good exposure. Most of my concentration when photographing birds is on keeping them in focus and in the frame. So, for my exposure I need as much help as possible from the camera.
I manually set the shutter speed and aperture, but use auto-ISO. The shutter speed is set at 1/1000s as baseline so that I can quickly move up or down as dictated by the situation. The aperture is typically set to the widest possible to gather as much light as possible. So in most situations it is only the shutter-speed that I control. Since my focus point is always on the bird (in as much as it’s possible), the camera exposes for it. Therefore to counter it, I dial in an exposure compensation of about +1 stop. More times than not, ETTR doesn’t matter here, since ISOs I’m working with are well above the camera’s base value. Almost all the keepers in my In the Wild series have been captured this way.
Nevertheless, in several situations the camera gets it wrong. The commonest is when the bird is not big enough in the frame and the camera ends up exposing for the background or elements around the bird, resulting in an under- or over-exposed bird. At other times when the bird is itself dark, say a raven or an ibis, the camera again overexposes increasing the ISO to the high thousands.
My experiments to nail exposure in these cases continue. Eventually I want to be using complete manual mode for bird photography as well. That requires a lot of experience to ‘know’ the exposure for a given situation and also some nimble camera work on it’s command dials without taking my eye off the viewfinder.
This subject could be a spot, a centered area of the frame or even the entire frame; each of these is a so-called Metering Mode. ↩
Perspective and quality of light are some other factors, which I don’t talk about in this post. ↩
I describe the modes on Nikon DSLRs. Canon and Sony have equivalent but perhaps differently names modes. ↩
Why this holds good only at base ISO is best explained with an example. Let's say at ISO 100 overexposing by 3 stops requires too high a shutter speed than feasible. So you jump to, say, ISO 400. At this ISO, you’ve lost out on dynamic range and have intrinsically increased noise. Also, depending upon your sensor, it might have been better to underexpose at ISO 100 and raise expose during post. ↩
Noise is a fiendishly complex topic. At the simplest level, there are two kinds of noise in DSLRs - Shot noise and Read noise. Shot noise is inherent in the light that is captured, while Read noise is that which is introduced by the camera when reading off the sensor. Maximizing light captured improve SNR with respect to shot noise, while lower ISOs help reduce the read noise. This is why ETTR works, overexposing just the right amount at base ISO maximizes SNR and minimizes read noise. Finally underexposing during post-processing reduces any visible noise even further. ↩