Why the Camera Gets It Wrong
Your camera has several exposure modes for a reasonit's not always easy to obtain the correct exposure. A selection of modes lets you choose the one that's most comfortable for you, or that suits a certain situation.
My way of working is to use evaluative metering. (This is Canon's terminology; other manufacturers use different terms such as pattern or matrix metering.) Evaluative metering is my camera's most advanced metering mode. The camera divides the photo into zones and analyses them, weighting the readings toward the active autofocus point, and makes an exposure decision according to what it "sees."
Whichever metering mode you use, check your camera manufacturer's website for a detailed article on how the metering works. The better you understand your camera's exposure modes, the more likely you are to recognize scenes that may mislead the camera into giving an incorrect exposure.
The most common situations that create problems are subjects with lots of dark or bright tones, because your camera's meter works on the basis that all the tones within the scene average out to mid-gray. The meter tells the camera what exposure settings to use to achieve this result. This approach works well most of the time, which is why cameras use this system.
But what happens when you're facing an unusual scene, such as a snow-covered landscape? In this case, the camera will give the exposure setting required to render the snow in gray, resulting in an underexposed photo. When this happens, as confirmed by looking at the histogram, it's up to the photographer to override the camera's settings.
The easiest way to override is by using the exposure compensation function. This feature varies from camera to camera, so check your instruction manual for details. On my EOS 60D, exposure compensation is easy. I just turn the Quick Control dial on the back of the camera with my thumb to increase or decrease the exposure levels. If I think the camera is giving me the wrong exposure reading, I can compensate without moving my eye from the viewfinder.
Raw Versus JPEG
The appropriate approach to exposure varies, depending on whether you're using the JPEG or Raw formats. If you're using JPEG, aim to make the exposure as accurate as possible, on the basis that you have limited leeway to adjust the photo in post-processing. The histogram is a good aid for this technique, as it's generated from the JPEG file and is therefore an exact match for the image.
On the other hand, things get a bit more interesting if you use the Raw format. The aim of exposure changes somewhat, from getting a precise exposure to getting one that includes as much information in the image as possible. You can think of a Raw file as the digital equivalent of a negative. The aim is to capture a Raw file that contains as much information as possible so that you can create the best possible image from it when you process it.
You do this with a technique called exposing to the right (often referred to as ETTR in online articles). This means that you give the photo as much exposure as possible without clipping the highlights. The histogram should be as close to the right side of the graph as possible without going past it. This technique works best with low-contrast subjects, giving you an image that in most cases is too bright. This problem is easily fixed when you process the Raw file with the exposure or brightness controls.
At the beginning of this article, I explained that if you underexpose an image and increase the brightness in post-processing you'll increase the noise levels in the darkest tones. Exposing to the right does the opposite: By overexposing and then reducing the brightness, you reduce noise levels. This technique comes into its own at high ISO levelsif you've never tried it, you'll be surprised at the increase in image quality you get.
Raw and the Histogram
I hope that the information in this article will help you to improve the accuracy of your exposures. Interpreting histograms is a bit of an art, and you'll get better with practice. The end result is better-quality files that need less work in post-processing to bring out their best.