Sep 7, 2004

Using that graph thingy - The Histogram

Many newer cameras come with a built-in histogram function. In this article, we take a look at what histograms are, and how they can be used.

Without a doubt, one of the most useful features of digital cameras is the LCD screen that allows the photographer to inspect a photo immediately after capture. Some cameras, such as the Sony F828 even have a live histogram. This instant view of the image is one of the reasons digital cameras have been so successful. But, while the LCD screen is near perfect for judging composition and framing, it is not so perfect for judging the exposure.

The reason that current LCD screens are unreliable with regard to exposure is because of the way the image is displayed. LCD screens use a relatively small number of pixels and a bright fluorescent light to reproduce the image. The result is that it tends to make photos look brighter than they actually are. I do not recommend using the image displayed in the LCD screen to judge exposure.
Histograms are a graphical means to verify the accuracy of the image preview provided by the camera's LCD screen. The graph represents the 2 extremes of the brightness of any photo: the shadows, on the left, and the highlights on the right.

The distribution of the graph where the spikes and bulges are clustered indicate whether the image is too dark, too bright, or well-balanced.
The brightness of an image is reflected by its histogram. The graphs actually contain spikes throughout. But what is important is to observe the position where the bulk of the brightness is situated. With an image depicting a subject with evenly distributed brightness, the image's histogram will have a distribution of brightness that will be most prominent near the centre part of the graph. An underexposed image's histogram will have a distribution of brightness that tends to be mostly on the left of the graph. An overexposed image's histogram will have the bulge showing the brightness distributed mostly towards the right of the graph.

It is important to understand here that not all images have to exhibit this bulge in the centre part of their histogram. Much depends on the subject of the image. In some cases, it might be entirely appropriate for the histogram to show a dominance at one end or the other, or both.
Generally, the histogram is a reliable way of deciding whether or not the camera's metering evaluated the subject accurately. Should the histogram show an over- or under-exposure, exposure compensation (EV) should be used to correct the situation.

Exposure compensation is a system that allows adjustment of the camera's metering for those situations that can lead it astray. Certain types subjects are notorious for fooling camera meters into producing an inappropriate exposure. Snow, for instance, often causes meters to under expose the image.

While some of the more advanced metering systems can take these situations into account, most metering systems do not, and every metering system will occasionally misbehave.

Should a histogram indicate an under-exposure, the exposure compensation should be set towards the positive (+0.5 or + 1EV) and the photo re-taken with the compensation. Ideally, the process should be incremental so that the result of each change can be observed.
Should the histogram indicate an over-exposure, then the reverse has to be applied and the exposure compensation should be set towards the negative side (-0.5 or -1EV), causing the camera to use either a smaller aperture, or a faster shutter speed.
It becomes obvious that if the histogram is analyzed in the context of the subject of the image, it can be very useful to adjust the obtain a correct exposure.