Viewing and using your camera's histograms

On your camera and in your image editing software, the most helpful tool for assessing your exposure settings is the histogram. A histogram is a graph of the tonal values of an image, from black on the left to white on the right. The height of the graph at each point represents the relative number of pixels in the image with that particular tone or brightness level. The higher the graph, the more pixels there are of that particular tone.

On your camera, you can select to view either the luminance histogram, which displays the range of overall brightness levels from black to white, or the RGB channel histograms, which show the range of tones for each of the three colour channels.

In a colour channel histogram, the more the graph rises to the right-hand side of the histogram, the more saturated that colour will be in the image. Conversely, if the graph rises to the left of the histogram, the more muted that colour is in the image. For judging the exposure of your image, the overall luminance display is more useful, unless you are shooting a very vibrant and colourful subject where you could easily clip one or more of the colour channels even if the overall exposure looks satisfactory – more about this shortly.

When you're reviewing an image on your EOS camera, repeatedly press the Info button to scroll through the different viewing options until you see a histogram. The first histogram view you'll see is the luminance histogram; to view the RGB colour histograms in addition, press the Info button again.

On the luminance (brightness) histogram, the horizontal axis represents the range of tones from black on the left to white on the right. The higher the graph at any point, the greater the number of pixels with that tone – or to put it another way, the larger the area of that tone in the image.

The graph shows the relative number of pixels rather than the absolute value. The more the graph skews towards the right-hand side, the brighter the image or the subject. Conversely, if the graph peaks towards the left, you are either photographing a very low-key subject or the image is underexposed. If the graph tails off before reaching the right-hand end, this means there are no areas in the image that are pure white. If it does not reach to the left-hand end, this means there are no pixels that are solid black.

The theoretical ideal histogram will be bell-shaped (a normal curve), with the graph just reaching the ends of the brightness scale and the main peak being around the centre. However, in real life it's likely to look quite different. With a landscape, for example, you're likely to have a peak to the right that represents the pixels that make up the brightest parts of the sky, and a peak towards the left end that indicates the pixels in the deepest shadows. In the middle, you'll probably see a number of peaks representing the mid-tone pixels which are likely to make up the grass and foliage in the scene.

Achieving the ideal bell-shaped curve is not critical. What is important is to check for tall peaks that are cut off at the extreme ends of the histogram, because this indicates that the shadows (on the left) or the highlights (on the right) are clipped. Clipping is when some tonal detail or gradation is lost and cannot be recovered. In a landscape shot, for example, it's easy to lose the detail in brightest clouds, which means that they are clipped to a uniform white with no gradation. This will be indicated by a cut-off peak at the extreme right-hand end of the histogram. You can address this by reducing the exposure a little and reshooting so that the peak in the histogram moves slightly to the left and is less cut off (and ideally not cut off at all) at the right-hand end.

Generally speaking, in the case of digital images, more detail is recoverable from shadow areas than from blown highlights, so a good rule of thumb when shooting with a digital camera is to expose for the highlights and recover the shadows. In histogram terms, this is often expressed as "expose to the right" – get the image as bright as possible without clipping (at least excessive clipping) in the highlights.

If you reduce the exposure a lot, all the peaks in the histogram will move much further to the left of the graph, indicating that the image is much darker than your first shot. Although shadows can be brightened post capture, it is not possible to recover colour and detail from areas that have been completely clipped to black. Hence, with very high contrast scenes, it's important to decide where you want the detail in your image – the shadows or the highlights.

The RGB histograms show the brightness of the three individual colour channels, and they are especially useful when you're shooting a scene with a subject that is predominantly one colour.

Because the luminance histogram shows the overall brightness values across all three channels, it may not make it apparent that your single-colour subject has some clipping – that is, some subtle variations of colour are being lost. By checking the RGB histograms as well as the luminance histogram, you can adjust the camera's exposure settings to ensure you get the best result possible.

This doesn't mean a perfect solution is always possible, of course: in the case of our example shot of a red flower on a yellow background, underexposing might reduce the clipping in the red channel but will also impact the green channel. In such cases the best answer may be to take a number of shots, bracketing the exposures, and produce a composite image in post-production combining the best-exposed versions of the different parts of the image (which you can assess with the aid of the RGB histograms). The Compositing tool in Canon's Digital Photo Professional (DPP) software can help here, as can similar tools in other image editing software.

When the Highlight alert is enabled in your EOS camera's playback menu, you may notice that some areas of your images flash when you review them. This flashing indicates where there are overexposed pixels – that is, those areas have been clipped to white and tonal detail has been lost. A quick glance at the histogram view will confirm this, and if necessary you can adjust the exposure accordingly. This display is very useful if you have opted to "expose to the right" because it helps you assess whether clipping is occurring in any important areas of the image or is confined to relatively small or unimportant areas, and is thus an acceptable trade-off for getting the image as a whole as bright as possible.