|
|
| The Digital Corner | ||||
| Mastering the Histogram (Part I) |
||||
| To understand digital photography, you need to understand
histograms. What is a histogram? What does it reveal about the image?
What information does it not provide? What role does the histogram play
in digital photography? These are all valid questions. Now let’s
get some answers. Where traditional photographs consist of negatives and prints, digital images consist of pixels. Each pixel is assigned a specific RGB (red/green/blue) value, rendering it as a specific shade of color on the monitor. The histogram is simply a graphical representation of the pixel contained within an image. In the simplest form, histograms do not provide color information; rather they provide an assessment of the tones contained within an image. Tones range from pure black (a tonal value of 0) to pure white (a tonal value of 255). Between these two extremes are the shades of grey that make up a photograph.
 To
clarify the idea of tones, think of a B&W image. Each pixel has a
specific tonal value ranging between 0 and 255. Lighter tones have higher
values while darker tones have lower values. A histogram is nothing more
then an attempt to display an image’s tones on a bar graph. The
left side of the histogram represents a tonal value of 0 (white). The
right side is black (tonal value of 255). The graph shows the distribution
of pixels between white and black. The higher the bar graph at a specific
point, the more pixels of that specific tone are found in the image. A
histogram that peaks in the center consists of primarily middle tones
(a bighorn sheep sitting on some rocks for example). A peak on the left
side of the histogram, on the other hand, represents an image that contains
many dark pixels.Now, let’s clarify something. Many digital photographers believe that a properly exposed image always produces a histogram that gradually rises from each side and peaks in the center, forming a small mountain. This is true, assuming that your subject is mostly middle tone. Neither dark nor light toned subjects will produce such a histogram if they are properly exposed. Each image will produce a unique histogram, and the correctness of that histogram is based on the exposure that you want, not how it compares to some ideal histogram. Middle tone images consist of mostly middle tone pixels. High key pictures lean heavily towards the right side of the histogram while low key images contain more tones on the left side of the histogram. Earlier in this article, I indicated that a histogram only provides information about tone, not color. This is not entirely true. While traditional histograms do not provide color information, Photoshop CS and CS2 users can apply an advanced histogram display to review color information. In the histogram palette (accessed through Windows ? Histogram if it is not already displayed on the desktop) there is a dropdown menu in the upper right corner. Selecting “all channels view” will display four different histograms. The first (top) histogram is a traditional histogram that displays tonal information. The other three histograms display tonal information for the red, green, and blue channels, respectively. By looking at the red channel histogram, for example, I can determine how prevalent the color red is throughout the tones of the image. While some people find this informative, I see little practical use for this information. As a photographer, I am much more interested in how an image appears then what the numbers say. For this reason, I keep the histogram set to compact view when working in Photoshop. Now that histograms have been defined, what real use are they? After all, why do I care about what the histogram tells me when I can look at the image? Did I really pay over $700 for a computer program so that I can look at a graph to know that my dark image is dark!?! Absolutely not! Well, actually yes. Photoshop can do that, but it can do more. It can also tell me that my light images are light! Seriously, for digital photographers, histograms play three important roles; reviewing exposure in the field, establishing a full tonal range, and global tonal adjustments. Let’s start in the field. DSLR cameras have the option of displaying image thumbnails and histogram information immediately after a picture is taken. While the thumbnail is often too small to provide anything but the most basic evaluation of composition, the histogram provides important exposure information. When looking at the histogram, I am not interested in how pixels are distributed between white and black. This can be adjusted in Photoshop if necessary. What I am interested in is how close the right side of the histogram comes to the edge of the graph. Remember, the right edge of the graph represents pure white without any detail. As more pixels accumulate against the right edge of the graph, more and more highlight detail is lost. As I approach a potential photographic opportunity, I like to take a test shot and review the histogram before getting closer. As long as the histogram approaches, but does not pressed up against, the right edge the image highlights will contain detail. If the test shot reveals that I am too close to the edge, I reduce the overall exposure and take another test shot. Once I am happy with the exposure, I can shoot without changing any settings unless the light falling on my subject changes. Photoshop can be used to bring out hidden details, provided that the information is contained in the original file. A highlight value of 255 records no detail and it is not possible to recover detail that was never recorded. By ensuring that the brightest parts of the image fall below the 255 value, no information is lost. So far I have spent a lot of time focusing on the right side of the histogram. What about the left side? Should I be concerned about pixels accumulating on the left? No. I don’t spend much time worrying about pixels on the left side for three reasons. First, digital cameras are exceptionally good at holding detail in the dark portions of an image. In most cases, a black shadow can be lightened in Photoshop to reveal detail that was not previously visible. The same is not true of highlights. Therefore, I expose for the highlights, not the shadows. Second, dark shadows can always be lightened with fill flash. The same can not be said for highlights. Third, I believe it is better to focus on maintaining detail in the highlights because of how the human eye works. When looking at high contrast scenes, the eyes respond to the brightest part of the scene, not the darkest. By contracting the pupil, the eye reduces the amount of light hitting the retina to see detail in the brightest areas of the scene. The trade off is that shadows are too dark to see any detail as long as our eye is adjusted to see the highlights. Photographically, images that contain highlight detail at the expense of shadow detail appear “correct” because they replicate how our eyes work whereas images that sacrifice highlight detail while retaining shadow detail look “wrong”. Ideally, I want to record detail over the entire tonal range, however, if that is not an option, I prefer to create a photograph that retains the highlights at the expense of shadow detail.
Histograms provide a photographer with a better understanding of how tones interact to create a photographic image. Histograms can also help provide a tool to assist digital photographers in fine tuning exposure in the field. Next month I will discuss how the manipulation of histograms in Photoshop can be used to transform average images into exceptional favorites. |
||||
| Editor's Comment: Let us know what you think! Please email the Editor to let us know your thoughts. |
| HOME FORUMS CHAT LINKS ID RESOURCES ARCHIVE HELP | 
|
