Friday Foto Talk: Dynamic Range III – The Histogram   7 comments

Elowah Creek, Columbia River Gorge, Oregon.

Elowah Creek, Columbia River Gorge, Oregon.

This is the third of four parts on Dynamic Range.  Check out the first two parts.  Part I is an introductory look, including what dynamic range actually is.  Part II goes into how your eyes and your camera see things differently.  Today we’ll look at a key tool you should be using and how it help you both understand and control dynamic range.

Beacon Rock, Washington side of the Columbia River Gorge.

Beacon Rock is a landmark on theWashington side of the Columbia River Gorge.

The Histogram & LiveView

With most any digital camera, when you capture an image the camera builds one or more histograms, which are attached to the image file.  Histograms are simply graphs with horizontal and vertical axes.  Most used (and useful) is the histogram that depicts tones (See Figure 1 below).  These show tone on the horizontal axis and the amount of pixels for that tone on the vertical axis.  On the far left is pure black (0, the black point); on the far right is pure white (255, the white point).

histogram_labeled

Fig. 1: Please click this image (which is not mine) to go to the source, a nice introductory blog post on histograms.

With most DSLRs, you can capture the image in LiveView.  In other words, you can use it like a camera phone, viewing the scene on the LCD instead of through the viewfinder.  This allows you to see a representation of what your final image will look like.  Unlike most camera phones, it also allows you to see a very accurate, live-action histogram of the scene before you.  This way you can see what the tonal distribution actually is before you fire the shutter.  For example a live histogram will show exactly how much of the scene is overexposed or underexposed (see Figure 2 below).  When I use LiveView I’m usually on a tripod.  I’ll frame up my composition by looking through the viewfinder.  Then I turn on LiveView to see how things look on the histogram.

Figure 2. The histogram on the left is of an image with a small area of underexposure.  The one on the right has a small area of overexposure.  Click on image to go to the source website.

Fig. 2. The histogram on the left is of an image with a small area of underexposure. The one on the right has a small area of overexposure. Click on image to go to the source website.

 

Dynamic Range & Your Histogram

When you capture an image with a lot of inherent contrast (a sun-dappled forest or the side of a sunlit barn with the door open), your histogram will show a curve that spans across the entire width of the graph.  And it will probably climb up the sides (see Figure 3 below).  For scenes with low overall contrast, the histogram of a properly exposed image will sit somewhere near the center of the graph with the curve dropping down to the bottom before it reaches the far left or right (see Figure 4).

 

Figure 2. The image would have a small amount of pure black (far left).  A little more of it is overexposed highlights (far right).  Click image to go to the source website for this.

Fig. 3. This high-contrast image would have a small amount of pure black (far left) plus an area of overexposed highlights (far right). Click image to go to the source website for this.

A casual shot in a city park at sunset.  This image has both clipped shadows and blown out highlights (the sun, which is okay!).  Thus its histogram is similar to Figure 3.

A casual shot in a city park at sunset. This image has both clipped shadows and blown out highlights (the sun, which is okay!). Thus its histogram is similar to Figure 3.

Figure 3. Three low-contrast histograms. Dark on left, bright on right.  Click image to go to the source website this.

Figure 4. Three low-contrast histograms. Dark on left, bright on right. Click image to go to the source website this.

This black and white image of a stormy Columbia River Gorge is mostly shadow and mid-tones, with modest contrast.

This black and white image of a stormy Columbia River Gorge is mostly shadow and mid-tones, with modest contrast.  Its histogram would be similar to the one on the left in Figure 4.

 

These two situations, low contrast and high contrast, are of course extremes in a continuum.  And one situation is no better than another.  But when faced with a scene that has a lot of overall contrast, a camera with good dynamic range will expose so that the histogram (again, of a properly exposed picture) comes close to but does not climb up either the left or right edges (see Figure 5 below).

Fig. 5. Histogram of a high-contrast scene.  Click image to go to the source website.

Fig. 5. Histogram of a high-contrast but properly exposed scene. Click image to go to the source website.

Garden of the Gods, Colorado.  This image has fairly high contrast, and so a histogram very similar to Figure 2.

Garden of the Gods, Colorado. This image has fairly high contrast, and so a histogram very similar to Figure 5.

When the curve of your histogram climbs up the edges, that means you are not properly exposing parts of the image.  You’re losing data and will wind up with either blocked (too dark) or blown-out (too bright) areas (see Figure 2 above).  The higher up the sides the histogram goes, the bigger are those areas, and the more likely they are to negatively impact the final image.  Of course if you want to do this then it’s probably fine.  And if something like the sun is overexposed, then it is more than fine, it’s natural.

A camera with good dynamic range helps with this problem of shadow or highlight clipping.  It allows you greater latitude to bring out details in very bright or very dark parts of your images.  It allows you to capture images with a great variety of overall contrast levels.

Pink bleeding hearts bloom in the forests of the Pacific Northwest during spring.

Pink bleeding hearts bloom in the forests of the Pacific Northwest during spring.

A Caveat

I would like to throw in a big caveat to this benefit.  Don’t think you always need to even out tones HDR-style.  How much brightening of shadows and darkening of highlights you do on the computer is, after all, up to you.  You can even go the opposite direction by letting things go to pure black.  Or you can purposefully allow bright areas to be blown out.

It’s probably best to not think so much about how much contrast you can squeeze from the scene without blowing out highlights or clipping shadows.  Instead, think about the mood of the scene, the emotions you wish to elicit, the story you want to tell.  Let that be your guide, not the ability of your camera to manipulate a wide range of tones.  It’s yet another case of “just because you can doesn’t mean you should”.  Okay, end of rant, and end of post!  See you next Friday, when I’ll wrap up dynamic range.

This shot of Oregon's Mount Hood was framed so as to show some of the surrounding forest.

This shot of Oregon’s Mount Hood was framed so as to show some of the surrounding forest.

Advertisements

7 responses to “Friday Foto Talk: Dynamic Range III – The Histogram

Subscribe to comments with RSS.

  1. I love your tutorials! I was just beginning to learn to read histograms so this post was timely. Thanks!

  2. This has been a great series. Thank you so much.

  3. Very helpful. I’ve been trying to use the histogram as a guide bur didn’t really understand it. I shall view it with renewed interest and understanding now!

  4. These tutorials of yours are really helpful for me. I am just starting to use- or at least look at- the histogram and so this came at a very opportune time. Thank you!

Please don't be shy; your words are what makes my day!

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: