Understanding and Controlling Color Contrast
In Chapter 3, we saw how luma contrast contributes to the punchiness, sharpness, and overall appeal of an image. Within the domain of the chroma component, color contrast plays a similar role, though perhaps a more nebulous one, in shaping what the audience sees and how it responds to the various subjects within a scene.
Simply put, color contrast is the amount of differentiation among the various colors found within an image. The more varied the colors of different elements of the scene are from one another, the more color contrast there is.
When there is too little color contrast, the result can appear monochromatic, as if there’s a tint washed over the entire scene. If there’s a high degree of color contrast, then the variously colored elements of the scene will likely pop out at the audience with greater distinction.
The significance of color contrast is also supported by research into the contribution of color to object segmentation—the separation of specific subjects from textures found in the background and surround of a scene. To quote from “The Contributions of Color to Recognition Memory for Natural Scenes,” “A possible evolutionary advantage for color over luminance-based vision may lie, however, in superior segmentation of objects from textured backgrounds.” One would imagine that there’d be a significant advantage to being able to spot a colorful ripe fruit (or a dangerous predator) among the dense foliage of the forest or jungle where one might have lived 150,000 years ago.
One last note: I often run into situations where clients ask me to raise the saturation even though I feel things are probably as saturated as they ought to be. In these situations, what usually fixes the problem is finding a way to increase the color contrast in the image by selectively boosting the saturation of a specific hue, rather than increasing the saturation of the entire image. The following section covers a variety of different strategies for doing just this.
Types of Color Contrast
From a creative perspective, Johannes Itten, in his landmark The Art of Color (John Wiley & Sons, 1961), identified several types of color contrast that are useful to us as colorists. I highly recommend Itten’s book, which focuses on fine-art examples from numerous old masters. However, the following sections summarize the primary categories of color contrast Itten describes with examples that I’ve adapted to the problems and issues I encounter in real-world scenes in film and video.
It’s worth reflecting that all of these color contrast effects are effective because of the opponent model of the human visual system described at the beginning of this chapter. Every color in an image is evaluated relative to the other colors that surround it. Whether you’re adjusting the entire scene at once with a primary color correction or a specific element within the scene via a secondary color correction, these principles will help you to understand why certain seemingly small adjustments can “bring the shot alive.”
In the following sections, you’ll also see how you can use the vectorscope to evaluate color contrast of different kinds.
Contrast of Hue
This is the most fundamental type of color contrast you can control. The tough thing about hue contrast is that its successful appearance requires deliberate color decisions and placement by both the wardrobe and production design departments.
In the following example, the art direction and lighting were deliberately monochromatic, to create the atmosphere of a dark, lush nightclub. The result is highly saturated, but it has low color contrast, since all the hues are within a very narrow range, as seen in the vectorscope (Figure 4.136).
Figure 4.136 An image with high saturation but low color contrast.
In the next example, you’ll see an image that’s actually a bit less saturated than the previous example, but it displays considerably greater color contrast, as seen by the multiple arms of the vectorscope graph extending in multiple directions from the center (Figure 4.137).
Figure 4.137 An image with wide color contrast, showing a variety of hues.
I’m fond of saying during a session that if a range of different colors isn’t in an image to begin with, I can’t really put them there. That said, there’s often ample opportunity to tease faint colors that aren’t immediately obvious out of an otherwise dull-looking shot by doing some, or all, of the following:
- Eliminating an excessive color cast from the image to center the shadows and highlights within the vectorscope and redistribute the various other color spikes of the vectorscope graph about this center at as many different angles as possible.
- Turning up the saturation of an otherwise desaturated image, pushing all of the hues within the vectorscope farther out toward the edges, increasing the distance between each different hue cluster and increasing the color contrast.
- Increasing saturation selectively, bringing as many different colors out of the background as possible, while perhaps slightly desaturating the dominant hue. This requires secondary color correction operations covered in subsequent chapters.
Figure 4.138 has a slight orange color cast (or is it the sunset?), as well as low saturation overall, which gives the impression of low color contrast.
Figure 4.138 Low color contrast as a result of a color cast and reduced saturation.
Taking a different tack with this shot to increase the color contrast, you can neutralize the color cast, turn up the saturation, and use a hue curve operation to tease some of the already existent reds and oranges out of the newly revealed greens of the forest in the background. The same hue curve operation lets you tease more blue out of the sky and the water reflections, all of which gives you a lot of color contrast and a more polychromatic image (Figure 4.139).
Figure 4.139 Expanding color contrast by neutralizing the color cast and selectively increasing saturation to boost a variety of different hues.
Looking at the vectorscope, you can see that the graph in Figure 4.139 has become more centered, has stretched out further, and extends in more directions.
Another type of color contrast is the narrower combination of warm and cool tones, as opposed to a hodgepodge of different hues. Cold-warm contrast is subtle, realistic as far as naturally occurring color temperature is concerned, and it frequently occurs naturally in shots utilizing mixed-light sources. Of course, it doesn’t hurt if the art directors went out of their way to keep the production colors cool and warm to reinforce the lighting scheme.
In particular, cold-warm contrast tends to be expressed as the interplay between the warm hues of human skin tone and background lighting or art direction choices.
In Figure 4.140, the interior of the van is deliberately bluish, with the result that it plays against the complexion of the actor to create this kind of contrast.
Figure 4.140 Cold-warm contrast as a result of art direction.
If you want to add cold-warm contrast to an image that’s otherwise lacking it, you can try the strategy of making opposing adjustments to the highlights and midtones of an image using the color balance controls, adding warmth to the highlights, and cooling off the darker midtones and shadows (Figure 4.141).
Figure 4.141 Cold-warm contrast as a result of lighting.
Notice how the warmth of the woman’s highlights makes her stand out from the bluish background. Meanwhile, the man’s face is still getting its color from the background illuminant, so he blends in a bit more with the background.
It’s well known among painters and designers that placing highly complementary colors adjacent to one another results in a high-energy interplay between the two hues.
This type of color contrast is a much more aggressive choice and may require a correction to eliminate it if the effect is too distracting. On the other hand, you may go out of your way to create this particular color combination if it’s to the benefit of the scene; it could go either way.
In Figure 4.142, the baby blue/cyan of the woman’s sweater is in almost perfect complement to the beige/tans of the surrounding couch (with a little help from a hue curve operation). The result is that the sweater adds significant color contrast to the image, even though it’s not really that saturated.
Figure 4.142 Complementary contrast: The blue sweater is complementary to the beige/warm couch surrounding it.
You know you’re dealing with complementary contrast when there are two distinct arms of the vectorscope graph that stretch out in almost opposite directions, as shown in Figure 4.142.
In Joseph Krakora’s documentary Vermeer: Master of Light (Microcinema, 2009), the painter’s rendering of blue fabric with yellow highlights is observed to have this effect. Similarly, Vermeer’s use of colors for shadows that are complementary to those of the subject casting them has a similar, if more subtle, effect.
In the following example, the yellows in the glass of liquor have been emphasized to play off the man’s blue shirt, adding visual interest to the shot (Figure 4.143).
Figure 4.143 Complementary contrast. The yellow highlights of the drink contrast with the blue of the shirt.
Unfortunately, if excess complementary contrast is in a shot by mistake, the result can be distracting, so measures may be needed to reduce the amount of contrast by either selectively shifting the hue or reducing the saturation, of the offending subject, typically via a secondary color correction of some kind.
In Figure 4.144, the orange basketball in the background just happens to be the complement of the blue-painted wall. The result is that it sticks out like a sore thumb.
Figure 4.144 Unwanted complementary contrast: The orange basketball calls attention to itself when surrounded by the blue wall.
This is an example of a detail you’d probably want to suppress using a secondary color correction of some kind, either using a hue versus saturation curve to reduce the oranges in the scene, an HSL Qualification to key the basketball to create a matte with which to make the adjustment, or a mask/power-window shape to isolate the basketball if it’s too close to the man’s skin tone to key cleanly.
This type of contrast refers to the effect that a dominant surround or background color has on an interior subject. More often than not, simultaneous contrast is the source of problems, rather than the solution.
It’s a difficult phenomenon to show in real-world shots, but the following example should illustrate. In the following three images (Figure 4.145), the shot has been modified so that the wall hue is different. Look back and forth among all of the shots. You may notice that the background color subtly affects the look of the woman.
Figure 4.145 Our perception of the woman’s face is subtly altered by the color that surrounds her.
Now consider that the hue of the woman’s face is identical in each of these three shots, and you can see how simultaneous contrast can often work against you when you’re trying to match two shots with different backgrounds together. If the reverse angle in a scene happens to have a dominant color in the background when the primary angle doesn’t, a face in the foreground may not seem to match the face in the previous shot, even though they may actually match quite closely!
In these cases, you may actually obtain a better perceptual result by making an adjustment to correct for the illusion of a color cast, as if it were real (and in fact it is, to your eye). This is a great example of when numeric accuracy isn’t as important as the perception of accuracy.
Contrast of Saturation
Even within a more or less monochromatic shot (an abundance of earth tones, for example), if nothing else, ideally you can at least extract some contrast between highly saturated and less saturated subjects. This can be a good strategy for helping differentiate a foreground subject from a background when they’re otherwise muddled together because of a similarity of hue.
In Figure 4.146, the vectorscope shows that the hue of the man and the wall are virtually identical. However, the man stands out from the background not just because his complexion is darker but because he’s actually a bit less saturated than the wall. This contributes to the distinction between the foreground and background elements.
Figure 4.146 Even though the man’s face is the same hue as the wall paneling, the differences in saturation and lightness make him stand out.
Figure 4.147 is similarly monochromatic, with a warm color tone across the entire frame. The vectorscope graph appears to be an indistinct blob of oranges.
Figure 4.147 In this image with low saturation contrast, the illuminant of the scene and woman are in the same range of hue.
To create a bit of distance between the woman in the foreground and the background, you can make a secondary correction to slightly desaturate the background (but not totally, you want to retain the warmth of the overall lighting), while increasing the saturation of the woman’s face (just a bit, you don’t want her to look like she’s got a spray-on tan). You can see the results in Figure 4.148.
Figure 4.148 Reducing saturation in the surrounding scene and increasing saturation in the woman’s face increases saturation contrast in this scene, bringing the woman more to the foreground.
Just to make a point, no hues were altered in Figure 4.148. The resulting change is subtle but helps the woman in the foreground stand out a bit more, which adds some depth to the image and focuses viewer attention on her, both of which are distinct improvements. Also, notice how the vectorscope graph in Figure 4.148 has changed from an indistinct blob to a more defined shape pointing in two distinct directions: reddish/orange and a warm yellow.
Contrast of Extension
This final aspect of color contrast can be a lifesaver when you have only a little bit of differentiating color in an otherwise monochromatic scene. For example, Figure 4.149 is awash in tans, browns, beiges, and orange, with warm lighting throughout. The one thing that keeps this image from being chromatically flat is the reflections of the vivid green lampshades.
Figure 4.149 The green lamp reflections, though small, add color and interest to an otherwise monochromatic scene.
It’s just a small dash of green, but the fact that it’s completely distinct from the general range of hues in the rest of the environment means that little bit of green matters a lot. This is the principle of contrast of extension that matters to us as colorists.
Notice that the difference between contrast of extension and complementary contrast is that contrast of extension can utilize hues that more closely neighbor the dominant hues within a scene, when viewed on a color wheel. Also, contrast of extension relies on increased saturation to allow the smaller feature to play off of the larger scene. Simply having an element of a suitably different hue isn’t enough; it needs to be vivid enough to catch the viewer’s eye.
With this in mind, there are often seemingly monochromatic scenes where, at second glance, you find that you’re able to pull a little bit of color out of someone’s shirt, or tie, or a bowl of fruit on a table. Anything that happens to have a bit of color that you can stretch out, when viewed in the vectorscope, can provide a way for you to introduce a bit more color contrast to an image that otherwise may seem flat.
The following shot is suffused with warm lighting that plays off the wood doors and the beige medical wallpaper. It’s saturated, but there’s not much pop (Figure 4.150).
Figure 4.150 Another scene with low color contrast.
By selectively boosting the color of the man’s tie as much as we can, this little bit of shimmering blue brings just enough additional color to the image to make it look more interesting. It also serves to call a little more attention to the man in the foreground (Figure 4.151).
Figure 4.151 Boosting the color of the tie, even though it’s a small element, creates contrast of extension because it stands out from the dominant hues in the scene.
Adjustments to extend the contrast of a specific element may be as simple as raising overall saturation. Other times, it’s possible to tease a little more color out of an element in the frame by using a hue versus saturation curve or by using HSL Qualification to isolate and saturate a specific feature within the image.
Whatever method you use, remember that a little splash of color can go a long way.