Color Correction Made Easier

By Ctein Back to

ctein1

I originally planned to give this article the admittedly glib title, “Color Correction Made Easy.” Upon modest reflection, I realized that title would be extremely misleading. Color correction isn’t easy. It’s never going to be easy.

Let’s face it, if good color correction (and, correspondingly, good tone correction) were easy, everyone would be a great photographic printer. There would be no need for custom labs, professional printers, and the myriad craftspeople out there whose business is converting decent photographs into more-than-decent prints. What I’m trying to say is that if you find color correction an occasionally frustrating business, that does not reflect ill upon you. It only makes you part of a very large community that really cares about the quality of its work and is always looking to improve.

Good color correction requires a minimum of a good eye and an understanding of basic tools, but sometimes cleverness is demanded. None of us ever stops improving our craft. Certainly not me. There’s one trick in this article that I only came up with in the last couple of years, solving a problem that had been vexing me for a lot longer than that.

These Curves turned Figure 1, top, into Figure 1, bottom. They were created “automatically” using the three eyedropper tools on the left side of the control panel.

Fine color printers have always relied upon tips, tricks, and tools. In the darkroom, we had ring-arounds, print viewing filters, reflection densitometers, and video analyzers, ordered by increasing cost and rarity. Few “home” printers owned reflection densitometers; hardly any owned video analyzers. One of the great things about digital printing is how level the playing field has become; any semi-serious hobbyist can afford a photographic-quality printer and a computer and software sufficient to the task.

Speaking of software… I rely heavily on mondo-expensive Photoshop, but this is my stock in trade, and I need a certain amount of insane capability. You don’t need Photoshop to do color correction. Although this article is Photoshop- oriented, you can do great color correction with free and low-cost programs like GIMP, Photoshop Elements, and my favorite, Picture Window Pro. The latter is a wonderful bargain for Windows users and has capabilities even Photoshop doesn’t. It’s fully 16-bit capable and has been for a long time, and it has the coolest color correction tool that I’ve ever encountered. This tool alone makes it worth the price of the program. More about that later.

Where to begin

For the purposes of this article, your source material doesn’t matter. You can be scanning color slides or negatives or even old color prints (I do that all the time in my restoration work). If you’re starting with original digital photographs, you can color correct JPEG, TIFF, and RAW files. Still, some things will make your tasks a lot easier and the results better.

If you have any choice in the mat- ter, start with 16-bit-per-channel files (a.k.a.,”16-bit color” or “48-bit color”). When starting with a digital photograph, that means RAW or 16- bit TIFF files. If you’re scanning, choose the appropriate setting in your scanner software.

If at all possible you want to be working in at least Adobe RGB color space; ProPhoto RGB color space would be preferable. Your color film and your digital camera can capture tones and colors that Adobe color space can’t handle (just forget about sRGB). The out-of-gamut pixels will come up as pure white or pure black in one color channel or more; it will be very difficult to do good color correction on those extreme colors, and they are often the ones that are in most need of help. Note that getting satisfactory results in ProPhoto RGB color space requires using 16-bit color; trying to work in such a large color space with 8-bit data is a guarantee of banding and contours.

You may not have any choice in these matters; your photograph may be an 8-bit JPEG in sRGB color space, in which case you make the best of an awkward situation. I’m just saying that if you’ve got the opportunity, more bits and a bigger color space will make your life a lot easier.

First Principles

A photograph with good color usually has a rich range of values from near- white to near-black. Somewhere in that color print you’ll find a bit of deep shadow that approaches black and a bit of a highlight glint that approaches white. They aren’t necessarily large or important parts of the photograph, but they’re usually there.

This is nearly always true when you examine each color channel separately. The photograph may not have any true whites or true blacks, but somewhere in the photograph there will almost always be pixels that will have values near 0 or 255 for one of the individual colors. This idea may take some getting used to, so here’s an example. The purest, most saturated yellow will have a blue value of 0 and red and green values of 255. You don’t need a white or black pixel to get extreme values in the individual color channels of a color photograph; any pure color will have some color values near black or white.

One more thing, somewhere in the photograph there’s likely to be some small area that is pretty close to a neutral gray. Put this all together and it makes one really useful technique for balancing overall color. If the darkest points in your photographs are near neutral black, and the brightest points in your photograph are near neutral white, and that patch of neutral gray in the photograph really is neutral, the overall color balance will be good 90% of the time. How do you achieve this? With curves and eyedroppers. Here’s how that works (with just about any software out there, including most scanner and RAW conversion software).

The first thing to do is to get the eyedroppers set properly. Usually double-clicking the eyedropper icon brings up a control panel in which you can set the values for that eye- dropper. Software defaults are RGB values of 0,0,0 and 255, 255, 255 for the black and white eyedroppers respectively. We want to pull those in a bit, to make sure that we don’t accidentally clip the highlights or shadows. Set the black eyedropper RGB values at 15, 15, 15 and the white eyedropper RGB values at 240, 240, 240. That gives some headroom so that your pixel choices don’t have to be perfect. Some software will let you make these new values permanent defaults, so you don’t have to set them every time you launch the program.

Normally the middle gray eye- dropper works well at its default RGB value of 128, 128, 128. If the neu- tral patch in your photograph is far from middle gray, you should adjust that. For a dark gray, try RGB values of 70, 70, 70. For a light gray, RGB values of 170, 170, 170 are good.

If there is a radius control for the eyedropper, set it for three pixels for digital photos and five pixels for film scans. That prevents digital noise or film grain from throwing off your settings. This is especially important with film scans, where the darkest parts of the scan are likely to be very grainy.

Now you’re ready to go. Click the black eyedropper on the darkest area you can find in the uncorrected photograph (shown in Figure 1, top). Click the wide eyedropper on the lightest air you can find. Very often, the overall color balance will look pretty good with just those two adjustments. If the photograph still has a substantial overall color cast, click the middle gray eyedropper on a mid-toned part of the photographs that you want to come out neutral (Figure 2). Most of the time, this will give you good overall color balance (Figure 1, bottom). Not perfect by any means, but refining that is where the art, skill and your keen eye come in.

(Figure 3) Photoshop’s Variations tool creates a ring-around that will help you evaluate the type and magnitude of color error in your photograph. Unfortunately, Variations only works on 24-bit color files.
(Figure 4) The Selective Color adjustment allows you to individually vary the color components of each of the primaries, as well as the blacks, whites, and middle grades.You can use this to tweak narrow color ranges in a photograph. In this case, I converted Figure 1, into Figure 5. Knocking 39% cyan out of the reds made them brighter and richer. Subtracting 6% magenta and adding 70% yellow made the yellows more brilliant.
(Figure 5) The results of applying the Selective Color adjustments in Figure 4 to the bottom photograph in Figure 1. The red and gold trim in the woodwork is clearer and stands out better, and the candle glow looks more luminous.

Individual Color Adjustment

Many times you’ll get the over- all color balance roughly correct and still find that you’re unhappy with the way colors are rendered. Further Curves adjustments will correct most of those color errors, but learning how to do this is a matter of practice and skill; there aren’t any simple principles and tricks that I can offer. Other tools, though, do provide some quick and easy fixes.

Photoshop contains two such tools under Image|Adjustments: Variations and Selective Color. Variations has one huge handicap; it is limited to working on 8-bit color in Photoshop CS4. But if that’s the kind of file you happen to be working with, it’s a good tool to try when you feel like the photograph has some overall color casts and you’re not sure how to correct them. It’s a computerized combination of the classic darkroom ring-around and print viewing filters. Ring-arounds are especially valuable when you’re having trouble telling which direction the color needs to go.

Launching Variations brings up a control panel that shows the original photograph in direct comparison to lighter and darker versions and versions shifted towards each of the primary colors (Figure 3). A slider controls the degree of the shift, from extremely slight to gross, while selection buttons control whether you’re emphasizing a color shift in shadows, midtones, or highlights. Clicking the variation that looks best makes it the central choice, and a new ring-around is created from that. It’s a way to fiddle with the color balance of a photograph that many people find comfortable and intuitive to work with.

If you feel that the overall color balance is good but certain colors didn’t come out the way you’d like, Selective Color is one way to attack that problem. Its drop-down menu will let you independently adjust the color qualities of all six primaries, as well as the blacks, middle grays, and whites. For example, suppose you have a photograph of foliage where the greens don’t seem to have the intensity they ought to. Launch Selective Color and pick Greens from the drop-down menu. You’ll see sliders for Cyan, Magenta, Yellow, and Black. If the greens look desaturated, move the magenta slider in the minus percentage direction. Magenta is the complement of green, so reducing the amount of magenta in the greens will make them purer. If you want, you can shift the greens towards the yellow by increasing the percentage of yellow and decreasing the percentage of cyan. If the greens are too light or too dark, moving the black slider will correct that. I used the Selective Color settings shown in Figure 4 to create Figure 5.

By far, though, my favorite color correction tool isn’t a part of Photoshop. It’s the color correction tool that’s built into Picture Window Pro, available as a Photoshop plug-in called Color Mechanic Pro. This is extraordinarily powerful and intuitive and unlike any other color correction tool you’ve ever used. If you already own Photoshop and have no desire to add Picture Window Pro to your ensemble, purchase that plug-in. You won’t regret it.

Color Mechanic Pro takes advantage of the way we naturally think about color correction. We can look at a photo and immediately identify the colors and tones that are off, but we can’t simply tell Photoshop, “That skin tone is too pink and that gray too blue – fix it!” Color Mechanic Pro comes close to doing just that with a remarkably simple way to correct color. It presents you with a view of your file and a color-space hexagon. Click on a point in the image, and the corresponding color is selected within the color hexagon. You can drag that color into any other place in the color space. The color space warps smoothly around that change, as if it were a rubber sheet change, so related colors adjust to fit; greatly different colors aren’t affected at all (Figure 7). Changes are immediately reflected in the “after” image and the color hexagon. There are also sliders to control the brightness and the strength of the color changes.

You can do this to as many color points as you like, custom-tuning the color palette to fit the photo. You can lock down a color so that it doesn’t change by adding a correction point but not dragging it to a new location. That pins the color at its original value, no matter what other warps you make to the color space.

(Figure 6) Fluorescent light is never a good choice for illuminating a subject, but sometimes one doesn’t have much choice. The top figure shows the photograph as it came from the camera. The bottom photograph shows the degree of color correction I achieved using the Color Mechanic plug-in settings shown in Figure 7.
(Figure 7) Color Mechanic is an intuitive way to selectively adjust colors in a photograph and improve its overall color balance and rendition. The top illustration shows the control panel as it opens up. In the bottom illustration, I zoomed in on the color space to give me finer control, and I selected points in the snow, white feathers, and silver gray feathers. I dragged all those points close to the neutral center of the color space to eliminate the color casts in the photograph. Then I added a point in the greenish rock steps and dragged it towards a more neutral yellow, to make the rocks more naturally tan. The results appear in Figure 6, bottom.

 Fixing Skin Tones

The Hue/Saturation adjustment is a powerful and underappreciated tool for modifying ranges of colors in subtle ways that are difficult to achieve by other means such as Curves. For example, used in a masked adjust- ment layer, Hue/Saturation works amazingly well for smoothing out blotchy Caucasian skin tones.

Caucasian skin tones can be blotchy for many reasons: the age or health of the subject, the harsh effects of on-camera flash, the dubious quality of old photographs that are being restored. Regardless, blotchy skin tones look that way because they have some patches that are too pink or red and other patches that are too yellow. If one were to make the yellows a little ruddier, they’d look a lot more like a proper skin tone. Similarly, making reds and pinks a little yellower would sup- press that flushed look.

(Figure 8) Blotchy Caucasian skin tones occur when some parts of the skin are rendered too yellow and others too pink. In this photo, it’s most evident in the knuckles and fingertips, the nose and the cheeks. This blotchiness is quickly and easily removed using selective Hue/ Saturation controls as shown in Figure 9, resulting in the photograph on the right.
(Figure 9) The key to eliminating blotchy skin tones is to shift the reds in the Hue/Saturation adjustment towards the yellow (a plus Hue adjustment) and shift the yellows towards the red (a minus Hue adjustment). The shifts will often be comparable in size, but you can favor one over the other, depending on how pink or tan you want the skin to come out overall. In this case, I corrected the yellows more strongly and slightly darkened them to keep the lightness of the skin tones the same.

Create a Hue/Saturation adjustment layer and set the control sliders for the red and yellow channels as shown in Figure 9. Moving the Hue slider to the left or right shifts the whole spectrum. By moving it to the left for the yellow channel, we’ve shifted the yellow colors into the red and the reds into the pinks. For the red channel, moving the slider to the right shifts redder colors into the yellow range.

It doesn’t take a lot of hue shift to get the desired effect (Figure 8, right). I exaggerated the changes here so they’d be visible in reproduction. Be careful not to overdo this in practice, or you’ll wipe out most of the color variations in the face and it will look like the subject is wearing too much pancake makeup.

The reason for doing this in an adjustment layer is that usually you won’t want to change all the colors in the photograph, just the skin tones. So, the final step is to fill the mask channel with black and use a white brush to paint in the mask areas for the skin that you want to alter.

Product Resources: Software: Adobe Photoshop CS4, Adobe Photoshop Elements, Picture Window Pro and Color Mechanic Pro by Digital Light Color, GIMP by GNU Image Manipulation Program


About the Author

Ctein
Ctein
Ctein is a technical writer and expert printmaker. He is also the author of Digital Restoration and Post Exposure—Advanced Techniques for the Photographic Printer.