Essentials of Photoshop Color Correction: Highlights and Shadows
Peachpit Press is offering this book at a discount to creativepro.com readers. Follow this link.
If you take a consistent approach to the color-correction process, your images will invariably look better, and you'll be finished faster. Regardless of why an original image looks wrong, you can always improve it by establishing proper highlight and shadow values, ensuring midtone neutrals are really neutral, enhancing contrast in the area of interest, and applying sharpening in a manner appropriate to the image content and output conditions.
- Highlights and Shadows
The most basic rule of color correction is to take advantage of all available tonal range, and you can only do this if the image's lightest points are as light as possible and its darkest points are as dark as possible. That's why setting highlights and shadows is always the first step in the color correction process (see below).
If other corrections are applied first, such as neutralizing a midtone color cast, it's almost certain they will need to be adjusted after highlight and shadow are properly set.
- Neutral Tones
If the neutrals in an image are wrong, everything else is wrong. The highlight and shadow points corrected in the first step not only ensure proper reproduction of the image's lightest and darkest points-they also make these points neutral.
This removes any color casts that might be affecting these areas, but for proper color balance the same thing has to be done throughout the image's tonal range. This requires carefully measuring, assessing, and sometimes changing midtone neutral or almost neutral colors.
Images just don't look right unless they have snap-that intangible quality of dimension that separates a flat and lifeless picture from one that jumps off the page. Basic contrast comes from the difference between an image's lightest and darkest points, and is therefore determined by its highlight and shadow values.
But extra contrast can be engineered into any image by applying curves that are steepest, on a channel-by-channel basis, in the areas of interest. In some cases, contrast can be borrowed from another channel in the image, or from one of the channels created by converting a duplicate of the image into another color space.
It's not good to be dull. Before being printed (or used online) images must be sharpened to compensate for limitations of the digitizing and printing processes. The optimal settings for the unsharp mask filter always depend on image content, size, and output medium.
There is no one-size-fits-all procedure for sharpening images, which means that it often requires two stages: images are sharpened immediately after scanning or capture with a digital camera to produce a master image, then application-specific sharpening is applied just prior to use, once the final size and other output conditions have been determined.
Now let's see about setting highlight and shadow aimpoints.
Setting Highlight and Shadow Aimpoints
One of the main objectives of color correction is to take full advantage of the available tonal range, which is achieved by mapping the brightest part of the image that contains reproducible detail -- the highlight -- to the whitest white of the output medium, and mapping the darkest part of the image-the shadow-to solid black in the output medium. When the final output device is not known at the time an image is corrected, target values are used that will allow it to be repurposed for the widest variety of output conditions.
Photoshop's Info palette provides an easy way of measuring the original color values in key tonal areas, such as highlights and shadows, while the Curves tool provides an excellent way of converting the original color values to the ideal color values.
The question is: What are the ideal color values?
First, the answer depends on whether the target output device accepts RGB or CMYK data. A theoretically pure white would be 255R 255G 255B in the RGB model (formed by combining the red, green, and blue primaries at full strength), but 0C 0M 0Y 0K in the CMYK model (the absence of all four inks).
More practically, the answer depends on the color characteristics of the output device or output medium, regardless of whether it's a monitor, desktop printer, or printing press. For example, consider the problem of creating bright neutral highlights on different kinds of RGB-based devices, such as monitors, film recorders, and desktop inkjet printers (which invariably take RGB input, even though they print with four or more inks).
At first glance, it seems that to get the broadest overall tonal range, the highlight must be set to the brightest possible value, which in the RGB model is 255R 255G 255B. However, a simple test reveals that almost all output devices have a practical limit, with slightly lower RGB values such as 250R 250G 250B or 245R 245G 245B.
Here's how to determine the actual RGB highlight value for a particular device, such as an inkjet printer. In Photoshop (or other imaging, drawing, or page-layout application), create a one-page document consisting of 12 small squares without borders and containing 1 percent increments from 255R 255G 255B to 244R 244G 244B (see below for PDF files for highlights -- aimpoint.hl.pdf -- and shadows -- aimpoint.sh.pdf).
PDF downloads. To determine optimal highlight values for your RGB output device, print a test page containing patches filled with incremental values from 255R 255G 255B through 244R 244G 244B.
Note: These files are large. For best results, do not open them in your Web browser. Instead download them to your hard drive before opening them in Acrobat or Acrobat Reader. To get Acrobat Reader, click here.
The first patch containing visible tone indicates the lightest area in which your printer can maintain highlight detail. Note that depending on the particular paper stock and printing conditions used for this book, it may not be possible to see anything in many of the patches, especially those that are close to the white point.
Print this page and look closely to see which square contains a just barely visible gray tint. For instance, if the squares filled with RGB values 255 through 251 appear pure white, but the square containing 250R 250G 250B is just slightly gray, then it's pointless to assign tonal values in the image higher than 250, as these will simply disappear. That could be a big problem, as there is sometimes crucial image detail in the highlight areas.
The solution is to either measure the RGB highlight values appropriate to your specific output device or use a standard value that will produce excellent results with the vast majority of output media. For output to monitors and to color inkjet and laser printers, the optimal values are typically between 245 and 248 (with equal amounts of red, green, and blue, to ensure that neutral tones stay neutral). Some devices, such as high-end film recorders, can maintain detail up to and beyond 250, and your target highlight value should be adjusted accordingly.
Another key reason for using a highlight value below 255 is that it leaves room for specular highlights-those caused by glare from shiny objects or surfaces. The real highlight isn't necessarily the brightest area in an image; it's the brightest area that contains detail. Setting this highlight to a value of 245 or 250 makes it possible to reserve 255R 255G 255B for an area so white that the background paper shows through, such as for the reflections from glass, metal, water, or snow. For instance, leaving room for specular highlights has made it possible to maintain essential detail in the antique car image (see figure 1).
Figure 1: Instead of using a pure white value of 255R 255G 255B as a highlight aimpoint, the standard practice is to use a value closer to 245R 245G 245B (1a), both to compensate for limitations in the output medium and to allow for specular reflections. The setting of 255R 255G 255B (1b) has slight better contrast).
For instance, the image of the antique car was corrected for a standard 245R 245G 245B highlight (figure 1a) and retains detail in the headlamps and the grille.
The same image separated with a 255R 255G 255B highlight (figure 1b) has slightly greater contrast, but some subtle highlight details are no longer visible. The difference isn't huge, but it's significant, and can't be ignored if you're committed to producing the highest possible quality color.
Given a choice between using a standard value of 245R 245G 245B or 250R 250G 250B, it's probably better to use 245 to protect highlight detail. Also, these corrections are being applied to correct tonal defects, not to compensate for the properties of a specific output device, so it's best to use a conservative value that can later be modified automatically at print time for any output device for which a profile is available. Highlight values of 250 or higher can routinely be achieved by the high-end film recorders used to produce "second-generation originals" for rescanning.
The tradeoff in using a highlight aimpoint of 245 instead of 255 is that compressing the image's tonal range into fewer shades of gray slightly increases the possibility of banding, because there's now a bigger gray difference from one shade to another.
Despite this tradeoff, we'll use 245R 245G 245B as the standard RGB highlight aimpoint throughout the exercises in this book. However, if the preceding test showed that your particular output device can attain a brighter highlight value, by all means use that value to make full use of the tonal range of your device and media.
With the highlight aimpoints properly established, it's time to turn our attention to the other end of the brightness scale, where the optimal shadow aimpoints are also determined by the characteristics of the output device.
In theory, the darkest areas in an image should be solid black, which would mean RGB values of 0R 0G 0B. In practice, however, virtually all output devices are unable to accurately reproduce shadow tones, so that everything darker than, say, 90 percent gray appears as black.
As a result, we need to use a standard RGB shadow aimpoint that's slightly lighter than black, knowing that all of the shadow tones will appear noticeably darker when printed. For instance, for display on a monitor or output to a desktop inkjet or color laser printer, you'll obtain the best results by using an RGB shadow aimpoint between 5R 5G 5B and 15R 15G 15B.
The color separation engine built into Photoshop understands that detail in the shadow tones can plug up if allowed to print too dark and so automatically converts RGB files so that shadow aimpoints in this range produce dark but detailed shadow areas.
To more precisely find the optimal shadow aimpoints for a particular output device, repeat the test procedure described earlier for determining highlight aimpoints, but this time use patches containing RGB values between 0R 0G 0B and 20R 20G 20B. The objective is to find the smallest value that retains important detail in the shadow tones.
Throughout the exercises in this book, we'll use 10R 10G 10B as the RGB shadow aimpoint, though if your tests produce better results with a higher or lower value, feel free to use that instead.
PDF Download. To find the optimal shadow point for an RGB-based output device, print a test target (see the PDF download aimpoint.sh2.pdf) containing color patches ranging from 0R 0G 0B to 22R 22G 22B in 2-point increments. At first, it may appear that all of the patches are equally black, but on closer examination you will usually be able to discern some lighter patches, gradually building to a maximum of solid black.
Now that we've set the highlights and shadows, we can move onto curves in our next installment.
Liked This? Read These!
Excerpted from "Photoshop Color Correction" (Peachpit Press). Read More
Excerpted from "Photoshop Color Correction" (Peachpit Press). Read More
GretagMacbeth, a leading innovator in color control technologies, will highlight its suite of digital workflow color control products for textile design, dyeing, printing and finishing in Booth 6C-50... Read More
The $4,599 Xerox Phaser 7800 is a laser printer aimed at graphic design agencies and other companies in need of color-accurate output at high speeds. The 1200 x 2400 dpi printer comes with the... Read More