Photoshop.CS.Bible [Electronic resources] نسخه متنی

Heightening Focus and Contrast

If you’ve experimented at all with Photoshop, you’ve no doubt had your way with many of the commands in the Filter Sharpen submenu. By increasing the contrast between neighboring pixels, the sharpening filters enable you to compensate for image elements that were photographed or scanned slightly out of focus.

The Sharpen, Sharpen More, and Sharpen Edges commands are easy to use and immediate in their effect. However, you can achieve better results and widen your range of sharpening options if you learn how to use the Unsharp Mask and High Pass commands, which I discuss at length in the following pages.

Using the Unsharp Mask filter

The first thing you need to know about the Unsharp Mask filter is that it has a weird name. The filter has nothing to do with unsharpening — whatever that is — nor is it tied into Photoshop’s masking capabilities. Unsharp Mask is named after a traditional film-compositing technique (which is also oddly named) that highlights the edges in an image by combining a blurred film negative with the original film positive.

That’s all well and good, but the fact is most Photoshop artists have never touched a stat camera (an expensive piece of machinery, roughly twice the size of a washing machine, used by image editors of the late Jurassic, pre-Photoshop epoch). Even folks like me who used to operate stat cameras professionally never had the time to delve into the world of unsharp masking. In addition — and much to the filter’s credit — Unsharp Mask goes beyond traditional camera techniques.

To understand Unsharp Mask — or Photoshop’s other sharpening filters, for that matter — you first need to understand some basic terminology. When you apply one of the sharpening filters, Photoshop increases the contrast between neighboring pixels. The effect is similar to what you see when you adjust a camera to bring a scene into sharper focus.

Two of Photoshop’s sharpening filters, Sharpen and Sharpen More, affect whatever area of your image is selected. The Sharpen Edges filter, however, performs its sharpening operations only on the edges in the image — those areas that feature the highest amount of contrast.

Unsharp Mask gives you both sharpening options. It can sharpen only the edges in an image or it can sharpen any portion of an image according to your exact specifications, whether it finds an edge or not. It fulfills the same purposes as the Sharpen, Sharpen Edges, and Sharpen More commands, but it’s much more versatile. Simply put, the Unsharp Mask tool is the only sharpening filter you’ll ever need.

When you choose Filter Sharpen Unsharp Mask, Photoshop displays the Unsharp Mask dialog box, shown in Figure 10-7, which offers the following options:

Figure 10-7: Despite any conclusions you may glean from its bizarre name, the Unsharp Mask filter sharpens images according to your specifications in this dialog box.

Amount: Enter a value between 1 and 500 percent to specify the degree to which you want to sharpen the selected image. Higher values produce more pronounced effects.

Radius: This option determines the thickness of the sharpened edge. Low values produce crisp edges. High values produce thicker edges with more contrast throughout the image.

Threshold: Enter a value between 0 and 255 to control how Photoshop recognizes edges in an image. The value indicates the numerical difference between the brightness values of two neighboring pixels that must occur if Photoshop is to sharpen those pixels. A low value sharpens lots of pixels; a high value excludes most pixels from the running.

The preview options offered by the Unsharp Mask dialog box are absolutely essential visual aids that you’re likely to find tremendously useful throughout your Photoshop career. Just the same, you’ll be better prepared to experiment with the Amount, Radius, and Threshold options and less surprised by the results if you read the following sections, which explain these options in detail and demonstrate the effects of each.

Specifying the amount of sharpening

If Amount were the only Unsharp Mask option, no one would have any problems understanding this filter. If you want to sharpen an image ever so slightly, enter a low percentage value. Values between 25 and 50 percent are ideal for producing subtle effects. If you want to sharpen an image beyond the point of good taste, enter a value somewhere in the 300- to 500-percent range. And if you’re looking for moderate sharpening, try out some value between 50 and 300 percent. Figure 10-8 shows the results of applying different Amount values while leaving the Radius and Threshold values at their default settings of 1.0 and 0, respectively.

Figure 10-8: The results of sharpening an image with the Unsharp Mask filter using eight different Amount values. The Radius and Threshold values used for all images were 1.0 and 0, respectively (the default settings).

If you’re not sure how much you want to sharpen an image, try out a small value in the 25- to 50-percent range. Then reapply that setting repeatedly by pressing Ctrl+F (z -F on the Mac). As you can see in Figure 10-9, repeatedly applying the filter at a low setting produces a nearly identical result to applying the filter once at a higher setting. For example, you can achieve the effect shown in the middle image in the figure by applying the Unsharp Mask filter three times at 50 percent or once at 250 percent. I created the top-row results in Figure 10-9 using a constant Radius value of 1.0. In the second row, I lowered the Radius progressively from 1.0 (left) to 0.8 (middle) to 0.6 (right).

Figure 10-9: Repeatedly applying the Unsharp Mask filter at 50 percent (top row) is nearly equivalent on a pixel-by-pixel basis to applying the filter once at higher settings (bottom row).

The benefit of using small values is that they enable you to experiment with sharpening incrementally. As the figure demonstrates, you can add sharpening bit by bit to increase the focus of an image. You can’t, however, reduce sharpening incrementally if you apply too high a value; you must press Ctrl+Z (Win) or z -Z (Mac) and start again.

Just for fun, Color Plate 10-1 shows the results of applying the Unsharp Mask filter to each of the color channels in an RGB image independently as well as in pairs. In each case, I maxed out the Amount value to 500 percent and set the Radius and Threshold to 4.0 and 0, respectively. The top row shows the original image on the left and Unsharp Mask applied to all three channels at once on the right. In the remaining rows, I alternated single-channel applications with applications to pairs of channels. You can see how the filter creates a crisp halo of color, especially around the outside edge of the lady’s face. Sharpening the red channel creates a red halo on the face and brings out blue-green details in the facial cream; sharpening the red and green channels together creates a yellow halo on the face and bluish details in the cream; and so on. Applying the filter to one or both of the red and green channels produced the most noticeable effects because these channels contain the lion’s share of the image detail. The blue channel contained the least detail — as is typical — so sharpening this channel produced the least dramatic results.

Setting the thickness of the edges

The Unsharp Mask filter works by identifying edges and increasing the contrast around those edges. The Radius value tells Photoshop how thick you want your edges. Large values produce thicker edges than small values.

The ideal Radius value depends on the resolution of your image and the quality of its edges:

When creating screen images — such as Web graphics — use a very low Radius value, such as 0.5. This results in terrific hairline edges that look so crisp, you’ll think you washed your bifocals.

If a low Radius value brings out weird little imperfections — such as grain, scan lines, or JPEG compression artifacts — raise the value to 1.0 or higher. If that doesn’t help, don’t fret. I include two sure-fire image-fixing techniques later in this chapter, one designed to sharpen grainy old photos and another that accommodates compressed images.

When printing an image at a moderate resolution — anywhere from 120 to 180 ppi — use a Radius value of 1.0. The edges will look a little thick on screen, but they’ll print fine.

For high-resolution images — around 300 ppi — try a Radius of 2.0. Because Photoshop prints more pixels per inch, the edges have to be thicker to remain nice and visible.

You can of course enter higher Radius values — as high as 250, in fact. Higher values produce heightened contrast effects, almost as if the image had been photocopied too many times, generally useful for producing special effects.

But don’t take my word for it; you be the judge. Figure 10-10 demonstrates the results of specific Radius values. In each case, the Amount and Threshold values remain constant at 100 percent and 0, respectively.

Figure 10-10: The results of applying eight different Radius values, ranging from precise edges to very gooey.

Figure 10-11 shows the results of combining different Amount and Radius values. You can see that a large Amount value helps to offset the softening of a high Radius value. For example, when the Amount is set to 200 percent, as in the first row, the Radius value appears to mainly enhance contrast when raised from 0.5 to 2.0. However, when the Amount value is lowered to 50 percent, the higher Radius value does more to distribute the effect than to boost contrast.

Figure 10-11: The effects of combining different Amount and Radius settings. The Threshold value for each image was set to 0, the default setting.

Recognizing edges

By default, the Unsharp Mask filter sharpens every pixel in a selection. However, you can instruct the filter to sharpen only the edges in an image by raising the Threshold value from 0 to some other number. The Threshold value represents the difference between two neighboring pixels — as measured in brightness levels — that must occur for Photoshop to recognize them as an edge.

Suppose that the brightness values of neighboring pixels are 10 and 20. If you set the Threshold value to 5, Photoshop reads both pixels, notes that the difference between their brightness values is more than 5, and treats them as an edge. If you set the Threshold value to 20, however, Photoshop passes them by. A low Threshold value, therefore, causes the Unsharp Mask filter to affect a high number of pixels, and vice versa.

In the top row of images in Figure 10-12, the high Threshold values result in tiny slivers of sharpness that outline only the most substantial edges in the woman’s face. As I lower the Threshold value incrementally in the second and third rows, the sharpening effect takes over more and more of the face, ultimately sharpening all details uniformly in the lower-right example.

Figure 10-12: The results of applying nine different Threshold values. To best show off the differences between each image, I set the Amount and Radius values to 500 percent and 2.0, respectively.

Using the preset sharpening filters

So how do the Sharpen, Sharpen Edges, and Sharpen More commands compare with the Unsharp Mask filter? First of all, none of the preset commands permit you to vary the thickness of your edges, a function provided by Unsharp Mask’s Radius option. Second, only the Sharpen Edges command can recognize high-contrast areas in an image. And third, all three commands are set in stone — you can’t adjust their effects in any way (except, of course, to fade the filter after the fact). Figure 10-13 shows the effect of each preset command and the nearly equivalent effect created with the Unsharp Mask filter.

Figure 10-13: The effects of the three preset sharpening filters (top row) compared with the Unsharp Mask equivalents (bottom row). Unsharp Mask values are listed in the following order: Amount, Radius, Threshold.

Sharpening grainy photographs

Having completed my neutral discussion of Unsharp Mask, king of the Sharpen filters, I hasten to interject a little bit of commentary, along with a helpful solution to a common sharpening problem.

First, the commentary: Although Amount and Radius are the kinds of superior options that will serve you well throughout the foreseeable future, I urge young and old to observe Threshold with the utmost scorn and rancor. The idea is fine — we can all agree that you need some way to draw a dividing line between those pixels that you want to sharpen and those that you want to leave unchanged. But the Threshold setting is nothing more than a glorified on/off switch that results in harsh transitions between sharpened and unsharpened pixels.

Consider the picture of Frederick Douglass in Figure 10-14. Like so many vintage photographs, this particular image of the famed abolitionist is a little softer than we’re used to seeing these days. But if I apply a heaping helping of Unsharp

Figure 10-14: The original photograph is a bit soft (left), a condition I can remedy with Unsharp Mask. Leaving the Threshold value set to 0 brings out the film grain (middle), but raising the value results in equally unattractive artifacts (right).

Mask — as in the second example in the figure — I bring out as much film grain as image detail. The official Photoshop solution is to raise the Threshold value, but the option’s intrinsic harshness results in a pockmarked effect, as shown on the right. Photoshop has simply replaced one kind of grain with another.

These abrupt transitions are quite out of keeping with Photoshop’s normal approach. Paintbrushes have antialiased edges, selections can be feathered, the Color Range command offers Fuzziness — in short, everything mimics the softness found in real life. Yet right here, in what is indisputably Photoshop’s most essential filter, we find no mechanism for softness whatsoever.

While we wait for Photoshop to give us a better Threshold — one with a Fuzziness slider or similar control — you can create a better Threshold using a very simple masking technique. Using a few filters that I explore at greater length throughout this chapter and the next, you can devise a selection outline that traces the essential edges in the image — complete with fuzzy transitions — and leaves the non- edges unmolested. So get out your favorite old vintage photograph and follow along with these steps.

STEPS: Creating and Using an Edge Mask

Duplicate one of the color channels. Bring up the Channels palette and drag one of the color channels onto the little page icon. Mr. Douglass is a grayscale image, so I duplicate the one and only channel.

Choose Filter Stylize Find Edges. As I explain in Chapter 11, the Find Edges filter automatically traces the edges of your image with thick, gooey outlines that are ideal for creating edge masks.

Press Ctrl+I (z -I on the Mac). Or choose Image Adjustments Invert. Find Edges produces black lines against a white background, but to select your edges, you need white lines against a black background. The Invert command reverses the lights and darks in the mask, as in the first example in Figure 10-15.

Figure 10-15: I copy a channel, find the edges, and invert (left). I then apply a string of filters to expand and soften the edges (middle). After converting the mask to a selection outline, I reapply Unsharp Mask with winning results (right).

Choose Filter Other Maximum. The next step is to thicken the edges. The Maximum filter expands the white areas in the image, serving much the same function in a mask as Select Modify Expand serves when editing a selection outline. Enter a Radius value and press Enter or Return. In my case, a Radius of 4 pixels worked nicely, but for best results you should experiment with different values based on the resolution of your image.

Choose Filter Noise Median. You need fat, gooey edges, and the current ones are a bit tenuous. To firm up the edges, choose the Median filter, enter the same Radius value you did for the Maximum filter, 4 in my case, and press Enter or Return.

Choose Filter Blur Gaussian Blur. Unfortunately, the Maximum filter results in a bunch of little squares that don’t do much for our cause. You can merge the squares into a seamless line by choosing the Gaussian Blur command and entering the same radius you entered for Maximum (4 in my case). Then press Enter or Return.

The completed mask is pictured in the second example of Figure 10-15. Although hardly an impressive sight to the uninitiated eye, you’re looking at the perfect edge mask — soft, natural, and extremely accurate.

Return to the standard composite view. In a color image, press Ctrl+tilde (Win) or z -tilde (Mac). In a grayscale image, press Ctrl+1 (Win) or z -1 (Mac).

Convert the mask to a selection outline. Ctrl-click (Win) or z -click (Mac) the mask name in the Channels palette. Photoshop selects the most essential edges in the image without selecting the grain.

Choose Filter Sharpen Unsharp Mask. In the last example in Figure 10-15, I applied the highest permitted Amount value, 500 percent, and a Radius of 2.0.

Whatever values you use, make sure the Threshold is set to 0. And always leave it at 0 from this day forward.

In case Figures 10-14 and 10-15 are a little too subtle, I include enlarged views of the great abolitionist’s face in Figure 10-16. The top image shows the result of using the Threshold value; the bottom image was created using the edge mask. Which one appears sharper and less grainy to you?

Figure 10-16: Enlarged views of the last examples from Figures 10-14 (top) and 10-15 (bottom). A good edge mask beats the Threshold value every time.

Using the High Pass filter

The High Pass filter falls more or less in the same camp as the sharpening filters but is not located under the Filter Sharpen submenu. This frequently overlooked gem enables you to isolate high-contrast image areas from their low-contrast counterparts.

When you choose Filter Other High Pass, Photoshop offers a single option: the familiar Radius value, which can vary from 0.1 to 250.0. As demonstrated in Figure 10-17, high Radius values distinguish areas of high and low contrast only slightly. Low values change all high-contrast areas to dark gray and low-contrast areas to a slightly lighter gray. A value of 0.1 (not shown) changes all pixels in an image to a single gray value and is therefore useless.

Figure 10-17: The results of separating high- and low-contrast areas in an image with the High Pass filter set at nine different Radius values.

Converting an image into a line drawing

The High Pass filter is especially useful as a precursor to Image AdjustmentsThreshold (covered in Chapter 17), which converts all pixels in an image to black and white. As illustrated in Figure 10-18, the Threshold command produces entirely different effects on images before and after you alter them with the High Pass filter. In fact, applying the High Pass filter with a low Radius value and then issuing the Threshold command converts your image into a line drawing.

Figure 10-18: Several applications of the High Pass filter with low Radius values (top row), followed by the same images subject to Image Adjustments Threshold and Filter Blur Gaussian Blur (middle). I then layered the second row onto the first and modified the Opacity and blend mode settings to create the third row.

In the second row of examples in the figure, I followed Threshold with FilterBlur Gaussian Blur (the subject of the next section). I set the Gaussian Blur Radius value to 1.0. Like the Threshold option in the Unsharp Mask dialog box, the Threshold command results in harsh transitions; Gaussian Blur softens them to produce a more natural effect.

Why change your image to a bunch of slightly different gray values and then apply a command such as Threshold? One reason is to create a mask, as discussed at length in the “Building a Mask from an Image” section of Chapter 9. (In Chapter 9, I used Levels instead of Threshold, but both are variations on the same theme.)

You might also want to bolster the edges in an image. For example, to achieve the last row of examples in Figure 10-18, I layered the images before applying High Pass, Threshold, and Gaussian Blur. Then I monkeyed around with the Opacity setting and the blend mode to achieve an edge-tracing effect.