USM got us started on our sharpening endeavor, but the one pass approach, so often used by many photographers, left us with a number of sharpening problems. One of the biggest problems with using a one pass USM sharpening approach is that, beyond the amount, radius, and threshold controls, there is little flexibility to adjust the sharpening to the specific needs of a particular image. With the one pass USM approach, everything in the image gets clobbered with the same USM tool and settings, regardless whether some areas of the image need more or less sharpening than those settings. The first tool in this article, Smart Sharpen (available on Photoshop CS2), addresses this issue by varying the sharpening depending on the tonal values. The second sharpening method, High Pass sharpening, sharpens in a manner that allows the sharpening to be modified, even at a later date, or even completely undone.
Once again, we must deal with the same old caveat: the quality of your monitor. In this article, sharpening effects are shown that are sometimes subtle (these subtle effects are what often differentiate a superb sharpening job from a mediocre one). How well you can see these effects will depend on your monitor. For instance, I can easily see all of the effects on my desktop, but some of the effects are impossible to see on my laptop.
The Smart Sharpen tool is one option that adds more flexibility that can improve the sharpening process. Smart Sharpen is opened by selecting Filter/Sharpen/Smart Sharpen. Figure 1 shows the Smart Sharpen tool.
As can be seen in Figure 1, Smart Sharpen has amount and radius controls. These controls function similarly as they do with USM. A closer look, however, reveals options that do not exist with USM. These options lead to two main advantages of Smart Sharpen over USM. The first advantage has to do with the Remove option. This option allows the user to select one of three blurring options (i.e., how the second copy of the image is blurred as described in the section on How USM Works). The different remove options produce different sharpening effects, which lead to more choices for the photographer. The second advantage has to do with the Shadow and Highlight menus. These menus allow the photographer to set different amounts of sharpening for the mid-tone, highlight, and shadow areas. The last choice is the More Accurate option. This will increase the accuracy of the sharpening, but will slow down the computer.
Figure 2 shows the Remove option expanded. Three options are revealed. The Gaussian Blur option is the standard blur option used by USM. Selecting this option will produce results similar to USM. The Lens Blur option produces perceived sharpening effects with a smaller radius that USM. This could be of particular benefit with images that have fine detail. The motion Blur option has to do with improving the sharpness of images where there was blurring due to motion (e.g., camera movement). This option will not be dealt with in this article.
Figures 3 through 5 show an image and the effects of the Gaussian and Lens Blur options. Figure 3 shows the unsharpened image. This image is a section from a close-up shot of some very small vegetation that resulted in relatively fine detail in the image. A radius setting of 2.9 was chosen. Figure 4 shows the result of sharpening with the Gaussian Blur option. The image is clearly oversharpened. The result is a loss of image detail. The primary cause of the oversharpening is a radius that is too large for the fine detail when used with the Gaussian Blur option. Figure 5 shows the same image sharpened using the Lens Blur option and the same 2.9 radius setting. The fine detail of the vegetation is far superior in this image. Despite the fact that the Lens Blur option used the same radius as the Gaussian Blur option, it was able to better capture the fine detail.
The second advantage of Smart Sharpen is the ability to set different sharpening amounts for the mid-tones, highlights and shadows. Figure 6 shows an early morning shot of the bank of the Feather River in Northern California. This image has a good range of tonal values from light to dark. Smart Sharpen will allow the sharpening to be optimized for each tonal area.
Sharpening an image in Smart Sharpen starts out the same as in USM. First, the sharpening radius and amount should be set for the mid-tones by working in the Sharpen menu (see Figure 1). The radius should be set according to the image content and image detail (see Part II of this series for a detailed explanation). Once the correct radius has been determined, the sharpening amount should be adjusted.
Figures 7 and 8 show a crop from one of the mid-tone areas of the Feather River image. Figure 7 shows the unsharpened image. Figure 8 has had a radius of 1.4 and an amount of 175% applied. The image shows an appropriate sharpening for this mid tone area.
Now that the mid-tone areas have been dealt with, the highlights need to be addressed. The highlights need to be protected from oversharpening because sharpening can cause some of the highlights to blow out. For instance, an area of highlights that has slight detail may be blown to pure white by the sharpening. Attenuating the sharpening in the highlight areas will prevent this.
Figure 9 shows the Highlight menu. This menu has three controls. The Fade Amount determines how much the sharpening is reduced. The higher the setting, the more the sharpening is decreased. A setting of 0 will result in no reduction of sharpening (i.e., the highlights will be sharpened just like the mid-tones). A setting of 100% will result in no sharpening of the highlights. The Tonal Width determines which tonal values will be affected by the reduction in sharpness. Smaller values of Tonal Width result in only the lightest highlights being affected by the sharpening reduction. Larger values of Tonal Width will result in a wider range of the highlights being affected. The Radius setting determines how many pixels, around a particular pixel, will be evaluated to determine to which tonal area that pixel belongs (i.e., is that particular pixel a highlight, mid-tone, or shadow).
Finally, the shadows need to be addressed. Personally, I prefer the shadows to have a bit less sharpening than the mid-tones for two reasons. First, the human eye is drawn to sharp objects. Generally, the shadow areas are not the center of interest in an image (there are exceptions). Rather, the shadows tend to play a backup role. In most cases, I do not want the shadows grabbing attention away from the main objects in the image. Second, the shadow areas are more prone to noise than the better illuminated areas. Reducing the sharpening of the shadows will make the noise in the shadow areas less prominent.
Figure 13 shows the Shadow Menu. The Shadow menu has the same controls as the Highlight menu only they are applied to the shadow areas (thus, they will not be repeated here).
Figures 14 through 16 show a crop of an area that includes some rock that is in shadow. Figure 14 shows the unsharpened image. Figure 15 shows what the area looks like with the same settings that were used for the mid-tones (i.e., no reduction in sharpness). The shadowed part of the rock needs to have the sharpening reduced a small amount. By setting the Fade Amount to 25 and the width amount to 45, the sharpness has been reduced an appropriate amount as seen in Figure 16
Smart Sharpen is certainly an improvement over USM. It provides different blur options and allows for some adjustment to the sharpening to account for the different sharpening needs of the mid-tones, highlights, and shadows. Nonetheless, we are still not out of the woods. Another look at the Feather River image (reposted as Figure 17) shows an issue that has not yet been addressed. Not only does this image have a significant range of tonal values; it also has a significant range of texture from the rather smooth texture of the water worn rock at the river's edge to the fine textural detail of the small leaves. It would be highly advantageous to be able to sharpen different textures differently. Areas of significant texture should get higher degrees of sharpening while smooth areas should get little or no sharpening. Unfortunately, Smart Sharpen does not address this issue.
In addition, the other problems still exist. While noise in the shadow areas may be decreased by attenuating the sharpening in those areas, noise in other areas will be sharpened with Smart Sharpening.
The sharpening requirements of the image content may still conflict with the sharpening requirements of the output device.
Smart Sharpening can also cause color fringing.
Smart Sharpen is also destructive and the sharpening is permanent -- it can not be undone.
In essence, the way Smart Sharpen has been applied in this example, it is still a one pass approach. While it may be better than USM, it still suffers from many of the shortcomings of the one pass approach.
While the first reaction to this news may be, "Darn, all that work and I still have sharpening problems", we are headed in the right direction. The next sharpening approach, High Pass sharpening, deals with one of these issues.
As we move up our hierarchy of sharpening methods, High Pass Sharpening takes us to a new level -- one that allows for a significant amount of flexibility. High Pass sharpening is not a tool (you will not find a High Pass Sharpen tool in Photoshop); rather, it is a technique that uses the High Pass filter on a duplicated background layer to sharpen the image.
High Pass sharpening has two big advantages. The first advantage is that the sharpening is done on a separate layer. With the previous tools, the sharpening was applied to the Background layer. Sharpening is destructive. When applied to the Background layer, it can not be modified or undone at a later time. With High Pass sharpening applied on a separate layer, not only can the sharpening be adjusted at a later date but also can be completely undone simply by deleting the layer that was added during the High Pass Sharpening process.
The second advantage is that the High Pass sharpening technique uses the High Pass filter, which isolates the edges. Thus, High Pass sharpening applies sharpening primarily to the edges where it is needed and protects smoother areas from the sharpening.
The High Pass sharpening technique starts with the Layers palette (see figure 18). The Background layer is duplicated. This can be done by dragging the Background layer down to the Create a new layer icon at the bottom of the palette. The new Background copy layer is selected and the bend mode is set to Hard Light (see Figure 19).
The High Pass filter is then run on the Background copy layer. The High Pass filter is opened by selecting Filter/Other/High Pass. Figure 20 shows the High Pass filter menu. The only control that needs to be adjusted is the Radius. Images with fine detail will require a small Radius (I used a radius of 3 on this image). Those with less detail will use a larger radius. As in the case of USM, increasing the radius up to a certain point will make the detail stand out without degrading it. Further increases in the radius will cause the detail to degrade. The goal is to find that point and set it as the radius (some experimentation may need to be done to find the correct radius). The Preview box shows the results of the High Pass filter on the Background copy layer. The screen can be viewed to see the effect that the High Pass filter is having on the image. Clicking the OK button closes the menu and carries out the High Pass Filter action.
Now the task of making the final sharpening adjustments begins. There are two adjustments that are made at this time (both are found on the Layers palette). The first adjustment that needs to be made is to determine whether to use the Hard Light or Soft Light blend mode. The Hard Light mode will create a much higher level of sharpening than the Soft Light mode. The second adjustment to be made is to set the Opacity control. After running the High Pass filter, the Opacity control will be at 100%. This creates the maximum amount of sharpening. Reducing the opacity will reduce the amount of sharpening. A little bit of playing around with these two controls should yield the desired amount of sharpening -- these two controls can always be changed at a later time if so desired.
Figures 21 and 22 show the effect of High Pass sharpening on one image. Figure 21 shows an image before any sharpening has been performed. Figure 22 shows the same image after High Pass Sharpening has been applied.
High Pass Sharpening has some real advantages over the previous sharpening tools. The ability to adjust the sharpening at a later time through the use of the blend modes and the Opacity control is significant. It should also give the photographer some peace of mind knowing that the sharpening can be quickly undone by deleting the sharpened layer. Perhaps less obvious is that by performing the sharpening on a separate layer, the High Pass sharpening method can be used in conjunction with more sophisticated sharpening techniques (to be covered in subsequent articles).
Even with the High Pass sharpening technique, we are still stuck with some of the other sharpening issues. On the other hand, we are picking up techniques, each of which solves a specific problem. As we continue to develop an arsenal of sharpening tools, we will be able to pick and choose those techniques, or combinations of techniques, that will solve our sharpening challenges.