The digital-camera industry is booming. That’s hardly surprising when you consider that a
typical midrange camera (priced between $500 and $1,000) not only is fun to use but also shoots million-pixel photographs that you can download, edit, and send to the far ends of the globe in a matter of seconds.
But working with digital cameras also has its drawbacks, especially for graphics and design pros who must prepare the image to suit the high-resolution standards of print media. The fact is, even the top 3-megapixel camera captures at best half the resolution
theoretically attainable with run-of-the-mill 35mm film. To make matters worse, most
digital photographs undergo a heap of JPEG compression, and all are upsampled from the mere 8 bits of color data recorded by the camera’s CCD (which captures the image) to the 24 bits that you see on your computer screen.
Although you can’t magically add resolution to a digital photo, or restore entire elements that the camera missed, you can often salvage what may at first appear to be a lost cause. Using a basic regimen of blurring, sharpening, and blending in Adobe Photoshop, you can smooth over even the most extreme imperfections and enhance fragile detail in an image.
In addition to writing Macworld Photoshop Bible (IDG Books Worldwide, 1999), Contributing Editor DEKE McCLELLAND hosts the 20-tape video training series Total Photoshop
(Total Training, 2000).
1. Photo Woes
Any number of factors can affect the quality of a digital image–from limitations in the camera’s hardware to the technical know-how
of the photographer.
This photograph suffers from common digital-image problems. The most noticeable–its dark tones–results from poor light metering. The image also has a preponderance of “noise,” caused by excessive JPEG compression and a relatively low resolution of just under 1 million pixels.
Although you could avoid most of these problems by using the right digital camera, often graphic designers have no control over the photographer’s equipment. By the time the image lands on your monitor, all you can do is repair the damage.
Following its Photoshop workout, the previously muddied image appears brighter, smoother, and more detailed. Even though you may not be able to cure your tainted pixels completely with this process, you can certainly get them out of intensive care and into something closely resembling a state of cheerful health.
2. Adjust Levels
It’s fair to say that every digital snapshot requires some degree of color correction. Levels and Hue/Saturation are arguably the two best tools for this purpose. The first of these, the Levels command, balances the image’s brightness and contrast.
Before you begin adjusting the image, take a moment to save it under a different name so you don’t overwrite the original. You can save in the JPEG format–it’s perfect for digital photos, after all–but make sure to increase the JPEG quality to 10 (the maximum setting) to reduce further loss of data.
Choose the Levels command (command-L) from the Adjust submenu of the Image menu. The Levels window displays a histogram of the image’s highlights and shadows.
Crop the histogram by pushing the outer sliders
to the first group of pixels on either end of the histogram.
The most important option in the Levels window is the middle Input Levels value
, known as the gamma value, which lets you lighten the midtones. Because the original photo is so dark, I raised the gamma of my image to 1.6–a huge leap.
3. Saturate Your Colors
The Levels command did lighten the photo’s
colors, but it also washed them out, making them appear grayer. To get more-vivid colors, use Photoshop’s Hue/Saturation command.
Using Hue/Saturation has the side effect of enhancing JPEG compression artifacts–often radically. So before applying the command, duplicate the image to a new layer. (Use command-A to select the image and command-J to copy it to a new layer.) Give the new layer a recognizable name, such as “Vivid.”
In the Layers palette, highlight the new layer
.Then choose Hue/Saturation (command-U) from the Adjust submenu of the Image menu.
Increase the Saturation value well beyond what seems sensible. For example, I raised the Saturation of this image to 70 percent
. Don’t worry if the image looks absurdly grainy–you’ll need an extreme effect when it comes time to blend this layer with the original image in Step 5.
4. Smooth the Photo
The next step is to reduce the unwanted grain and enhance the highly desirable edge detail. You achieve the first goal using the often overlooked Median command and its more common buddy, Gaussian Blur; you achieve the second using that most essential of all filters, Unsharp Mask. Together, these three operations will melt away random pixels and flow them into recognizable forms.
Choose Median from the Noise submenu of the Filter menu. In the Radius box, enter a value of 3 pixels or more; increase the value until the compression artifacts are almost entirely smoothed away
. Don’t worry about the fact that the Median command makes the photograph appear doughy or indistinct; you’ll recover the image’s focus later.
Median has a habit of generating its own inaccurate edges. To blur them away, return to the Filter menu and open Gaussian Blur from the Blur submenu. Apply the filter with a Radius value of 1.0 pixel
. A tiny bit of blurring is all you need.
To regain the crisp edge detail, choose Unsharp Mask from the Sharpen submenu under the Filter menu. When adjusting the controls, bear in mind that you’re going for an exaggerated effect. I generally crank the Amount value to its full volume of 500 percent
. (After so much blurring, you’ll need all the sharpening you can get.) Then set the Radius value to 1.0 pixel, matching the value of the Gaussian Blur filter
The result is by no means a perfect image. The photograph should look oversaturated and gummy, as if molded out of brightly colored plastic. You’ll probably also continue to see compression artifacts; in fact, they may be more conspicuous than ever.
5. Blend Layers
You now have an extremely dark and an extremely bright image. To find a happy medium, you need to blend the exaggerated layer with the original image behind it by lowering its Opacity value.
In the Layers palette, adjust the Opacity value of the corrected image’s layer
. If your original photo had few faults to begin with, make the opacity of the new layer low–say, 5 percent or less. (A low Opacity value favors the original image over the filtered layer.) Just a little dab of detail enhancement can make a perceptible difference.
Problem-prone images, such as this one, warrant higher Opacity values. A value of 50 percent is about as high as you’ll want to go: placing more weight on the extreme adjustment layer than on the original image generally results in a surreal effect. Bear in mind that what you’re trying to achieve is not an image that seems radically corrected, but rather one that looks as if it didn’t require correction in the first place.
6. Finishing Touches
Once you’ve found the best balance between the two images, you can merge them into a single layer for further adjustments.
Merge the two layers into one by pressing command-E or selecting Merge Down from the Layer menu
. This affixes the corrected layer to the original, creating a base image that will respond better to standard enhancements.
Now you can apply color and focus adjustments as if you had scanned the image from a high-quality film source. I again boosted the saturation of my colors using Hue/Saturation and reapplied the Unsharp Mask filter–albeit far more subtly than before.
Depending on how your image looked before you started, you may end up with something that approaches absolute perfection. Although my snapshot remains rife with compression artifacts and strange color aberrations, the image now looks far better than it did before.