Regardless of your level of photographic expertise, there will likely come a time when you'll want to display your work on the Internet - either for self-promotion or marketing purposes, or simply for sharing with others to obtain their comments or critiques. Naturally, that means having to scan your prints, negatives, or slides. This article will describe some of the basic elements of scanning. We won't cover all of the technical nuances, nor all of the available options or techniques. Hopefully, however, describing my personal scanning techniques (using the Silverfast scanning software with an Epson 3200 scanner) will get you started toward making acceptable scans of your images. If you are using different software with a different scanner, the concepts explored here will still apply, but you'll need to make adjustments in terminology and procedures for your equipment and software.

You should note that I am not a scanning expert, nor do I claim to have all of the answers. Over the years, however, I've figured out ways to get my scans to reasonably represent the quality of my prints, negatives, and slides, given the limitations of Web display. There are numerous other ways to achieve the same or similar results, so you shouldn't consider the procedures discussed here as being the only way to do things.

Software, Terminology & Tech Stuff

The nature or quality of your scans will largely depend on:

• your scanner's capability
• the scanning software you use, and
• your image editing skills

Scanners are available in a variety of types, sizes and capabilities, the combination of which translates to a variety of prices. Although delving into comparisons of specific scanners is well beyond the scope of this article, there are a few technical issues that are worth noting - especially if you haven't already purchased your scanner.

Scanner types can be boiled down to:

• flatbed scanners for prints only
• flatbed scanners with film/transparency adapters
• dedicated film scanners
• drum scanners

The key features to consider include the maximum optical resolution, the D-Max, and the scan area size. The scanner's maximum resolution, along with its maximum scan size, will determine how large a scan you can produce, and how detailed that scan will be. That, in turn, will control how large of a digital print you can make from the scan, should you choose to do digital printing. The D-Max spec depicts the dynamic range the scanner is capable of "seeing" in your original. Higher is better here, and scanners having a D-Max of 3.2 or higher are preferred for good photographic work. Most flatbed scanners are limited to letter-sized originals (8½x11" or A2), meaning you'll be limited to scanning 8x10" prints. There are, however, "tabloid" sized scanners capable of scanning 11x17" (more expensive), and smaller ones aimed at the lower-end of the consumer market (much less expensive). Choose whatever is appropriate for the type of work you do.

It's good to recognize (before you buy a scanner) that the better dedicated film scanners will often offer the highest resolution capabilities, and usually higher D-Max specs, than flatbed scanners. Technological advances usually show up first in dedicated 35mm film scanners, and then migrate to flatbeds, and then to medium-format film scanners. Dedicated 4x5 film scanners are usually the last to see advances in technology. Personally, I opted for a dedicated film scanner capable of doing both 35mm and medium format (a Polaroid SprintScan 120) for use with those films, and the Epson 3200 for prints and 4x5 film. Even though the Epson 3200 will do 35mm and medium-format film, the dedicated film scanner produces much better scans for the smaller sized film.

Software - unless you purchased your scanner on the used market, it likely came bundled with scanning software. Ideally, the scanning software should allow you to control the size or resolution of the scan, the "exposure" and the contrast curve. If the software that came with your scanner doesn't provide sufficient control over the scanning process, you may want to consider separate software.

My Epson 3200 came bundled with the Silverfast software, which works as a plug-in for Photoshop. Silverfast is usually regarded as one of the best scanning packages available, but if you have to buy it separately, it's fairly expensive. An inexpensive package that often gets good reviews is Hamrick Software's Vuescan.

Input, Output & Screen Resolution

Depending on the scanning software you use, the control over the "resolution" of the scan may vary. Often, this is expressed in terms of DPI (dots per inch), a term borrowed from its printing origins. Technically, this should probably be "PPI" - pixels per inch - but most people equate a dot to a pixel, so that term seems to have stuck. If DPI is used by the scanning software to determine the size of the scan in relation to the original, the number of pixels in the scan will be derived by multiplying the Size of the original (in inches) by the chosen DPI. Thus, a borderless (edge to edge) 8x10 scanned at 300 DPI would produce a digital image that was 2,400 x 3,000 pixels in size. Some software, like Silverfast, takes sort of a reverse approach by allowing you to set the size of the desired output print and the resolution (DPI) of that print. The software then selects the scan resolution required to create that file based on the area selected to be scanned.

In between scanner input and printer output is screen or monitor display. The size at which any given digital image will display is controlled by the number of pixels in the file, and the size and resolution of the monitor. The DPI value associated with the file is carried in the informational portion of the binary file header, but has no affect on the size of the displayed image. Thus, the often-heard suggestion of setting the image parameters to 72 DPI for Web display is at best misleading. For Web display, it is more straightforward to think in terms of the pixel dimensions of the image. This issue gets a little confusing, however, as some image editing software will allow the use of the DPI setting to control the re-sizing and re-sampling of the base image. We'll return to the issue later.

Monitor Calibration

The importance of having a reasonably well-calibrated monitor can't be stressed too much. Calibration controls both the tonal range your monitor displays, along with the color balance. If your monitor's calibration is "off", your image editing will be off by a corresponding amount. Similarly, those with properly-calibrated monitors will see a different image than you're seeing. Photo editing software often contains auxiliary routines for roughly calibrating your monitor. With the full edition of Photoshop, that is an auxiliary program called gamma.exe, accessed during the initial configuration of the software, and later through the Windows menu system. Third-party calibration software, some of which uses a screen sensor (usually called a "spider"), can also be used for more precise monitor calibration. These programs often use ICC Profiles as part of the calibration process, enabling close correspondence between what you see on the screen and what will be printed by specific printer/paper combinations. The use of ICC Profiles, however, is well beyond the scope of this article.

Scanning Procedures

If your scanner has a transparency adapter (usually a lid with a built-in light source), your software will likely offer an option for selecting the type of original - either "reflective" or "transparency", or some equivalent wording. In Silverfast, that's part of the "General" tab, and shows up after starting Photoshop and selecting File->Import->Silverfast (Epson) to start the Silverfast plug-in. While you're on the General tab, I also suggest turning off sharpening and other filters within the scanning process. Once you've selected the appropriate type of original, you can click on the navigation icon for doing a preview scan. The scanner will then do a quick scan of the original, and display a reduced-size verion in the preview screen. You can then adjust the area to be scanned by grabbing the corners of the dashed-line selection area and moving them as needed.

This is a good point to discuss scan types, as well. You'll notice in the example, grayscale images are described as 16-bit or 8-bit, and color images as 24-bit or 48-bit. How many bits are used to define the tone or color of each pixel is also called the "color depth" of the image file. For Web purposes, grayscale images use 8 bits to describe the shade of gray (including pure white and pure black), so each pixel will consume one byte of disk space when saved. As such, the "palette" of a grayscale image is limited to 256 shades. Color images, in contrast, are usually 24-bit images, resulting in a much larger palette of tones and colors. Some scanners will often scan at a higher color depth, and then convert to the color depth commonly available in image editing software. The latest version of Adobe Photoshop, Photoshop CS, for example, will handle some operations in 16-bit grayscale, and most operations in 24-bit color as an RGB file. It is often beneficial to scan B&W images as RGB color, just to make use of the larger palette of tones. Otherwise, large areas of subtle gray tones may show "color banding", rather than a smooth transition of tones.

The next step is to decide how large you want the scan to be. My approach is to assume that I might want to make a digital print of the scan, and to scan each original only once for various potential uses. Thus, even though the initial use of the scan may be aimed at a reduced-size version for Web display, I designate the maximum size of the digital print I might want to make, along with the resolution I want for that print. For my Epson 2200 printer, I usually use 300 DPI as the desired output resolution. In this example, I selected a scan area that measured 3.1x4.0 inches, the output resolution ("Screen") of 300 DPI (LPI), and designated a print size of 11x14 inches. The scanner software then calculated that it needed to scan at 1058 DPI to produce a file of that size. Although not shown in these examples, this was an 8-bit grayscale scan, so the resulting file would have been 13.23MB in size. A color image file would typically be about 30MB in size. These files are not intended for external use, but rather function as my digital "master" files, from which I create other versions for other uses. I save these master files in native Photoshop PSD format for maximum quality.

Scanner Controls

Silverfast provides extensive control over the scanning process, including setting black, mid-tone, and white points, as shown in the example to the left, and control over the contrast curve, as well. The histogram (the graph with the peaks and valleys) shows the relative distribution of pixels of various densities. It's usually best to set the black and white points so as to include all of the available image data. To adjust these points, just grab onto (on PCs, left-click and hold) the little triangle below the histogram with your cursor, and move it where it needs to be. The preview of the scan will change correspondingly. Also, pay attention to the "Color Space Compression" display. I prefer to move those controls to both extremes, so that no compression of the tonal range takes place. Again, just grab onto the triangles with your cursor and move them. Using this procedure, you'll want both sets of Color Space Compression controls (above and below the gradient bar) at the extremes. Once you've made any necessary adjustments, just click on OK to proceed.

Adjusting the contrast curve is a little more delicate. Silverfast will save and store the last contrast curve you used, and reload that curve for the next scan. If the last curve isn't appropriate, click on the reset button to restore the default straight line contrast curve. You can then re-adjust as needed. For example, to reduce the intensity of highlights in the preview, grab a point on the upper portion of the line and drag it downward. You'll see the result reflected in the preview, and you'll notice that subtle changes in the contrast curve will result in significant change in the preview image. Adjustments can also be made to the darker tones in the image. Tinker with these adjustments until the preview image looks about the way you want it. Additional adjustments can be made in Photoshop once the image is scanned. It's good, however, to get the original scan reasonably close to what you want.

Key Photoshop Adjustments

The scope of this article is not to provide a full tutorial on Photoshop, but there are a few controls in Photoshop that I consider essential to creating good, representative scans of photographs. These include Levels and Curves (similar to the histogram and curve adjustments within Silverfast), the unsharp mask filter, and image size adjustments.

To adjust Levels in Photoshop, select Image->Adjust->Levels from the top menu bar. A histogram like that shown here will be displayed. I usually adjust the black point so that is immediately below the beginning of the black data at the left of the histogram. Within Photoshop, the white point is really white with no detail (pure white). So, I usually adjust the white point to be slightly to the right of the image data, so as to retain highlight detail. In adjusting either of these, you'll notice they also adjust the mid-tone triangle. So, you may need to readjust that after setting the black and white points.

Obviously, adjusting black, mid-tone, and white points also affects overall contrast. Photoshop also provides adjustments for Brightness and Contrast, as well as a Curves adjustment. I prefer to use Curves rather than Brightness and Contrast, as it provides a finer level of control. I also use the Curves adjustment after doing the Levels adjustment. After experimenting with these controls, you'll develop a feel for how to best use their inter-relationship.

Unsharp Masking - Almost all scans will require some type of sharpening to approach the quality of the original. The trick is to do so without introducing digital artifacts. For this purpose, I prefer Photoshop's Unsharp Mask filter, accessed via the Filter->Sharpen->Unsharp Mask selection from the menu bar.

The sample image above shows the settings I use for unsharp masking. The Photoshop manuals provide a formula for calculating the optimum Radius and Threshold settings based on the size of the image. Rather than changing the settings with each new image, I prefer to leave the settings at what will work for an image of around 500 pixels on the long side. As Photoshop stores the last filter used within the menu bar, this is convenient for speeding up the digital workflow when editing multiple images. For larger images, the filter can simply be applied multiple times (two or perhaps three times) in succession. This approach also seems to produce better overall results.

Spotting - As in the "wet" darkroom, dust gets everywhere when scanning, too. To "spot" scans, I use the Photoshop Cloning tool (the rubber stamp in the tool box). First, I enlarge the image to 100%, so I can see all the dust, move the image display to one corner, and then select the Cloning tool. Pressing the ALT key allows you to select the area from which to clone, so I select an area that is the same tone or color as the area around the dust speck, and select a soft-edged brush of an appropriate size - something between 13 and 19 pixels is usually about right. Selecting an opacity of less than 100% will also allow for better blending of your spotting efforts. Then, just scroll through the image, spotting as you go.

Reducing Image Size - After all of the above adjustments have been made, and the "master" scan has been saved as a Photoshop .PSD file, you can make a reduced-size version for Web display. There are a couple of factors to consider here - the dimensions of the image in pixels, and the level of JPEG compression to minimize the resulting file size (and, the time it takes for people's browsers to download the image). In most Web venues, an image size of about 550 pixels along the long side of the image fits the desired parameters for a "large" version, and a final file size of around 70KB is considered "polite".

Better image quality results if the resizing is done in steps of about 50% or less. A light touch of unsharp masking at each step also helps to preserve the quality of the original. Thus, Assuming I'm starting with an image of 3000 pixels on the long side, I'll resize to 1500 pixels in the first step, apply the unsharp mask filter using the values mentioned above, then move to 900 pixels, unsharp mask, and finally the end-size of 550 pixels.

Note that for this purpose, all three of the check boxes at the bottom of the Image Size window should be checked. Constraining proportions will keep the dimensions in the proper ratio, and resampling will cause the software to reduce the number of pixels in the file.

If you un-check the Resample Image box, you can also change the DPI setting without changing the size of the image, or the amount of space it requires on-disk. However, increasing the DPI setting with the Resample box checked will force the software to increase the number of pixels, "manufacturing" enough new pixels to fit the requirement. Doing so usually results in unexpected distortions and digital effects unless done carefully. If you really need additional pixels for a larger print, or some other reason, it's usually better to rescan the image at a higher resolution.

Saving as JPEG - After you've finished re-sizing your scan, it should be saved as a JPEG file for Web display. The JPEG file format provides an adjustable amount of image compression which produces a smaller file. The level of compression used is a trade off between image quality and file size. Within Photoshop, this is accomplished by clicking the "Save as" option in the File menu-bar selections. I like to append the file size (pixels) to the file name, so I can easily distinguish between versions when looking at directory listings. So, if I started with "lighthouse.psd", I'd save the 550-pixel version as "lighthouse-550.jpg".

Once you've added the "-550" (if you choose to), and click on "Save", a JPEG dialog box will appear. You can then select the level of compression you want to use. Within Photoshop, I find that a level of "8" is usually a reasonable compromise between image quality and file size. For an image of 550 pixels on the long side, that usually keeps the actual on-disk file size to around 70KB or so. Files too much larger than that usually take too long to download, use too much bandwidth, and are generally considered to be "anti-social" on the Internet. After you've selected the compression level, Photoshop will calculate how large the resulting file will be, and displays that information, along with how long it will take to download over a 56Kbps dial-up connection. The on-screen image also displays the resulting image quality. Once you're satisfied with your settings, just click on the "Save" button, and you're done.

Developing a Work Flow - As you can see, scanning is not all that difficult, although there is a learning curve and some skill involved that can be developed through experience. Don't be afraid to try different techniques, but remember to save your work at critical points. That way, if the experiment goes awry, you can still recover the previous good version. In doing so, you'll also develop a workflow that streamlines the process when working on a series of images. Photoshop provides various features that help in this regard. And, as with photographic experiments and testing, taking notes along the way helps, too.

Web Display

In order to display images on the Web, they must be uploaded to a server that provides Web services. It is not possible to display files directly from your PC. Some Web communities, like APUG provide a limited amount of space for using the images within that community. External links, such as posts on other forums, are generally precluded by the server's settings. If you want to post images on a variety of forums, or want to use your images to promote your photography business, it's a good idea to contract with a service provider, obtain your own domain name, and create your own photography site.

Happy scanning!