Scanning Art in Photoshop
for Use in Video Programs

Scanning is as much an art as a science. Thousands of pages have been published about the mysteries of scanning, and I won't try to summarize them here -- instead, I'll confine myself specifically to drawings and photos that will be transferred to video via a digital editing system. But these tips will only by useful if you're already familiar with the basic principles of scanning. (For a thorough introduction to the theory and practice of scanning, I recommend Real World Scanning and Halftones by David Blatner and Steve Roth, published by Peachpit Press.)

One overall rule worth mentioning, however, is to do as much work as possible before you scan. Adobe Photoshop is a wonderful tool for enhancing and retouching images, but it can only work with the information that is present in the original scan. So do as much preparation as you can to make that scan as high quality as possible.

(NOTE: Photoshop has no standard scanning module; different scanners come with different scanning software. Therefore, depending on what scanning software you're using, the details of the following steps may vary.)

Step 1: Cropping
Step 2: Resolution
Step 3: Image Mode
Step 4: Black Point/White Point
Step 5: Gamma
Step 6: Processing and Retouching
Step 7: Scaling and Cropping
Step 8: Saving the File

In your scanning software's Preview window, use the rectangular marquee tool to frame the portion of the image that you want to scan. The absolute dimensions don't matter, but try to make the aspect ratio of the rectangle is as close as possible to 4:3. (That is, the width divided by the height should be approximately 1.33.) Keep in mind that the outermost edges of the image won't be visible on the TV screen, so make the rectangle about 10% larger than the portion of the image that you'll want to see.

You can also use the marquee as a guide to make sure that the art is lined up properly on the scanner -- that the horizontal lines are truly horizontal and the vertical lines truly vertical. (You can always rotate the image later in Photoshop, but that will require Photoshop to interpolate pixels -- that is, to "guess" about the position and color of certain portions of the image. The less guesswork Photoshop has to do, the better the quality of your video output will be.)
Return

No matter what the size or dimensions of your original art, scan it at the highest optical resolution of your scanner. "Optical resolution" refers to the number of pixels per inch that your scanner can capture mechanically, without digital enhancement. For example, most inexpensive scanners have a maximum optical resolution of 300 dpi (dots per inch) at 100% magnification. The scanning software that came with your scanner may allow you to scan at higher resolutions (say, 300 dpi at 400% magnification), but this is merely a trick: the scanning software is taking the data from the 100% magnification and interpolating pixels to create the 400% magnification. The only way to find out your scanner's maximum optical resolution is to consult the documentation that came with it, or to check with the manufacturer.

Keep in mind that resolution is a ratio, not a fixed quantity. So, for instance, if you scan an image at 300 dpi at 50% magnification, the resolution is exactly the same as if you had scanned the image at 150 dpi at 100% magnification. Either way, your image will look exactly the same on the video screen.
Return

If your original art is in color, scan it in RGB (millions of colors) mode. Otherwise, scan it in Grayscale mode. Never scan an image in Lineart mode, even if it's a black-and-white line drawing -- you'll lose too much information.
Return

Video has a much more limited tonal range than the eye can perceive. When scanning, you want to make the most efficient use possible of this limited range. Therefore, you need to pay special attention to setting the "black point" and "white point" of your image.

By definition, the black point is a very low level of gray that you define as black. Anything as dark as the black point -- or darker -- will appear black in the scanned image. Similarly, the white point is a very high level of gray that you define as white. Anything as bright as the white point -- or brighter -- will appear as white in the scanned image.

To demonstrate why the black point and the white point are important, imagine you're scanning a photo of a black cat against a white wall. The black cat isn't pure black, of course -- it contains many shades of gray, plus there are subtle shadows that allow you to see, for example, that one leg is partly behind another. Ordinarily, if you were to scan this image, you'd be squeezing in the entire tonal range from pure black to pure white, and the cat would appear as a high-contrast silhouette -- all the detail would be lost. But if you were to set the black point to the darkest part of the cat, and the white point to the lightest part of the cat, those subtle shading differences would have room to spread out over the whole tonal range, and the scan of the cat would include much more detail.

At the same time, the white wall -- which is much brighter than the white point -- would appear as a solid white area, with no shading or detail. That's the tradeoff you have to make: by allowing one part of the image to show more detail, you often have to sacrifice detail in a less important part of the image. (If there's a great disparity between the brightness of the background and the foreground, you may want to do the scan twice -- once using the background to set the black and white points, and once using the foreground -- and composite them in Photoshop, as demonstrated in the photos below.)

Most scanning software allows you to set the black point and white point by clicking eyedroppers on appropriate parts of the image. If your scanning software doesn't allow you to set the black and white points before you scan, you can always do it later in Photoshop -- but the result will be less satisfactory, since some of the necessary information will have been lost in the scanning process.
Return

"Gamma" refers to the level of emphasis given to the midtones in an image. Normally, the levels of gray increase in a straight line from the black point to the white point -- this is referred to as a gamma of 1. If the gamma is set higher than 1, all of the levels of gray will become a bit brighter, but those closer to the black point will get more of a brightness boost than those closer to the white point. Since video images tend to lose detail in the darker areas of the screen, you'll usually need to adjust the gamma higher than 1.

If your scanning software allows you to adjust the gamma manually, experiment with different settings to see which gives you the best detail for a particular image. Generally, you'll want a gamma of at least 1.8, though sometimes you may need to go as high as 4 or even 5. (You can see the difference in the photos to the left -- the top one was scanned with the default setting; the bottom one was scanned with a gamma setting of 4.) When in doubt, scan the image at several different gamma settings and see which looks best when imported into your video editing program.
Return

After you've scanned the art, you can use Photoshop's tools to clean up the scan. The first thing you'll probably want to do is color-correct the image, since scanners -- especially inexpensive scanners -- are notorious for skewing color values. The easiest way to color-correct is to place the mouse pointer over a white area in your image and look at the RGB settings on the Info palette: there should be relatively equal percentages of red, green, and blue. If the amounts of red, green, and blue aren't equal, you can use the Levels or Curves window to make the necessary adjustments to each channel.

You may then want to use the Rubber Stamp tool to retouch flaws in the image and the Unsharp Mask filter to bring out detail. (For information on using these and Photoshop's other tools, see the Photoshop manual or one of the many third-party books on Photoshop techniques.) Finally, use the NTSC Colors filter to make sure your image doesn't contain any colors that are technically unsuitable for video.

Original scan	Same scan after being color-corrected, balanced, retouched, and sharpened in Photoshop

Return

Most digital editing systems define a full-screen image to be 640 pixels wide and 480 pixels high. To avoid unexpected results when you import your graphic files into the editing program, it's a good idea to do all your scaling and cropping in Photoshop before you do your video editing.

If you followed Step 2 and scanned your art at your scanner's highest optical resolution, the scanned image will be much larger than 640 x 480. This is good, because certain flaws in the image will disappear when you scale it down. For example, any less-than-expert retouching efforts will look much more natural when the image size is reduced. More important, however, scaling down the image eliminates many instances of "jaggies" and flicker that would otherwise occur, especially with black-and-white line art. This is because Photoshop automatically antialiases an image when it reduces it in size, making it look smoother and more stable on the video screen.

To scale your image down, use the Image Size window (on the Image menu). Make sure the Proportions and File Size boxes are both checked. In the Resolution field, type "72" (if your video editing software is Macintosh-based) or "75" (if your video editing software is Windows based), and click OK. Then open the Image Size window again and uncheck the File Size box. Make sure the units for Height and Width are both set to pixels. Type "640" in the Width field. The number in the Height field will change automatically -- make sure it's 480 or greater. If it's not, type "480" in the Height field. The number in the Width field will change to something greater than 640, but that's okay for now. Then click OK, and your image will be scaled properly.

Now, if the width of your scaled-down image exceeds 640, or if the height exceeds 480, you'll have to do some final cropping. You can do this with Photoshop's Crop tool, or you can use the Canvas Size window (on the Image menu) to enter the proper dimensions.
Return

You may want to modify your image in other ways -- for example, by adding extra channels or layers. These things are fine, but keep in mind that you'll most likely have to save your final image in a standard graphics format -- such as PICT or TIFF -- in order to import it into a video editing program. These file formats don't support layers, and they only support one extra channel: the alpha channel, which is used as a mask to determine which areas of the image should be transparent.

Therefore, if you've added layers or channels, or if your image contains transparent areas, you'll need to do the following things before you save the file: First, in the Channels palette, delete all channels other than Red, Green, and Blue. Second, in the Layers palette, select "Merge Layers" from the popup menu to condense all the layers into a single layer. Third, if you want your PICT file to include an alpha channel, select "Load Selection" from the Selection menu and choose "Layer Transparency Mask." All the transparent areas of your image will be highlighted. Then choose "Save Selection" to save the transparency mask as a new channel.

Now you can save the image to the graphic file format of your choice.
Return

The photos that illustrate this article were taken in China by Debra Goldentyer in 1986.

Go to previous page | Go to home page | Go to next page