Scanning Tips



Resolution determines the level of detail recorded by the scanner, and is measured in dots per inch (dpi). The greater the dpi number, the higher the resolution. Image quality improves with higher resolution, but only up to a certain point, after which increasing resolution simply makes file sizes larger (and sometimes unmanageable) without yielding any visible improvement to the image.

There are two types of resolution. Optical resolution is the "native" hardware resolution of your scanner as determined by the optics in your scanner hardware. On the other hand, interpolated resolution is resolution enhanced through software (either in the scanner or the scanning software), and is useful for certain tasks like scanning line art. A typical desktop scanner may produce 600x600 dpi optical resolutions and interpolated resolutions of 4,800x4,800 dpi or higher.



When choosing a resolution setting for B/W photographs, consider the type of image you're scanning and your printing method. An easy way to determine the best resolution for your intended output is to find out the lines per inch (lpi) halftone/screen capability of your platemaking or printing device and multiply it by 1.5 to 2.0.

600 dpi Laser Printers produce good halftones to 85 lpi

1200 dpi Laser Printers produce good halftones to 120 lpi

1200 dpi Imagesetters produce good halftones to 133 lpi

2400 dpi and higher Imagesetters produce good halftones to 200 lpi

Example: To tailor a scanned image of a B/W photo to a typical newsletter printing press that prints at 100 lines per inch, multiply 100 x 1.5 or 2.0. This gives 150 or 200. The optimal resolution setting for your image would then be 150 dpi to 200 dpi (depending on how high the output quality will be). Lpi will vary, depending on the quality of the printing job. A newspaper web press uses approximately 85 lpi; newsletters use 100 to 120 lpi, brochures/magazines from 133 to 150; and fine art books may go as high as 200 to 300 lpi.



If you're outputting images to a monitor (such as doing web work), you don't need to scan images higher than 72 dpi, as monitors are capable of only showing images up to 72 dpi. A higher-resolution image will not be any clearer on the monitor and will simply create larger files.

Remember that the higher the resolution, the larger your image file will be and the longer it'll take to send over the internet. An 8.5" x 11" color photograph scanned at 75 dpi takes up about 1.6 megabytes (MB). Doubling resolution to 150 dpi will increase the file size four times -- to approximately 6.3MB! Going to 300 dpi will increase file size to 26.2MB.

What you need to do then is to select the lowest possible resolution that still gives you good image quality to keep file sizes manageable.

Note, however, images created at low resolutions for monitor viewing will appear fuzzy at resolutions of 100+ lpi when output for offset printing. It is not possible to increase the resolution of low dpi scans by resampling the image in a software application like PhotoShop (also known as software interpolation). The higher quality resolution must come from the original hardware scan’s optical resolution.



High resolution for line art is important if you're processing an image through a high-end output system that carries data from the scanner through the final output. This is because high resolution can improve the sharpness and clarity of the dots that make up the image.


Interpolated resolution may be useful for scanning line art or enlarging small originals. Line art is any black-and-white or single-color graphic, such as a logo, ink sketch, or mechanical blueprint (not a photo or screen).

For line art: Set the resolution equal to that of your output device. For instance, if you're producing line art to be printed by a 1200-dpi imagesetter, you can interpolate resolution up to 1200 dpi for superior results. This will produce smoother lines and eliminate some of the jaggedness characteristic of poor line art scans.

For enlarging small originals: Let's assume that you scan a 1" x 2" line art drawing at 600 dpi using a 600x600 dpi scanner with maximum optical resolution of 600 dpi. To enlarge the image to two times the original size without loss of detail, interpolate the resolution to 1200 dpi. This way, the image retains clarity and sharpness when the print size is doubled.



Scaling is the process of creating larger or smaller images in your scanning software so that you need not resize the images later when they are delivered to your image-editing program. Scaling has an inverse relation to resolution: The lower the resolution, the larger the image can be scaled. At the highest resolution, images can only be scaled smaller.

To illustrate the use of scaling: Assume you scanned a 2" x 2" photographic image at 200 dpi. To double the image size to 4" x 4" without loss of detail, increase scaling to 200% and maintain resolution at 200 dpi. This is also the same as scanning the image at 400 dpi at 100% scaling on the scanner - and then using your DTP software to enlarge the placed photo to 4” x 4”.

Important Concept: Both these methods will produce the same file size – i.e. doubling the scanning resolution has the same effect as doubling the image size. The resulting file size, though, will be quadruple the original, not double.



(Applies to Some Scanners and Some Software – for Photos only)

Resolution might add detail, but it can't add sharpness, they're different things. The scanners digital samples every couple of 1/1000 of an inch are not as sharp as a good camera lens. The standard tool of choice for sharpening a photo is the Unsharp Mask filter. It's magic! A little tweak can make a big improvement. The Unsharp Mask filter has parameters that allow it to affect only the edges in the image, and to exclude the smoother low-contrast areas. It is NOT a simple on/off choice. And it is these parameters that make the USM so powerful and useful for photographic images.


Radius controls how wide the edge rims become, and Radius = 1.0 is about the right ballpark, with 0.6 to 1.5 often being useful. Radius values around 3 or 4 can cause halos at the edges, a detectable faint white rim around objects. Radius units are not the same as pixels, the units step in tenths, the width is usually at least 4 pixels, but you will see various effects. Radius is a very important parameter, and the easiest way to ruin a good scan is with too much Radius. Inanimate objects can use the most radius, human faces can tolerate the least, and landscapes fall in between. But it really depends on the size of the details. Fine detail needs a small Radius, or you lose detail. Large images need more Radius. One suggestion for printing is perhaps a Radius Setting = Resolution / 200, i.e., 0.5 for 100 dpi, 1.0 for 200 dpi, 1.5 for 300 dpi, where Resolution is the scaled output resolution (on paper when printed). Radius and Amount interact, reducing one allows more of the other.


Threshold specifies how far apart adjacent tonal values have to be (values of 0…255) before the filter does anything to the edges, before it is judged to be an edge at all. This lack of action is important to prevent smooth areas from becoming speckled. Low values should sharpen more because fewer areas are excluded. Higher threshold values exclude areas of lower contrast. Human faces want values greater than 1 or 2, like perhaps 5 or more. For inanimate objects, perhaps 0 or 1 is useful. General work, try 3 or 4. This control has little effect at high values, but has more effect changing between low values of 0 to 5.


Amount is like a volume control, exaggerating the edge differences (how much darker and how much lighter the edge borders become). Amount interacts with Radius as to degree of sharpening, but it does not affect the width of the edge rims. Amount has a large effect, and values of 80 to 120 are normally usable if the Radius isn't too large.


Again, fine details and small images need a small Radius and more Amount, else sharpening can be counter-productive. Large or featureless objects can use much more radius. How it looks on the continuous tone video monitor is not the same as it will print on paper using a halftone screen. At 150 dpi, Start at about A=100, R=1.0, T=3 for objects, and less for portraits, say A=60, R=0.6, T=6. There are many combinations, so experiment, and the macho types must learn not to over do it.


Adobe Photoshop has the definitive USM, and Adobe PhotoDeluxe works the same. PaintShop Pro 5.01 is similar, but 4.14 is lacking a USM filter. Ulead PhotoImpact also has an Unsharp Mask filter, but the parameters are different and it works a little different too. PhotoImpact 3.01 defaults are WAY too big. It will work if you back off considerably, start at the bottom end. Sharpen Factor = 20 to 25, and Radius = 2 or 3 seem useful. Again, PhotoShop has the definitive USM, and it can use slightly higher numbers, perhaps 120 for Amount. If using the USM in the graphics application, and if you don't like the results of the USM, don't simply do it again on top of the previous try. Instead, use menu EDIT - UNDO to go back to the original image, and then try different USM parameter values. Also, repeating the UNDO / REDO menu back and forth will quickly show the results of the USM.


USM should be the final operation, after all other adjustments (with the one exception of the histogram Output Level settings to limit the levels for halftone printing, which must be done last).



Scanned images are generally saved as graphics files, and several graphic file formats are available for use. To get the best scans, be familiar with the pros and cons of each file format and how they are compatible with your D-T-P software and print method.



(listed in order of preferred format for printing halftones):


TIFF: Short for Tagged Image File Format. TIFF is probably the most popular file format. It's good for storing bitmaps in many different resolutions, color models and compression types, and is supported by most commercial applications. Use the TIFF format whenever possible, since this is the most widely used.


EPS: Short for Encapsulated PostScript. EPS is good for storing vector drawings but not for line art (with the exception of Photoshop EPS). EPS is ideal for print applications because it offers more control and compatibility when printing to a PostScript printer. While vector drawing EPS files are high quality and small file size, bitmapped eps files like Photoshop EPS are usually very large.


PSD: Adobe Photoshop's internal image format not usually supported in PC or MAC dtp applications. These files retain PhotoShop only attributes (like layer information) and can easily be changed to the TIFF format with no loss of quality.


PCX: Developed by Z-soft for use in a few paint programs. Also suitable for scanned images and is widely supported for PC use in many D-T-P programs.



(not usually suitable for offset print halftones):


JPEG: A compression algorithm format used to store color or grayscale files. This format allows the user to trade off image quality for a smaller file size when saving the scanned image. The degree of compression for this format can result in sever degradation of image quality, In contrast to GIF, JPEG is the recommended file format for compressing images with continuous tone or complex color shadings.


GIF: A format used to store images with 256 colors or 256 shades of gray. GIF is a good file format for compressing images to be used on the Web, as long as the images are made up of solid blocks of color -- like logos, or lettering. It's not as good for images with continuous tone or complex color shadings -- like skin or images with rich shadow / highlight details.


PICT, PICT2: PICT is generally used for line art with limited color (usually 256 colors). PICT2 is for 8-bit grayscale or 24-bit color images. This format is not widely supported by PC applications.



Good scans produce good looking, easy to print halftones.

Poor scans and low resolution files produce group ulcers and general disapointment to all involved with the project.

No Comments Yet.

Leave a comment