Graphic elements can be used to enhance the readability of a publication and to illustrate specific points in the text. They can include rules under a running head, a printed tab at the page edge, a map, a photograph, etc. They can be created in the page layout program, created in and imported from an electronic illustration program such as Adobe Illustrator or Photoshop. Two types of graphic formats are used in electronic publications: object-oriented and bitmapped. Using these two formats appropriately can mean the difference between problem-free imaging and not imaging at all.
 
Object-Oriented (Vector) Images
Vector graphics are a series of points in electronic space describing a path and what is contained in that path (a fill, for example). Vector graphics are like string art: each pin is a control point, and manipulations to that point determine how the thread following it will behave.
 
• Keep it simple! The more control points associated with a vector drawing, the bigger the file and the greater the chances for imaging failure. Autotraced graphics often contain an enormous number of control points; many of these can often be deleted, simplifying the file.
• If you have created several small elements to be placed on various pages, export or save them individually. Cropping out extra items in the page program does not delete the extra data from the file.
• Nested graphics (an EPS graphic placed into another EPS graphic) can cause major imaging headaches and should be avoided.
• Masks or clipping paths over TIF scans should be used with caution. Too many of these in a graphic can cause a page to "choke" the RIP.
• Color names must be identical to those used in the page layout program, and should be designated as spot colors and not process colors to produce correct output (unless you are producing process color art).
• Make sure you use percentages of one spot color as tints, and not different spot colors, to prevent a 2-color job from imaging as a 6-or-more-color job!
• Degradés, vignettes and blends created in drawing programs work well in small areas. Setting the target printer for a high-resolution device will help to prevent banding.
• For larger areas, such as backgrounds, create vignettes in Photoshop or similar programs as high-resolution bitmaps.
• Fonts used in EPS graphics can be converted to graphic elements, eliminating the need to supply the font along with the graphic. This option, however, can degrade the appearance of small text and/or "delicate" faces and should be used with care.
• PICT and PAINT files and fills should be avoided. They were not designed for high-resolution output.
• Metafiles are similar to PICT files. Some PC clip art is supplied in this format. These should carefully be converted to EPS files for placement in the document.

Scans and Bitmaps (Raster) Art see scanning help for more info on ppi, dpi, lpi
Raster graphics are like tiles in a mosaic. Resolution (pixels-per-inch or ppi) determines the size of each "tile."  Monitors have a typical resolution of 72 ppi (where each pixel is a "dot"), laser printers have a typical resolution of 300 ppi, and image-setters output at resolutions from 1,200 to 3,600 or greater dpi. Desktop flatbed scanners typically offer hardware resolution of up to 300 dpi and use software interpolation to achieve resolutions of up to 1,200 ppi. Professional drum scanners can achieve real resolution of 4,000 ppi. Higher resolution = smaller "tiles." Smaller tiles = smoother transitions. Each tile contains a specific amount of data (even the "blank" or white ones).
• Resolution should be determined at the time the image is created or scanned. Attempting to image a 72-ppi scan at high resolution (1,200 ppi or greater) will result in obvious bitmapping or "pixelizing" of halftones and "jaggies" in line art.
• Cropping and sizing should be done first at the scanning stage. Any additional cropping, sizing, editing or rotation should be done in an image editing program such as Photoshop and not in the page layout program. Extra white space around an image should be cropped away in an image editing program because each pixel of blank white space increases both image size and RIP time.
• Converting color images to grayscale will often require manipulation of the various color channels to get good quality reproduction. Colors in the purple-to-red range tend to turn too dark, and colors in the yellow-to-blue range tend to disappear.
• What-you-see-is-NOT-what-you-get!  The image that looks great on your monitor is not necessarily what will print the best. Low-resolution (laser) proofs are not adequate for halftone proofing. Minimum dot value and dot gain compensation issues need to be worked out in advance through test files output at high resolution. Good halftone scans require high-quality equipment and a thorough understanding of press requirements.
• Halftone scans should be made as close to the final printed size as possible, at a resolution of 1.5 to 2 times the desired line screen (a 133-line printed image should be scanned at 12 times the line screen ruling to avoid moire or screen clash 12x133 =1596 ppi, then in photoshop resample the image to 266ppi grayscale or higher for color). Scanning at too low a resolution will result in obvious bitmapping of the image and/or loss of detail. Scanning at too high a resolution will pick up unwanted artifacts, create unnecessarily large files and longer RIP times.
• Line art should be scanned at the highest resolution possible, but with some thought given to the type of art. A soft, sketchy piece may reproduce beautifully scanned at 600 ppi and won't improve at 1,200 ppi. A text-intensive piece, such as a logo or a schematic drawing, requires at least 900 to 1200 ppi to reproduce well.
• In many instances, it would be better to re-create line art as an object-oriented drawing than use it as a scanned image. A full-page scanned border could easily take up 12 megabytes of space and 40 minutes to image. The same border, re-created as a vector image, might only be 120 kilobytes and image in 3 minutes.
• Many DTP programs offer some form of picture replacement using low-resolution FPOs during document production and replacing them with high-resolution versions before imaging. This can save time and disk space for the publisher, but should be discussed first with the printer. Not all imaging centers have this capability.
• At this time, electronic retouching of FPO images cannot be automatically reflected in the high-resolution version. Any retouching must be done to the final high-resolution version.