In most cases, Printing Services prefers to scan all photos and line-art graphics in house that are to appear in pieces that we are to print. This way, we can make sure our customers receive the highest and most consistent quality in their printed pieces. This also gives Printing Services more control over a critical aspect of printing in which we are judged by the audience, whether we scan the images or not. This tutorial is for those customers who insist on scanning their own photos and artwork, and for those who are just interested in the process. This tutorial also includes information about scanning for other purposes besides printing such as WEB publishing.
The following information and guidelines are intended to allow someone to get the maximum quality from a given scanner. This is not to say that the quality is going to be adequate for printing. Whether or not a given scanner has high enough pixel and/or color or tonal resolution has to be determined for each particular job. These guidelines should help in making these determinations. If there are any doubts that a given scanner can do the job, it is strongly advised to let Printing Services do the scanning for printing purposes and that the scanner in question just be used to produce images used to convey positioning and cropping information only. Printing Services has a CD-ROM burner that can be used for saving high resolution photos and giving the CD back to the customer, so they can be cropped and placed into their layout packages. This provides a streamlined way to process and handle photographs, eliminating the need for FPO scanning by the customer. (Why scan it twice?)
Types of Scanning
While Printing Services is mainly concerned about scanning issues that involve the conventional printing process, our customers may need to scan images for a variety of needs. They will commonly include
- scanning for half-tone images to be printed,
- line-art images to be printed, but can include
- scanning for alternate printing methods such as color copier reproduction,
- scanning for presentation slide manufacturing, or
- scanning for computer screen presentation for use in programs or WEB pages.
Each final output method requires its own consideration. This tutorial will discuss each of the different output methods and considerations going into the scanning process for each.
Scanning Halftones for Printing
The most common question about scanning photos for halftone reproduction concern the resolution of the scan or PPI (Pixels Per Inch.) Determining the appropriate resolution for your scan involves three factors; the size of the original, the size of the final image, and the line-screen frequency at which it will be printed. Before we look at these factors in more detail, let us look at the basic process of taking a photo or slide and producing a printed image from it.
Greyscale Scanning: Continuous VS Halftone
Most source photography enters the DTP arena as a photographic print or as a slide or negative. These are all called continuous tone images. Unless you looked at these images through a microscope, you can not see the picture elements that make up the image.
Tone Image (photographic print or slide):
No pixilation, no tonal stepping, and no resolution.
To the naked eye there is no pixilation or tonal stepping visible in a continuous tone image. It appears to be continuous in tonal range and image resolution. If you did look at a continuous tone image under a microscope, then you would see that they too are not truly continuous. It is just that the grain of a continuous tone image is too fine to detect with the naked eye. Since the printing process can come nowhere near this resolution, and since images are printed using a monotonic ink, namely black, a "trick" has been devised for printing these images so they appear to have a tonal range, but you don't need a microscope, or usually even a magnifying glass to see that a printed photo, or halftone, is made up of discrete picture elements called halftone dots. These dots vary in size according to the relative tonal value in the source continuous tone image. The darker the area on the photo, the larger the halftone dot. This "trick" has been employed by the printing industry for a long time, and is only now beginning to be challenged by a new process (possible with computer image processing) called stochastic screening.
Pixilation determined by resolution (PPI), maximum 256 gray levels.
Before computers and DTP, image resolution was not an issue. Since each image scanned into a computer has to have a resolution, we now have to understand how image resolution relates to halftone screen frequency. This relationship is also tied to the resolution of the output device used to create plates, which are used to print the piece. The crux of these relationships is in the plate production. The plate is in some ways an intermediate final product in the printing process. What is on the plate is what will be on the printed piece. To understand the required resolution you need to scan your images at, you have to know how these halftone dots are created on plates.
Halftone Dots on Modern Imagesetters
Let us look at an individual halftone dot and see how it is constructed on an imagesetter. Each halftone dot is constructed from smaller dots known as imagesetter device dots which are arranged in a grid. A typical resolution for an imagesetter is 2400 DPI (Dots Per Inch.) Most imagesetters are driven by a language called Postscript. Postscript has routines that construct these halftone dots. These routines are designed to create a maximum of 256 different sized halftone dots. So the maximum number of tonal values possible on a Postscript device using halftone dot production is 256. Since each halftone dot is produced in a grid of device dots, to get the maximum number of grey levels, or different sized halftone dots, there should be 256 cells in the halftone dot grid. Since the grid is square and the total number of cells should be 256, it should be 16 by 16 device dots large (16 * 16 equals 256.) If we divide 2400 by 16, we get 150. This means there can be as many as 150 halftone dots per inch and still be able to get 256 greylevels per dot.
|Size of halftone dots determined by Line-screen frequency (LPI), maximum 256 gray levels.||Typical halftone dot construction (16 by 16 grid allowing 256 different dot sizes.)|
In contrast to greyscale images are bitmapped, black and white images. You still have the same structure for the image as with a greyscale image, but each pixel is either black or white, not one of 256 tonal values. Typically, your resolution for a bitmapped image needs to be much higher than for a grayscale image. See the table below.
|Typical settings for scanned images|
|DPI of Imagesetter = 600||Image Type||LPI of Halftone||PPI of Scanned Image|
|Bitmapped (Black and White)||Not Applicable||600|
|DPI of Imagesetter = 2400||Image Type||LPI of Halftone||PPI of Scanned Image|
|Bitmapped (Black and White)||Not Applicable||2400|
The reason that bitmapped images need to be a higher resolution than greyscale images is because when they print on the imagesetter, you get just what you scanned. A greyscale image, on the other hand, goes through a halftone algorithm and your PPI needs only be 1.5 times your LPI.
So typically, you want to scan your greyscale images in so that the image will be 225 PPI after any resizing or cropping is done. The best way to accomplish this is to scan the original image at the optical resolution of your scanner. Then, you can set the size and cropping information in a program like Photoshop. Photoshop will do a much better job of changing the image size and resolution than the scanner software that comes with desktop scanners.