Color scanning services are available at the Computer Graphics Lab (CGL). This lab is run by Scientific Computing and Visualization, a group within the Office of Information Technology. Our services are available to the entire Boston University community. Original artwork should be brought to the CGL, located at 111 Cummington Street, Room 203, during normal lab hours This service requires three business days to complete, with an associated cost of $2.00 per scan.
On the scanning form you have a choice of specifying the scanning resolution in terms of either the number of pixels desired in the resultant raster image file or the number of dots per inch (DPI) to be digitized on the original artwork. The default is to create an image which has 600 pixels in the X/horizontal axis and a proportional number of pixels in the Y/vertical axis, as determined by the size (or more precisely, the aspect ratio) of the area to be scanned. For example, if you were scanning a 6"x10" area using the default resolution (600 pixels in X), your resulting raster image file would contain 600x1000 pixels (1000 pixels in Y to maintain the original image's X/Y aspect ratio of 6/10). This would be equivalent to specifying a 100 DPI scan. The maximum resolution of our scanning hardware is 300 DPI. Specifying a higher resolution will increase the amount of detail in the image and, correspondingly, create a larger file. The amount of resolution you need will depend on your application.
If you are still confused about resolution, see Question 3 below on what to do if you do scan your image in at the wrong size for your needs. For a simple answer, though, the following rules of thumb should generally produce reasonable results. If you are scanning in an image for displaying it on your screen or in a Web page, you generally want to use a fairly low resolution (example: specify a 100 DPI scan). If you wish to print your scanned image or do anything else with it where you want/need maximum detail and you don't care about the image's file size, then choose the maximum scanner resolution of 300 DPI. Finally, if you know exactly what size your final image should be, specify the resolution in pixels in either X or Y (example: specify a scan which results in an image which is 500 pixels in the Y/vertical direction).
Image files can be very large. Size is determined by three things: number of pixels (resolution), number of different colors available in the image (color precision or color depth) and file compression. If you need your file to be smaller, you can:
Spatial resolution, as described above, has the greatest impact on the size of your digitized file. Use the minimum resolution that meets your needs.
The number of different colors available to your image, referred to as color precision or color depth, will also affect the file size. The default color precision is 24 bits per pixel (i.e. 3 bytes, one 8 bit byte for each of the primary color components of red, green and blue). Other options are 8 bits per pixel and 1 bit per pixel. The 8 bit per pixel option is particularly useful for grayscale images, while the 1 bit option is adequate for black and white drawings. 8 bit and 1 bit precision will reduce the size of your file by factors of 3 and 24 respectively. You should select the minimum color precision you can, while making sure your application supports the precision you select and you are keeping as much detail as you need.
Finally, as a general rule, all image files should be stored in compressed form to reduce space. Many utilities, including the SDSC Image Tools, accept compressed image files, so you may never need to actually store the uncompressed file. The Unix commands compress, uncompress, and zcat may be used to compress and uncompress files.
You have two options. You can use software tools to resize the image or have the image rescanned at a more appropriate resolution. Software tools such as imscale, one of the SDSC Image Tools which are available on SCV systems, can be used to resize the image. This, however, is not the highest quality solution. Pixels will be added, removed or interpolated with these software tools. Because of this, there will be some level of image degradation. The best way to avoid these problems is to gain a better understanding of the image resolution which you require and then have the original picture rescanned at a resolution that better suits your needs.
If you are having your images scanned in primarily for screen display, the first thing you want to do is check the resolution of the monitor you are using. If you are using X Windows, the command xdpyinfo will provide you with the dimensions and DPI resolution of your screen. Here is an example:
cgl% xdpyinfo | more
The information which is relevant to you will be formatted as below:
... screen #0: dimensions: 1280x1024 pixels (325x260 millimeters) resolution: 100x100 dots per inch ...
This example xdpyinfo information indicates that your screen displays 1280 pixels in X and 1024 pixels in Y with a display resolution of 100 dots per inch (DPI). Most monitors have a display resolution between 72 DPI and 100 DPI. If you would like your image to display across approximately 1/3 of this example screen, request that it be scanned in at 425 (approximately 1280/3) pixels in X rather than the default which is 600 pixels in X.
Note, however, that the size that the image appears on some particular display is not in general the same as the size the image will appear to be on a display with different characteristics or when printed on paper. If you are importing the image into a document preparation package, such as FrameMaker, you must specify the DPI at which to import the file. If you wish the image to be the same size on the printed page as it is on your X display, you should set the import DPI to match that of the X display. In the example given above, you would specify an import DPI of 100 to accomplish this.
By default, we give you your scanned images in the Boston University Computer Graphics Lab (BUCGL) format. This format supports 24 bit color and is generally a very good format to use, except that it is only used here at BU. The SDSC Image Tools, which we use to manipulate images support many different formats and we can give you your scanned images in any of them.
Be warned that choosing an image format is a complex decision. If you need the maximum color precision of 24 bits, don't use a format like GIF (.gif) which only supports up to 8 bit color. Also, certain formats have a number of different variations with the same name. For example, TIFF (.tiff) files generated by most IBM PC applications can not be understood by our image tools. However, the image tools do understand most TIFF files created in Unix applications. If you are unsure which format to use, ask the graphics consultant at the Graphics Lab for help. Be sure to describe, as best you can, what you will be doing with the image.
If you wish to have your image files scanned into a raster image file format other than the BUCGL format, you must indicate your preferred format on the work order you submit. If you are not sure which format to use, ask the graphics consultant on duty for help or just use the default BUCGL format. On the SCV supported machines, you can always convert your images to another format on your own using the SDSC image tool imconv.
The Computer Graphics lab offers a variety of hardcopy services. Please see the documentation on 35mm slides, the thermal color printer, the color laser printer, and the dye sublimation color printer for more details.