Introduction

This workshop introduces linguists to some basic concepts related to using computer graphics for professional purposes. The purpose is both to introduce you to general concepts useful in a wide range of situations and applications, and to give instructions for specific tasks that you're likely to need to perform in teaching and presenting your research.

Concepts

There are two types of graphics that computers know about: "bitmap" or "raster" graphics, usually created and edited in "paint" or "photo" programs, and "vector" graphics, usually created and edited in "draw" or "illustration" programs.

Raster Graphics

A bitmap graphic is composed of a grid of "pixels", which correspond to the pixels (dots) that make up the image on your computer screen or on a printer. Each pixel is assigned a color.

A 16-color bitmap image, and a region enlarged

Graphics acquired with a scanner, a digital camera, or by capturing part of a computer screen will always be bitmaps. Editing a bitmap consists of changing the color of selected pixels. Windows provides a simple bitmap-editing program ("Paint", found in the "Accessories" folder); other popular programs which provide much more power, flexibility and creativity include Adobe Photoshop, Corel Photo-paint, and many others.

The amount of memory it takes to store or display a bitmap is determined by two things: its size (width times height, in pixels) and its "color depth" (the number of colors available for each pixel). Color depths correspond to powers of two, since they correspond to the number of bits (0s and 1s) the computer uses to store the pixel; common choices are 2 (black and white: 1 bit), 16 (4 bit), 256 (8 bit), 16,777,216 (24 bit). Different bitmap editing programs use different file formats and extensions, but the most common are ".bmp", ".jpg" or ".jpeg", and ".gif".

Since a bitmap image is composed of a fixed number of pixels, its apparent size is determined by the resolution of the device you use to display it. Depending on display settings, a computer screen has about 72 pixels per inch; a printer has about 300-1200. The number of colors that appear in the image also depend on the device it is displayed on; if your computer display is set to 16 colors, a 24-bit image will display with only 16 colors. And different printers and monitors will display the "same" colors differently. Thus the same picture can look very different on different computers and printers. (To change the color depth and resolution of your monitor on a Windows system, right-click the desktop, choose "Properties", and choose the "Settings" tab.)

In general, bitmaps take a lot of memory to display. This means that working with large bitmaps can result in very large file sizes; they can also slow down your computer and even cause it to crash. It is possible to compress bitmap graphics so they take up less space. The most common compressed file formats are ".gif" (compresses 256 color images without data loss) and ".jpg" (gives very high compression ratios with some loss of detail; you can specify the amount of compression you want). For comparisons, see Web for Linguists.

Vector Graphics

With vector graphics (also called "metafiles"), the computer stores information about "objects" such as lines, curves, rectangles, and ellipses, and their colors, sizes, positions, and fills. You edit a vector object by changing these properties. (A metafile can also contain other types of objects, like text and bitmaps; this is why it is called a "metafile". We'll see some examples below.) Vector drawings are easier to modify than bitmap graphics, because the program knows about the objects you draw. You can easily draw a line and then change its endpoint, for instance. In a bitmap program you can draw a line, but this simply has the effect of changing the colors of a collection of pixels; you can't later go back and change the line, because it isn't an object.

Vector drawings are edited using "draw" programs. Microsoft Office applications contain some simple vector drawing tools. Other popular drawing programs include Visio (good for creating technical drawings and diagrams), Corel Draw (good for "artistic" drawing and illustration), Adobe Illustrator, Micrografx Draw, and many others. Excel graphs are basically vector drawings. Vector drawings can't be displayed in Web pages without special browser plugins.

Because a vector drawing only contains information about the objects it contains, the amount of memory it requires depends on how complex the drawing is, but not the size or number of colors. In fact, size is irrelevant to a vector drawing; it can be scaled up or down without affecting the contents of the drawing or its size in memory.

Figure 2: A vector drawing with 48 objects, and a region enlarged

Different vector editing programs use different file extensions, but the most common one is ".wmf" ("Windows Metafile"). Most "clip art" consists of vector drawings.

Vector graphics can easily be converted to bitmap graphics (in fact, displaying a vector graphic on the screen or printing it implies this conversion), but the reverse is not true. However, there are programs (called "trace" programs) which perform this function; they work by finding regions in a bitmap with similar colors, and translating them into vector objects.

For illustrations in linguistics papers, you will usually want vector drawings, unless you want to include a photograph, a still from a video, or a screen capture image from the computer screen.

Compound Documents

Linguists will usually want to use graphics in conjunction with text, as part of a paper, a handout, or a set of overheads. This involves creating a "compound document", i.e. one which contains data in more than one format. Windows contains some very rich tools for creating and editing compound documents; the aim is to make several different programs act as though they were a single program. This is achieved using a technology called "Object Linking and Embedding", or "OLE". Applications can be designed to be either OLE clients - that is, they can accept objects from another program - or OLE servers - that is, they can provide objects which can be inserted into another program - or both. Microsoft Word and Excel can be both clients and servers; Paint can be a server but not a client. Microsoft Draw, Graph and WordArt are "server-only" programs - they can act as servers but can't run by themselves!

In this document we will be concerned with the most common situation, where you are using Word as a client application and other applications as servers: that is, pasting data from other programs into a paper or handout.

Note: while most of the procedures described here are available in all recent versions of Word (and in most cases other Office programs), the interface (how you access the commands) and terminology has changed quite a bit from one version to the next. In particular, more recent versions make Word's handling of images more similar to the options in HTML. The description here is for Word 2000; in other versions you may have to poke around to find the equivalent commands.