How DOS Games Drew Their Worlds

How DOS games drew their worlds

The appearance of a DOS game reflects both the display hardware and the choices made within its limits. A small pixel grid gives a character little room for detail. A restricted palette makes each color useful in several places. Repeated scenery can turn a handful of pictures into a long level. These constraints help explain the blocky figures, patterned shadows and clear silhouettes that remain recognisable on a much larger screen.

There was no single DOS look. Graphics standards changed, and each supported several modes. Even games using the same resolution could look very different because of their palettes, animation and drawing style.

CGA, EGA and VGA

Three IBM graphics standards provide a useful outline of that change. The table gives one representative mode for each, rather than every option a game could use. The EGA row uses its higher-resolution mode because the available color choices depend on the mode and display.

CGA: Introduced: 1981; Example resolution: 320 by 200; Colors in that mode: 4, with restricted palette choices
EGA: Introduced: 1984; Example resolution: 640 by 350; Colors in that mode: 16 selected from 64 on an EGA display
VGA: Introduced: 1987; Example resolution: 320 by 200; Colors in that mode: 256 selected from 262,144

CGA's familiar 320 by 200 mode used four colors at once. Common combinations included cyan, magenta and white, or green, red and brown, with a background often set to black. Intensity and display differences could change the result. Its 640 by 200 graphics mode used two colors. Strong outlines and alternating dot patterns helped artists suggest shapes and shades within those limits.

Composite output could produce another appearance. On a compatible color television or composite monitor, fine CGA patterns could create additional apparent hues through the way the color signal was decoded. The pattern, signal and display affected the result. Those hues were not extra independent palette entries available in the ordinary four-color RGB display mode.

EGA expanded the range. At 640 by 350 with an EGA display, sixteen colors could be chosen from sixty-four. Many action games instead used 320 by 200 with the familiar sixteen-color set. It is misleading to assign every EGA mode the same freedom to choose colors. The screen mode was part of the drawing decision.

VGA's common 320 by 200 mode allowed 256 colors at once. Each pixel referred to an entry in a palette; the standard VGA palette offered 64 intensity steps for each of red, green and blue. Multiplying those choices gives 262,144 possible mixes. The game could choose 256 of them for the screen, allowing smoother ramps for metal, fire, water or a shaded face.

VGA also supported 640 by 480 with sixteen colors. A higher resolution meant more pixels to draw and store, so the choice involved more than sharpness. Plenty of action games used the smaller grid, but fast action was not confined to it.

Why a 320 by 200 picture can change shape

A 320 by 200 image contains 64,000 pixels. Draw each as a square and the image has a width-to-height ratio of 8:5. That is wider than the 4:3 display area commonly used on the CRT monitors of the time.

To fill a 4:3 area, that image needs relatively taller pixels. Keeping the width at 320 square pixels makes the corresponding height 240, so the vertical stretch is six fifths. This is the basis of a common aspect-ratio correction. A circle drawn to look round on the stretched display can look squat in an uncorrected screenshot.

Not every game or modern viewer makes the same choice. A browser can preserve square pixels, reproduce the older display shape, or let a player choose. Electro Man's Classic view on DOSHD uses whole-number enlargement of its original 320 by 200 picture, keeping those displayed pixels square.

Standard VGA output in the common 256-color mode repeated each image row over two scan lines and refreshed at about 70 times a second. Row repetition and aspect correction describe different things: one concerns the signal sent to the monitor, the other concerns the proportions of the displayed picture. Neither tells you how often a game's rules advance.

Tiles and sprites make a larger world

Many games stored reusable tiles and a map of where to place them. A tile might be a brick, ladder or patch of ground. Drawing the same tile in fifty places needs one stored picture plus placement information, instead of fifty separate copies of the picture. Tile sizes varied; 8 by 8 and 16 by 16 pixels were common choices, not universal rules.

A sprite is a separate image used for something such as a character, enemy or pickup. A common drawing method places scenery first and sprites over it, leaving selected pixels transparent so the background remains visible. Games could implement that transparency in different ways. It was not a special transparent color supplied automatically by VGA.

Animation uses a sequence of pictures. Short sequences conserve memory and drawing work, though memory alone did not dictate their length. A few well-chosen poses can make movement readable: a crouch, an upward stretch and a landing tell the player more than extra detail on a face.

A palette is a shared set of choices

In a typical indexed-color screen, scenery, characters and status displays use the same palette. Reserving many entries for a sunset leaves fewer distinct entries for other objects. A color can also be reused across unrelated pictures, which helps give a scene a consistent appearance.

Changing a palette entry changes every pixel that refers to it. Games could use palette changes for fades or animation without redrawing every affected pixel. Dithering offered another tool: alternating colors in a pattern suggests an intermediate shade, especially when viewed from a distance.

On DOSHD, Classic keeps these visual decisions available for comparison. HD redraws a scene with more detail, where that view is offered. The HD article explains why sharper art still needs to respect the ledges, gaps and hazards underneath it.