The Playfield#

If you come from almost any later machine, your instinct is a frame buffer: a grid of pixels at (x, y) that you poke to draw a background. The VCS playfield is not that, and the distance between what you expect and what you actually get is the whole story of this chapter.

There is no background bitmap. The playfield is three registers — PF0, PF1, PF2 — holding 20 bits that describe one chunky row of background, which the TIA paints in real time as the beam crosses each scanline. Those 20 bits stretch across a 40-pixel-wide screen, and each “pixel” is a virtual block 4 color clocks wide and one scanline tall. Nothing is stored; there are only register values the beam reads as it goes.

And those registers are quirky in three ways that trip up everyone:

• PF0 is half a register. Only its high four bits (D4–D7) are used; the low nibble is ignored entirely.
• The bit order isn’t uniform. PF0 and PF2 are read one direction, but PF1 is reversed — so a pattern that looks continuous in your source has to be laid out with PF1’s bits flipped.
• You only describe half the screen. The 20 bits cover the left 20 pixels; the right 20 are generated automatically, as a mirror or a copy.

Why so strange?#

Because the obvious design was impossible in 1977. A true (x, y) bitmap of even this coarse 40×192 screen would need about 960 bytes at one bit per pixel — and the VCS has 128 bytes of RAM, total. And there would be no time to use it anyway: at 76 CPU cycles per scanline, the processor cannot fetch and push a stream of pixels fast enough to fill a line. So Atari’s designers traded resolution for cost: a handful of register bits, set once per line, that the TIA expands into a blocky background essentially for free. Every quirk above is a fingerprint of that bargain — each one saved chips, memory, or cycles.

For all that, it remains the cheapest way to put graphics on screen — one set of writes describes an entire line — which is why backgrounds, mazes, borders, and large static shapes are almost always built from the playfield.

In this chapter#

• The Three Registers — exactly which bit of PF0/PF1/PF2 lights which pixel, with the half-nibble and reversed-PF1 quirks laid out in a diagram. (Left half only.)
• Symmetry: Reflection & Repetition — where the right half of the screen comes from: mirrored or copied from the left.
• The Asymmetric Playfield — escaping symmetry by rewriting the registers mid-scanline, with the cycle-by-cycle safe-write schedule.
• Building a Scoreboard — a worked example: digit graphics, bit-masking two digits into a byte, ×5 indexing, and BCD scorekeeping.

The playfield’s defining quirk is its bit order. Getting a shape to look right means accounting for PF0’s half-nibble and PF1 being reversed relative to PF0 and PF2.