8-Bit vs 16-Bit What the bits actually meant

What a "bit" actually is

A bit is a binary digit — a single one or zero, the smallest unit a computer can hold. The "bits" on a console's box referred, at least in the honest early days, to the width of the number its processor could handle in a single operation. An 8-bit CPU works eight binary digits at a time; a 16-bit CPU works sixteen. (Our retro glossary unpacks this jargon and plenty more.)

Width compounds, because every extra digit doubles the range. Eight bits can express 256 different values — the numbers 0 to 255. Sixteen bits can express 65,536 — the numbers 0 to 65,535. That is not twice the room. It is 256 times the room.

Why does width matter? Because game code is arithmetic all the way down. Scores, coordinates, timers, physics — they are all just numbers being pushed around, and a processor that handles bigger numbers in one step spends fewer steps juggling. Wider registers usually came with wider data paths too, so more information moved per tick. But the CPU was never the whole machine: clock speed, dedicated graphics hardware and sound chips did most of the visible work — which is exactly where this story later goes wrong.

The same scene, twice

Here is what the jump looked like from the sofa. Both panels below show the same made-up landscape. The left one speaks strict 8-bit dialect — four flat colours, square sun, stepped hills. The right one says the same sentence with the vocabulary the 16-bit machines opened up: gradients, curves, glow.

The artists didn't get better overnight. The maths underneath them changed. When your palette holds a handful of colours, every hill is a decision about which two of them meet at a hard edge; when it holds hundreds, the edge itself can become a slope of in-between shades. Almost everything people remember as "16-bit graphics" — the sunsets, the parallax skies, the shading on a character's face — is that palette arithmetic made visible.

What actually changed

Put the four machines that defined the era side by side — Nintendo's pairing and Sega's — and the generation gap stops being a slogan and becomes a column of numbers.

Two details in that table repay a closer look. The first is sound. The NES synthesised its music live from five simple channels — two pulse waves, a triangle, a noise generator and a crude sample channel. The SNES instead shipped a Sony-designed SPC700 sound system playing eight channels of compressed samples — recordings, effectively — which is why its soundtracks suddenly had orchestral stabs, breathy flutes and slap bass where the previous generation had bleeps.

The second is that "16-bit" never meant "equal". The Mega Drive's 68000 ran at roughly twice the clock speed of the SNES's CPU, while the SNES answered with a far bigger palette and that sound chip. Same number on the box; very different machines underneath it. Which brings us to the trouble with the number.

Why "bits" became marketing

The moment bits started selling consoles, they stopped measuring them. Exhibit one: NEC's TurboGrafx-16, sold in Japan as the PC Engine. Its HuC6280 CPU is an 8-bit chip — a souped-up cousin of the NES's processor — bolted to a pair of genuinely 16-bit video chips. NEC named the console after the video half; critics called the label misleading, and when Sega launched the Genesis in America its advertising openly mocked NEC's claim to be the first 16-bit console.

Exhibit two arrived in 1993, when Atari marketed the Jaguar as the first 64-bit console under the slogan "Do the math". The maths in question involved 64-bit internal buses inside its custom "Tom" and "Jerry" chips — while the processors themselves executed 32-bit instructions and the machine's general-purpose CPU was a Motorola 68000, the same family as the Mega Drive's. The press did the math and laughed: Electronic Gaming Monthly quipped that by Atari's arithmetic the Sega Saturn would be a 112-bit machine.

That was really the end of it. Once the number could mean the CPU, the graphics chip, or a bus somewhere inside, it no longer told a buyer anything — and when CD-based machines arrived in the mid-1990s, the sales pitch moved on to polygon counts and disc capacity. The bit wars didn't end with a victor. They ended when the units stopped mapping to anything you could see.

Britain's 8-bit era ran on tape

One more wrinkle, and it matters if you grew up here: Britain mostly lived its 8-bit years on home computers, not consoles. The Sinclair ZX Spectrum launched in April 1982 at £125 for the 16K model and £175 for the 48K — pointedly undercutting the BBC Micro — and went on to sell more than five million units, making it one of the best-selling British computers ever built. Alongside the Commodore 64, it put the same class of 8-bit processor as the NES into millions of front rooms, fed by cassette tapes and pocket-money games from the newsagent. (Lee's own memory of the era survives in Wayne, Martin and the Indestructible Spectrum.)

So when the 16-bit consoles reached these shores — the Mega Drive arrived in September 1990 at £189.99 with Altered Beast in the box, distributed by Virgin Mastertronic — the generation gap felt different in the UK. We weren't upgrading from one console to another; we were watching the living-room television take over from the rubber-keyed computer on the desk. You can trace the whole handover year by year on our interactive timeline.

If this was your childhood — either side of the gap — Retro Delights is a members' club built around exactly these memories. Have a look at what membership includes.

8-bit vs 16-bit — FAQ

Is 16-bit twice as powerful as 8-bit?

No — the step is bigger than that. Each extra bit doubles the range a processor can handle in one operation, so 8 bits give you 256 values while 16 bits give you 65,536. In practice, though, most of what players noticed came from the hardware around the CPU: bigger colour palettes, more sprites, dedicated sound chips and more memory.

Why did the SNES look so much better than the NES?

Mostly colour maths. The NES could show about 25 colours at once from a hardware palette of roughly 54, while the SNES could hold 256 palette entries at once drawn from 32,768 possible colours. It could also throw around 128 sprites to the NES's 64, and its Sony-designed SPC700 sound system played eight channels of sampled audio rather than synthesised bleeps.

Was the Mega Drive really 16-bit?

Yes — more honestly than some rivals. Its Motorola 68000 is a genuine 16/32-bit processor clocked at about 7.6 MHz, noticeably quicker than the SNES CPU's roughly 3.58 MHz. It even carries a second, 8-bit Z80 processor alongside, used for sound duties and for backwards compatibility with the Master System.

What were 8-bit computers?

In Britain, the 8-bit era largely happened on home computers rather than consoles. Machines like the Sinclair ZX Spectrum — launched in April 1982 at £125 for the 16K model — and the Commodore 64 used the same class of 8-bit processor as the NES and Master System, but you loaded games from cassette tape and could program them yourself.

When did the bit wars end?

When the number stopped describing anything you could see. Atari marketed the Jaguar as 64-bit in 1993 on the strength of its internal buses, while its main CPU was from the same 68000 family as the Mega Drive's — and the gaming press openly mocked the arithmetic. Once CD-based machines arrived in the mid-1990s, the conversation moved on to polygons and disc space, and bit counts quietly fell off the box.

Keep exploring

The era either side of this gap is mapped across the site: the golden age of the arcades covers the coin-op world these consoles were chasing, the timeline walks the release history year by year, and the glossary decodes the rest of the jargon.

More from the guides

• Retro game values
• CRT retro gaming
• Most expensive games
• The kill screen