A coastline, a cloud, a mountain ridge — none of them are tidy, yet none look like television static either. They're random in a softer way. That softer randomness has a name: noise, and it's the secret ingredient behind almost everything that looks natural.
Plain computer randomness jumps around wildly — every value ignores its neighbors. It looks like static, and our eyes read it as fake.
Noise is different: each value stays close to the one beside it, so the line rolls instead of jumping. Slide from one to the other and watch it calm down.
A single layer of noise has exactly two knobs. Frequency sets how stretched it is — low for broad, lazy swells; high for tight, busy ripples. Amplitude sets how far it climbs.
These two dials are the whole vocabulary — everything richer is built by combining them.
Now combine those dials. Take a big slow layer, add a copy at double the frequency and half the amplitude, then another, and another. Each is an octave — the faint lines are the layers, the bold line is their sum.
Watch smooth swells gain rocks and ridges as the layers pile on. This fractal stacking is where noise gets its natural detail.
Do the same thing in two dimensions and the rolling line becomes a rolling surface — which, depending on how you colour it, is either drifting cloud or a map of hills and valleys.
It even drifts on its own, because we can slide through noise in a third direction: time.
Here's the beautiful part, and a callback to Chapter One: feed noise into an angle at every point and you've built a vector field — one that nobody had to design. Set specks loose on it and they trace the hidden weather of the noise.