ben ebsworth
Generative Art

Flow Field

Particles advected by a value-noise vector field.

knobs

How it works

A field of particles drifts across the canvas. Each frame, every particle samples a deterministic value-noise field at its current position to get a flow angle (noise(x·scale, y·scale + time) · 2π), then steps one unit in that direction. Because neighbouring points sample nearby noise values, the particles fall into smooth, river-like streams. As with the orbits, we never fully clear the canvas — a translucent fill each frame leaves a fading trail, so the streams read as flowing lines rather than dots. Lower Trail = longer streaks.

Knobs

  • Particles — how many points are advected.
  • Noise scale — zoom of the underlying field; small values = broad, sweeping currents, large values = tight eddies.
  • Speed — step length per frame.
  • Trail — per-frame fade; lower = longer streaks.
  • Color — line colour.