Blue Noise

Poisson-disk beside uniform random — the difference is the lesson.

patterns.poisson · v1 · renderer canvas

Blue noise (left) vs uniform random (right) · r = 17.0 px · 429 × 572 per panel

Both panels hold the same number of points and both are random. The left one keeps a minimum separation of 17.0 px; the right one keeps no promise at all, which is why it clumps and leaves holes. The clumping is not a fault in the randomness — independent points clump, and a scatter without clumps would be the suspicious one. It is simply that "random" and "evenly spread" are different things, and most of the time what people want is the second.

[data-canvas-stage] { display: none; }

Drawn in the browser

This generator does not send you an image. The server sends a short description of one — the parameters below, plus a 32-byte key — and the drawing is done here, from the same byte stream the server would have used. That needs JavaScript, so with it switched off there is nothing to show.

The description itself is still public: /api/v1/g/patterns.poisson returns it, and the maths is in resources/js.

Blue noise (left) vs uniform random (right) · r = 17.0 px · 429 × 572 per panel

Controls · from schema()

Width

200 2048

Height

200 2048

Show

Minimum separation

5 90

In pixels. No two blue-noise points are ever closer than this; the uniform panel gets the same number of points and no such promise.

Candidates per point

4 60

Bridson's k. Each accepted point gets this many attempts at a neighbour before it is retired. Low values leave gaps.

Dot size

0.6 12

Colour dots by

The measurement that separates the two panels: blue noise has almost none of it, uniform random has all of it.

Colour

Palette size

2 12

Entropy source

Picking a source changes the story on the receipt, never the quality: anything below class A is mixed with the OS CSPRNG before it reaches you. This result actually came from NIST Randomness Beacon.

seed token FR3ZKHRHHHD300F4R240JY2ZPG

Re-render keeps this token. Generate replaces it. That is the whole idea.

Burned into the corner before the file is written, so the image keeps its provenance wherever it ends up:

entropy in

16 bytes

stream used

4 bytes

rejections

duration

353 µs

palette

estimated points

706

guarantee

No two blue-noise points closer than 17.0 px.

cell size

12.02

Seed receipt

Class B · Public beacon

NIST Beacon pulse #2,005,730 — chained to every pulse before it, and signed.

Published openly — anyone can look this value up after the fact.

source: NIST Randomness Beacon
observed: 2026-10-30T11:12:00+00:00
latency: 116ms
mixed with: csprng — always

https://beacon.nist.gov/beacon/2.0/chain/2/pulse/2005730 ↗

Permalink

The token is the seed, so this link recomputes the result rather than looking it up. No database, and it works for as long as v1 does.

This page and /api/v1/g/patterns.poisson run the same code. There is no private endpoint.

Blue Noise

Result

Measurements

Seed receipt

Permalink