Perlin Noise Visualizer
Generate and inspect Perlin gradient noise with adjustable octaves, persistence and lacunarity, and see the measured value distribution against the theoretical bounds.
Last reviewed by the Radiatus Cloud team
Need this done properly for your business?
Radiatus delivers secure cloud, DevOps & compliance engineering.
Gradient noise is smooth because it interpolates slopes, not values
Perlin noise assigns a random gradient vector to every point on an integer lattice and asks, for any position in between, what each surrounding gradient would predict. Those predictions are blended with a fade curve that has zero first and second derivative at the endpoints, which is what makes the result look like rolling terrain rather than a lattice of bumps. A direct consequence is that the noise is exactly zero at every lattice point, since a gradient predicts no change at its own location.
Octaves turn one smooth field into landscape
A single octave is too smooth to look natural. Fractal Brownian motion adds copies at doubling frequency and shrinking amplitude, so broad hills come from the first octave and boulders from the last. Persistence sets how quickly amplitude falls, and it is the main control over the character of the result: below about 0.4 gives smooth hills, around 0.5 gives realistic terrain, and above 0.7 gives noisy crags because the fine detail never fades out.
The value range is bounded and worth knowing
With unit gradient vectors, two-dimensional Perlin noise cannot exceed the square root of two over two, roughly 0.707. How close a given view comes to that depends on how much of the lattice it covers: an image spanning a handful of cells typically peaks near two thirds of the bound, while one spanning a few thousand reaches it almost every time. This tool measures the actual minimum, maximum, mean and standard deviation rather than assuming them, which matters when you map noise onto terrain heights or thresholds and need to know where the water line will fall.
Related tools
- Data Collection Analysis — Analyze app description to infer data collection.
- Dark Pattern Detector — Scan UX text for manipulative patterns.
- AI Risk Disclosure — Generate disclosure text for AI features.
- Maturity Radar — Generate a radar chart of security maturity.
Frequently Asked Questions
Why is Perlin noise zero at integer coordinates?
Because each lattice point contributes a gradient dotted with the offset from itself, and at its own location that offset is zero. Every corner contributes nothing, so the interpolated result is exactly zero.
What does persistence control?
How fast the amplitude of each successive octave falls. Below 0.4 gives smooth hills, around 0.5 looks like natural terrain, and above 0.7 keeps fine detail loud enough to look noisy.
What is lacunarity?
The frequency multiplier between octaves, normally 2. Raising it spreads the octaves further apart in scale, which leaves visible gaps between the coarse and fine structure.
What is the maximum value Perlin noise can reach?
With unit gradients, two-dimensional Perlin noise is bounded by the square root of two over two, about 0.707. How close a real field gets depends on how many lattice cells you sample: a few cells peaks near two thirds of the bound, a few thousand reaches it. This tool reports the measured range and the cells in view.
How is this different from value noise?
Value noise interpolates random values placed at lattice points and shows visible grid artefacts. Perlin interpolates random gradients, which removes the axis-aligned bias and is why it became the standard for procedural terrain.
Privacy & Security
Everything runs in your browser; nothing is uploaded.
How to Use
Adjust the scale and octaves, then generate.
Disclaimer: This tool is provided "as is" without warranty of any kind. Results are for educational and utility purposes.