Gaussian Noise Generator

What it does

The Gaussian Noise Generator draws reproducible samples from a normal distribution with the mean and standard deviation you set — up to 10,000 values per run. It’s the workhorse for adding realistic noise to signals, simulations, and test data: most natural measurement error is approximately Gaussian, which is why this is the default noise model almost everywhere.

Common use cases

• Noising clean data — add measurement error to synthetic sensor readings, prices, or coordinates so test data stops looking suspiciously perfect.
• Signal processing tests — known noise into filters, smoothers, and denoisers, so effectiveness is measurable.
• Simulation inputs — perturbations for Monte Carlo runs and robustness checks.
• Teaching — the 68–95–99.7 rule, visible in a generated sample.

Settings

• Samples — 1 to 10,000 values, exportable as CSV, JSON, or text.
• Mean — the center of the distribution (0 for pure noise to add onto a signal).
• Standard deviation — the noise amplitude; ~68% of values fall within one σ of the mean.
• Seed — identical seed + parameters = identical noise, so noise-sensitive tests stay deterministic.

Privacy note

Samples are computed locally in your browser and never uploaded. The output is synthetic mathematical noise — not measurements of anything real.

FAQ

How do I “add noise” to my data? Generate with mean 0 and your chosen σ, then add the values elementwise to your clean series. σ controls how corrupted the result looks.

Why Gaussian instead of uniform noise? The central limit theorem: sums of many small independent errors trend normal, so Gaussian noise mimics real-world measurement error. Uniform noise (see White Noise) is better for dithering and some DSP cases.

Is the sample exactly normal? It’s pseudorandom draws from a normal distribution — small samples will wobble, large ones converge. For other distributions, use the Distribution Sampler.