gaussian01

gaussian01 is a small AI-generated visual demo about Gaussian distributions, covariance, and diffusion, built with C and Raylib.

This project is intentionally focused on visual presentation and interactive explanation rather than code quality.

gaussian.mov

What It Does

It presents a three-page interactive atlas of Gaussian-related ideas, starting on page 3 by default. You can pan, zoom, inspect values, drag interactive controls, and switch between overview and deeper geometry / diffusion views.

What It Visualises

• A progression from Gaussian PDF sampling and histogram convergence through addition, covariance, Brownian motion, higher-dimensional intuition, the central limit theorem, and diffusion
• A second page of interactive Gaussian geometry panels covering mean vectors, distance geometry, Gaussian hills, multiple means, covariance, 3D ellipsoids, dimensional growth, Gaussian operations, and correlation
• A third page focused on Mahalanobis distance, whitening, log density, score fields, Gaussian noise walks, and reverse diffusion
• An inspector panel for hovered interactive regions
• A movable minimap for navigating the current page
• Live sliders, draggable probes, and play / pause toggles inside the atlas panels

Install

You need a C compiler, make, pkg-config, and Raylib installed.

On macOS with Homebrew:

brew install raylib pkg-config
make run

Controls

• Space: pause or resume the animated simulations
• R: reset the camera and all simulations
• 1 / 2 / 3: switch pages
• Q: quit
• Mouse wheel: zoom
• Left click and drag on the world: pan the camera
• Middle click and drag: pan the camera
• Left click on tabs: switch pages
• Left click and drag sliders / handles inside atlas panels: adjust parameters
• Left click the minimap body: jump the camera
• Drag the minimap header: move the minimap