About

Representations all the way down.

sinusoidal.space is an independent technical notebook about signals, models and the computation joining them.

The editorial idea

A useful explanation should reveal what representation an algorithm receives, what information it retains, what it discards, and what the mathematical notation becomes in memory. Interactive figures here are arguments you can test, not decoration.

Standards

Examples expose assumptions and units. Numerical claims are tested against analytical cases. Performance results appear only when they have actually been measured with a disclosed method and environment. Analogies between signal processing and machine learning are useful only when their limits are explicit.

Colophon

The site is statically generated with Astro. Its diagrams are SVG, its interactive numerical layer is plain TypeScript, and pages ship no application framework.