The Sampling Theorem: Hearing Aliasing

In this idealized model, a digital recorder stores uniformly spaced samples of a sound wave, taken fs times per second. The Nyquist–Shannon sampling theorem says that the signal can be reconstructed exactly when all its frequencies are below fs/2.

1. The signal and the sampler

tone f1 000 Hz
Nyquist frequency fs/24 000 Hz
samples per period fs/f8.0
ideal reconstruction1 000 Hz

2. Time domain — the samples and what they reconstruct

original tone f samples at spacing 1/fs reconstruction (sinc/ideal low-pass)

3. Listen

sweep position:

Audio uses the Web Audio API — works in any modern browser. No data leaves your device.

Back to Collection