Recording & signals / Side by side

Aliasing and clipping: two different limits

One loses frequency identity during sampling. The other changes amplitudes that exceed a permitted range. Diagnose them separately.

Too few time samples

Sampling records a signal at discrete instants. If frequency content is not restricted below half the sample rate before sampling, different continuous signals can produce the same sample sequence. That ambiguity is aliasing. A higher-frequency input can appear as a lower-frequency component after reconstruction.

For an original calculation, sample a 900 Hz sine wave at 1200 samples per second. It lies above the 600 Hz Nyquist frequency. Its samples correspond to a 300 Hz sine with an appropriate phase or sign relationship. Turning down the input level does not solve that timing ambiguity.

Evidence: 1

Too much amplitude

Clipping occurs when values exceed a system’s allowed range and are limited at a boundary. A sine wave with requested peaks of ±1.4, hard-limited to ±1, acquires flat portions. Its shape is no longer sinusoidal and its frequency content changes. Reducing level before the limiting stage can prevent this distortion.

Increasing sample rate does not, by itself, create more amplitude headroom. Conversely, a perfectly unclipped signal can still alias. The remedies address different causes: control bandwidth and sampling for aliasing; control level and headroom for clipping.

Evidence: 2

Why the problems sometimes meet

A nonlinear process such as clipping creates additional frequency components. In a digital system, newly generated components above Nyquist can then alias unless the processing handles them appropriately. This is one reason some effects use oversampling. It does not mean aliasing and clipping are the same operation.

In the aliasing lab, the low sample rate is an explicitly selected teaching model. The browser still plays audio through its actual device rate. The plotted samples and calculated folded frequency demonstrate the ambiguity; they do not measure your audio hardware or certify a recording format’s quality.

Evidence: 1 · 2 · 3

DiagnosisAliasingClipping
Limit exceededRepresentable frequency bandAllowed amplitude range
Example900 Hz sampled at 1200 HzPeaks ±1.4 limited to ±1
Turning down firstDoes not remove frequency ambiguityCan avoid overload
Increasing rate aloneCan expand the available bandDoes not create amplitude headroom
Can occur together?YesYes
Keep this in mind

Sampling rate controls a timing constraint; headroom controls an amplitude constraint.

Make the connection

When samples tell another story

Set a frequency above half the model sample rate. Predict the folded frequency, then compare the displayed value and sampled points.

Open the experiment ↗

Sources & further reading

These sources support the explanations and model assumptions. Recording credits are listed in the listening panels and the source register.

  1. Basic concepts behind Web Audio API ↗MDN contributors / Mozilla · 2026-11-11
  2. Peak Clipping — Handbook for Acoustic Ecology ↗Barry Truax, editor / Simon Fraser University · Undated
  3. Sound Synthesis — Handbook for Acoustic Ecology ↗Barry Truax, editor / Simon Fraser University · Undated
Full source and rights register ↗