Examples

Every waveform on this page is computed from a real recording by the library itself: the C++ core analyzed the audio, and the JS display functions turned the result into SVG paths when this site was built. The recordings are in the public domain (sources).

Voice message — 12 s of speech

waveformPresets.chat

useWaveform(url, { ...waveformPresets.chat, barCount: 40 })

peak metric

resample(peaks, 40, { metric: 'peak' })

decibel scale

normalize(resample(peaks, 40), { scale: 'db', reference })

raw levels

resample(peaks, 40) // not normalized

Reading — 63 s

Overview, waveformPresets.player

resample(peaks, 120) // waveformPresets.player

Seconds 20–30

resample(peaks, 120, { start: 20, end: 30 })

Music — 35 s

RMS

resample(peaks, 120)

Peak

resample(peaks, 120, { metric: 'peak' })

Signed envelope

normalizeMinMax(resampleMinMax(peaks, 360))

Signed envelope, seconds 10–12

resampleMinMax(peaks, 360, { start: 10, end: 12 })

Example apps

The repository contains a bare React Native app and an Expo app with the same demo: a chat bubble drawn with the chat preset, and a full waveform with every knob — bar count, metric, scale, reference and zoom — that never decodes again.