Skip to content
Cymatics.Open the app

Understand interference patterns in voices and music

By Cimatica · Updated

A voice, a gong and a concert can all contain many frequencies. The useful question is which parts remain stable enough to follow, not whether a beautiful image proves the sound is music.

Interference, standing waves and beats

When waves overlap, their displacements add. Depending on phase, they reinforce or partly cancel. A Chladni plate reveals nodal lines of its own vibration; a microphone at one position does not measure that entire spatial pattern.

Two nearby steady tones can produce audible beats: 440 Hz and 442 Hz have a 2 Hz difference, so their combined loudness pulses roughly twice per second. This is different from tempo, and different from the geometric pattern chosen by a visualizer.

One voice can produce many peaks

A sung vowel contains a fundamental and harmonics. Peaks at multiples of a pitch can belong to the same source. Counting every peak as a separate note can turn one singer into an apparent chord.

Ordinary speech also contains breaths, consonants and moving pitch. Brief patterns are expected. A sustained sung vowel is a better single-note test than a sentence; singing remains music even when only one voice is present.

Concerts are not just stable chords

A live performance may contain drums, applause, distorted instruments, overlapping notes and room echoes. Some musical passages are mostly percussive or noisy, while a fan can produce a stable tonal hum. Tonality alone cannot identify a concert.

Cimatica's activity feedback describes acoustic evidence and uncertainty. It is not a semantic music-recognition service and does not identify an artist, song or individual speaker. Use the input context you know to interpret what the display follows.

Read a changing figure

Look for repeated structure during a held note, then watch what changes when another instrument enters. A stable shape suggests a sustained match to the model; a dispersing shape can mean silence, changing pitch, competing tones or insufficient confidence.

Several detected tones can blend into a composite figure. The blend is an artistic mathematical interpretation, not a reconstruction of the pressure field, phases or geometry of a physical concert hall.

A short listening experiment

First hum a steady note. Next speak a sentence, then clap once. Finally play a short musical passage at a comfortable level. Keep the microphone position fixed and compare sustained structure with fast-changing activity.

For interference itself, compare one steady tone with two nearby tones from a suitable source. Do not use the animation alone to infer beat frequency: listen to the pulsation or use a waveform analyzer when a numerical measurement matters.

Frequently asked questions

Can the app separate a singer from background music?

No. A microphone captures their mixture. Activity estimates can help describe that mixture, but they do not isolate a voice or an instrument into separate audio tracks.

Does an unstable pattern mean the concert is noise?

No. Percussion, short notes, multiple performers and echoes may all produce unstable patterns. A detector's uncertainty is not a judgement about the performance.

Are these the exact interference figures in the room?

No. They are simulated patterns driven by the sound at the microphone. A single microphone cannot recover the room's full spatial sound field.

Technical references

Keep reading

Sound interference patterns: voices, noise and live music | Cimatica