Evidence-led voice reference

Methodology

What the detector measures, what it rejects, how the classifier ranks matches, and where its evidence stops.

Voice-type references

Sustained-note measurement

Measure the range you can hold

Audio stays on this device

Current note
Microphone is offClarity gate ready
ReadyYour audio stays on this device.Microphone is off
Confirmed low Confirmed high

The browser analyzes small microphone buffers locally with a McLeod-style normalized autocorrelation detector. Only clear, stable notes held for at least 300 milliseconds can alter the final range.

Measurement pipeline

  1. Detect. A 2,048-sample microphone buffer is analyzed with a McLeod-style normalized autocorrelation method.
  2. Reject. The detector skips the central correlation lobe, requires high clarity, and gates low-energy frames.
  3. Confirm. A pitch must be stable for at least 300 ms before it can change an endpoint.
  4. Classify cautiously. Range overlap and center proximity rank broad categories; close scores are labeled borderline.

Accuracy boundaries

Microphone quality, room noise, vibrato, vocal fry, register transitions, and device audio processing can affect readings. The classifier does not measure timbre, passaggio, vocal weight, or comfortable tessitura over repertoire.

Editorial review status

Not yet independently reviewed by a named vocal coach or speech-language pathologist. No reviewer identity is claimed in the page or structured data.

Technical source

Philip McLeod, “Fast, Accurate Pitch Detection Tools for Music Analysis,” University of Otago (2008).

References

Ranges are approximate conventions, not biological boundaries.

Questions and limitations

Has this content been reviewed by a vocal coach or speech-language pathologist?

Not yet. No named clinical or pedagogical reviewer is claimed, and no reviewedBy structured data is emitted. The implementation cites its algorithm and range references, documents limitations, and should receive independent expert review before making stronger pedagogical claims.

Is this a medical or diagnostic test?

No. It estimates sustained pitch range and compares it with broad reference categories. It cannot assess vocal health, diagnose a condition, or replace a speech-language pathologist, laryngologist, or qualified singing teacher.

Before the browser asks

Your microphone stays local

The browser sends short audio buffers to this page’s pitch detector. Audio is not recorded, uploaded, or used for statistics.