◆ AIUNITES|

Live Vocal Resonance

Sing or speak a sustained vowel and VoiceStry listens through your mic, estimates your resonance peaks (formants) in real time, drops them onto the vowel space, and maps them to VRN. Your resonance is personal — vocal-tract length shifts everyone's numbers — so pick your voice type to calibrate the map.

Mic off. Click start and allow microphone access.
[ waiting for a voiced sound… ]
Pitch F0
Hz
F1 · peak 1
Hz
F2
Hz
F3
Hz
Squillo band
%

Where your voice sits

Grey = reference vowels for your selected voice type. Cyan = you, live. F2 → left/right, F1 → top/bottom.

Projection — formant tuning

When a vocal-tract resonance (a formant) lines up with a harmonic of the note you're singing, that harmonic rings — a loud, carrying tone for very little breath. Pick a pitch (or follow your own), sustain a vowel, and watch your formant bars snap onto the gold harmonic lines. Turn on the drone to match the pitch by ear (headphones recommended, so the mic hears you and not the tone), and hold a lock to build a streak.

Sustain a vowel to see formant–harmonic alignment.
Target F0
Hz
F1 → harmonic
F2 → harmonic
Output
Lock hold
[ projection reading appears here ]

Resonance, not force. You noticed it yourself — in a hard room (a nave, a glass-walled lobby) a sound can ring for almost no airflow. That's the room reflecting strongly while a tuned formant amplifies the tone: the resonance does the work, so you don't push air. An F1·H2-style tag below means a formant locked onto that harmonic.

How the reading becomes VRN

An acoustic estimate from your mic, in VRN (VSV c06). It reads the sound, not the sensation — a mirror, not a verdict.

[O]Oral — bright, forward, high F2
[H]Head — lifted, strong upper formants
[C]Chest — low, dark, low formants + F0
[Sq]Squillo — energy in the 2.8–3.2 kHz ring
[Ch+]/[Sc]Chiaro (bright) / Scuro (dark)
[F1↑][F2↓]Formant shift vs the target vowel
// example — a bright, forward /i/ might render as [O++, H++, Sq+, Ch+] "/i/ (heed)"

What the mic can and can't see: formants (F1–F3), pitch, and the squillo ring are directly measurable, so [O] [H] [C] [Sq] [Ch+] [Sc] [F1↑] [F2↓] are grounded. VRN's muscular / sensation symbols — [D] [Ap] [Zy] [Cov] [N] — can't be read from audio; they're the singer's felt input. This tool fills the acoustic half of a VRN line and leaves the body half to you.

Why it's different for every person

A formant is a resonance of your vocal tract, and resonance frequency scales with tract length. A longer tract (typically lower/male voices) pushes every formant down; a shorter tract (higher voices, children) pushes them up. The same vowel /i/ can read ~270 Hz F1 for a bass and ~440 Hz for a child — same vowel, different numbers. The relative layout of the vowel space is shared; the absolute Hz are yours. That's why the voice-type selector rescales the reference vowels — and why serious tools let you calibrate to your own voice.

Desktop apps that measure this for real (deeper than a browser demo): Praat (free, LPC formants)VoceVista Sing&SeeSygyt / Overtone Analyzer They use longer, cleaner analysis windows, so their numbers are steadier — but the method here (LPC spectral-envelope peak-picking) is the same idea.