Vocal Analyzer
How to use the resonance test
- Read the text in your usual voice and press Stop. The first take is saved on this device as your starting point.
- Read the same text again with the resonance you are practising — for example a brighter, “smaller” sound — and press Stop.
- Read the result: how many percent higher (brighter) or lower (darker) your resonance sits than at the starting point, and your pitch now and then.
- Keep everything else the same: the same device and microphone, the same distance, a quiet room. A different phone or headset is a new starting point — press Make this my starting point.
The comparison needs at least 5 seconds of voice, so read the whole text at a relaxed pace.
What “resonance” means here
Your vocal folds set the pitch; the throat and mouth above them shape the sound into a pattern of peaks called formants. A shorter or narrower vocal tract — or a raised larynx, a smile, a forward tongue — moves those peaks up, and the voice sounds brighter and “smaller”. This is what voice trainers usually mean by resonance, and listeners use it alongside pitch to judge whether a voice sounds male or female (Hillenbrand & Clark, 2009).
The test averages the shape of your spectrum over every voiced moment of a reading, with the pitch harmonics smoothed away, and finds how far that shape slid up or down in frequency since your starting point. Pitch is measured separately, so you can see when one changed without the other — for example a higher pitch with unchanged, dark resonance.
Why it only compares you with yourself
There is no score for how feminine or masculine your resonance is, on purpose. Microphones, rooms and the shape of each person’s vocal tract change the spectrum as much as training does, so one reading on its own cannot be placed against other people. Two readings by the same person, on the same device, of the same text, cancel most of that out — what is left is mostly what you changed.
How reliable it is
- Pitch does not leak in. On 80 real recordings with the pitch moved 4 semitones up or down and the resonance left alone, the measure moved less than 0.5%.
- It follows real changes, but understates them. With the formants moved 10% up or down and the pitch left alone, it moved +6.4% and −5.7%. Read the percentage as a direction and a rough size, not a calibrated amount.
- Take-to-take wobble. Two ordinary readings by the same person (40 speakers, two different texts each) differed by 1.2% typically and by less than 3.4% nine times out of ten. Inside ±3.4% the result says “about the same”.
- Not yet checked against people deliberately changing their resonance, which is what this tool is for. That is why it is marked experimental — tell us whether the direction matched what you heard.
Train safely
Resonance work should not hurt. Stop if your throat feels tight or your voice goes hoarse, and rest. A speech-language pathologist with gender-affirming voice experience is the best guide. This page was written by the site’s developer from the sources below; it has not been reviewed by a clinician and is not a training programme.
What stays on your device
Your audio is analysed inside your browser and is never uploaded — this site has no server that could receive it. The starting point is kept in this browser’s local storage as 160 numbers describing the averaged spectrum shape, plus its median pitch and date — not audio, and the speech cannot be rebuilt from it. Clear starting point removes it; clearing your browser data does too.
Frequently asked questions
What is voice resonance?
The way your throat and mouth shape the sound your vocal folds make. It sets where the formants — the peaks of the voice’s spectrum — sit. Higher formants sound brighter and “smaller”, lower ones darker and “bigger”.
Why does my voice still sound male at a higher pitch?
Pitch is only one cue. Listeners also use resonance, and a raised pitch with unchanged, dark resonance is often still heard as male (Hillenbrand & Clark, 2009). This test shows the two separately.
Can this test tell me if my resonance is feminine?
No. It only compares your own readings on the same device. Microphones and individual anatomy shift the numbers too much for a fair comparison with other people or an average.
Why do I need to read the same text each time?
Different words have different vowels, and vowels move the formants. Reading the same text keeps that constant, so the difference you see comes from how you shaped your voice.
Is my recording uploaded or stored?
No. The audio is analysed in your browser and never leaves your device. Only the starting point’s summary — 160 numbers describing the spectrum shape, the median pitch and the date — is kept in this browser.
Sources
- J. M. Hillenbrand & M. J. Clark, “The role of f0 and formant frequencies in distinguishing the voices of men and women”, Attention, Perception, & Psychophysics 71(5), 2009. doi:10.3758/APP.71.5.1150
- Resonance checks: 80 read-aloud recordings by 40 speakers from LibriSpeech (V. Panayotov et al., ICASSP 2015; CC BY 4.0), with pitch and resonance changed separately in Praat (P. Boersma & D. Weenink) — details on the about page.
- A. de Cheveigné & H. Kawahara, “YIN, a fundamental frequency estimator for speech and music”, Journal of the Acoustical Society of America 111(4), 2002. doi:10.1121/1.1458024 — the pitch detector this tool uses.
By Soluto, independent developer, not a clinician · Published · How this tool works and where the numbers come from