Voice Rater — How Does Your Voice Sound?

Record five seconds and get a 1-10 score, your estimated voice age, and a gender read-out. Everything runs inside your browser.

Press record and speak normally for five seconds — read a sentence out loud, or just say what you had for breakfast.

Nothing is uploaded. Your recording never leaves this browser tab.

How the Voice Rater Works

Most "rate my voice" sites either upload your recording to a server or quietly hand you a random number. This one measures your voice on your own device, using the same acoustic parameters that speech clinicians and voice coaches actually work with. Here is exactly what happens in the two seconds after you stop recording.

Your microphone signal is split into 40-millisecond frames. Silent frames are dropped. For every remaining frame the tool runs the McLeod pitch method — a normalised autocorrelation — to find your fundamental frequency, the physical rate at which your vocal folds are opening and closing. That single number, usually written F0, is the backbone of everything else on the results panel.

From there four more measurements are taken:

JitterHow much each vibration cycle differs in length from the one before it. Under 1% is typical of a healthy voice; above 2% is what makes a voice sound rough.
ShimmerThe same idea applied to loudness instead of pitch. High shimmer is heard as a wobbly, unsupported tone.
Harmonic-to-noise ratioHow much of your sound is clean vocal tone versus breath and turbulence. Healthy voices sit above 15 dB.
Formants F1 and F2The two loudest resonances of your vocal tract, found with LPC spectral envelope analysis. These describe your timbre and vocal tract size.

Those readings are combined into the 1-10 score using fixed, published weights: clarity 24%, pitch stability 20%, volume control 14%, expressiveness 14%, tone brightness 14%, delivery flow 8%, and range 6%. Nothing is randomised, and recording the same passage twice in the same room will give you the same score to within a few tenths. You can open the Raw acoustic measurements panel under any result to see every underlying number yourself.

Rate Your Voice in 3 Steps

The whole thing takes under a minute, and there is nothing to install.

1 · Allow the micPress the record button and accept your browser's microphone prompt. The permission is per-site and you can revoke it any time.
2 · Speak for five secondsTalk normally at a natural volume. A full sentence works far better than a single "aaah" because the tool needs to see your pitch move.
3 · Read your resultsYou get a score out of 10, an estimated voice age, a masculine-to-feminine read-out, your average pitch in hertz, and a breakdown of all seven sub-scores.

Want a longer take for singing instead of speech? Use the singing voice rater, which records for eight seconds and weights pitch range far more heavily.

What Your Voice Score Means

The score is deliberately calibrated so that an ordinary healthy voice on ordinary consumer hardware lands in the sevens. If everyone scored a nine the number would be meaningless, and if everyone scored a four the tool would just be insulting people. Here is the actual distribution the scoring curve was tuned against, using published clinical reference values for each voice type.

9.0 - 10 · OutstandingTrained speaking voice, very low jitter, recorded in a quiet room with a decent microphone. Rare on a laptop mic.
7.5 - 9.0 · StrongClear, steady, well-projected. This is where most people who record carefully in a quiet room end up.
6.5 - 7.5 · SolidA completely normal speaking voice recorded in normal conditions. Nothing wrong here.
Below 5.5 · Check conditions firstAlmost always background noise, a distant microphone, or a cold — not a bad voice. Re-record before drawing conclusions.

Because the biggest single weight sits on harmonic-to-noise ratio, your recording environment influences the score more than most people expect. A fan, an open window, or sitting a metre from your laptop can each cost you a full point. If you want to compare two takes fairly — before and after a warm-up, for instance — record both in the same spot, at the same distance, within a few minutes of each other.

Voice Age & Gender Analysis

Alongside the score you get two estimates that people usually come here for. Both are honest acoustic readings, and both have real limits worth understanding before you take them seriously.

Voice age

Voices genuinely do change measurably with age. Cycle-to-cycle irregularity — jitter and shimmer — tends to rise across the lifespan, while harmonic-to-noise ratio falls. Pitch follows different paths for different voices: feminine voices drift gradually downward after early adulthood, while masculine voices dip through middle age and then rise again later. The estimate combines a roughness score built from jitter, shimmer and HNR with a pitch-based term, then clamps the result to a believable 14-70 range rather than extrapolating to absurd numbers.

The honest caveat: a head cold, a late night, a cheap headset or a noisy kitchen all push the same measurements in the same direction as ageing does. A single take can easily read a decade off. It is a fun, physically-grounded guess — not a measurement of you.

Masculine and feminine resonance

The gender read-out places your voice on an acoustic scale rather than putting it in a box. Average pitch carries most of the weight, with your first two formants contributing the rest — formants matter because vocal tract length shifts resonance independently of pitch, which is why pitch alone misreads a masculine falsetto or a low feminine voice. Voices between roughly 155 and 175 Hz sit genuinely in the middle and are reported as androgynous, with a confidence figure attached so you can see how close to the boundary you are.

If that is the specific number you came for, the dedicated voice gender detector shows the pitch and formant evidence side by side.

Getting an Accurate Reading

Five practical things make a bigger difference to your result than anything about your actual voice:

Close the windowsTraffic and fan noise directly reduce harmonic-to-noise ratio, the heaviest-weighted component of the score.
Sit 15-25 cm awayToo far and the room dominates; too close and plosives clip the signal.
Speak a full sentenceSustained vowels give the tool nothing to measure for expressiveness or range, which costs you two sub-scores.
Skip the headset noise cancellingAggressive noise suppression and auto-gain flatten exactly the micro-variations being measured.

The tool deliberately requests your microphone with echo cancellation, noise suppression and auto-gain switched off, precisely so those processing steps do not erase the signal it needs. Some system-level or headset-level processing sits outside the browser's control, though, so a wired microphone or your laptop's built-in mic will usually give a more faithful reading than a Bluetooth headset.

Your Recording Never Leaves Your Device

This matters enough to state plainly. Voice Rater performs every calculation locally using the browser's Web Audio API. There is no upload, no server-side processing, no account, and no storage — your audio exists only in your tab's memory and is discarded the moment you close it or record again. That is a deliberate design decision rather than a policy promise: several popular voice rating sites transmit your recording to a backend for analysis, and once audio leaves your device you have no way to verify what happens to it.

The trade-off is that everything has to be computed on your own hardware, which is why the analysis is tuned to run in well under a second on a phone.

Frequently Asked Questions

Is the voice rater free?

Yes. Voice Rater is completely free with no signup, no account, and no usage limit. You can record as many times as you like.

Is my voice recording uploaded anywhere?

No. All analysis runs locally in your browser using the Web Audio API. Your microphone audio is never sent to a server, never stored, and disappears the moment you close the tab.

How accurate is the voice age estimate?

Treat it as a rough guess, not a measurement. Voice does change with age in ways that are measurable: cycle-to-cycle irregularity rises and harmonic-to-noise ratio falls over time. But a cold, tiredness, a cheap microphone or a noisy room shift the same numbers, so a single take can easily be off by a decade.

Can the voice rater tell if I sound male or female?

It reports where your voice sits on an acoustic masculine-to-feminine scale, based on your average pitch and your first two formants. That is a description of sound, not of identity. Voices in the 155-175 Hz band genuinely sit in the middle and will read as androgynous.

What is a good voice score?

Most healthy adult voices recorded on a normal laptop or phone land between 7 and 8.5. Above 9 usually means a trained or naturally very steady voice recorded in a quiet room. Below 5 usually points at a cold, a very noisy room, or a microphone placed too far away.

Why did I get a low score?

The three most common causes are background noise, sitting too far from the microphone, and speaking too quietly. All three lower the harmonic-to-noise ratio, which carries the most weight in the score. Open the raw measurements panel: if the harmonic-to-noise ratio is under 10 dB, it is your recording conditions, not your voice.

Do I need to install anything?

No. Voice Rater runs in any modern browser on desktop, Android and iOS. There is no app, no plugin and no extension. You only need to allow microphone access when your browser asks.