To install: tap Share, then Add to Home Screen.
Free · runs entirely in your browser

Do you sound monotone — or engaging?

Record a short speech sample for a real acoustic analysis of your prosody: pitch range, expressiveness, pauses, and pacing. Useful for public speaking, podcasting, interview prep, and language practice.

⚠ This computes genuine acoustic measurements, but it's an indicator, not a lab-grade voice assessment — background noise and recording quality affect the numbers.
SCAN · local analysis0 data sent
Tap to record 20–40 seconds of speech
or
—
Overall score /100
—
Pitch range (st)
—
Pitch variation (st)
—
Articulation (syl/s)
—
Time in pauses
—
Loudness var. (dB)
Pitch in semitones around your median (blue), loudness (grey), pauses (red bands), detected syllables (ticks)
Multi-language translator device and noise-reduction voice recorder
Voice tools · Loch Ness selection

Let your voice speak every language

A pocket translator for conversations across languages, and a noise-reduction recorder for clear voice notes.

TranslatorVoice recorder
Discover voice tools

What is prosody, and why it matters

Prosody is the melody of speech — the rise and fall of pitch, the rhythm, the pauses, the pace. It's separate from word choice: the same sentence can sound confident, bored, or anxious depending purely on how it's delivered. Listeners pick up on prosody almost instantly, often before consciously processing the words themselves, which is why it has an outsized effect on how engaging, credible, or persuasive a speaker sounds.

What this tool measures

Pitch range

The spread between your typical low and high pitch, measured in semitones — a wider range generally reads as more expressive.

Expressiveness

How much your pitch varies moment to moment relative to your average — low variation reads as monotone.

Pauses

Silent gaps of at least 300ms, detected from the energy envelope — their number and placement shape pacing and emphasis.

Speaking pace

Syllables per second, counted from energy peaks in voiced speech — with and without pauses.

Flat stretches

The share of time spent in 2-second windows where your pitch barely moves — where listeners tend to drift.

Phrase endings

Whether each phrase ends rising, level or falling — frequent rises can make statements sound like questions.

Loudness & emphasis

How much your volume varies, and whether stressed syllables get both higher and louder.

Rhythm & voice clarity

Contrast between long and short syllables (nPVI) and an approximate harmonics-to-noise ratio.

How the pitch contour is built

The recording is resampled to 16 kHz and cut into 25 ms frames every 10 ms. Speech and silence are separated with a threshold that adapts to your microphone's noise floor, so a louder or quieter mic doesn't change the pause count. Pitch is tracked with the YIN algorithm, octave errors are corrected against a local median on both sides of each frame, and all pitch measures are expressed in semitones so low and high voices are compared on the same scale.

Why two takes never give exactly the same numbers

Nobody says the same thing twice with identical melody, speed and pauses — natural variation between takes is real. On top of that, a short recording is a small sample. The tool shows a ± margin for pitch measures, a confidence level for the recording, and keeps your recent takes in this browser so you can compare averages instead of single runs. For the most stable comparison, read the standard passage each time.

Who this is useful for

Limitations

Frequently asked questions

What is prosody in speech?

The rhythm, stress, and intonation patterns of speech — how pitch rises and falls, how fast you speak, and where you pause. It's how you say something, not what you say.

How is a monotone voice measured?

Via the standard deviation of your pitch in semitones, the semitone range between the 10th and 90th percentile of detected pitch values, and the share of time spent in flat 2-second stretches.

Why do two recordings of the same voice give different results?

Your delivery naturally changes from one take to the next, and short or noisy recordings add measurement noise. Record at least 20 seconds, read the same passage, keep the same mic distance, and compare the average of several takes.

Is this analysis scientifically precise?

It uses real acoustic measurements, but background noise, mic quality, and a short recording introduce variability — treat it as a useful indicator, not a lab-grade assessment.

Is my recording uploaded anywhere?

No. Everything runs locally in your browser via the Web Audio API. Nothing is transmitted, stored, or shared.

How can I use this to improve public speaking?

Record a practice segment, check your range and expressiveness score, then consciously vary your pitch and add deliberate pauses before key points. Re-record and compare.