Skip to content
Tinnitus Clarified

Free tool

Digits-in-Noise Test

Runs entirely in your browser using your device's built-in voice. No audio is recorded and nothing is uploaded; your last result is kept on this device so you can compare later. Not a diagnosis; see sudden sensorineural hearing loss if your hearing changed suddenly, is one-sided, or came with pain, pressure, or dizziness.

You'll hear three spoken digits over background noise, repeated 18 times. Type back what you heard each time. The noise gets a little louder every time you get it right, and a little quieter every time you don't — so it naturally zeroes in on the noise level where understanding speech gets genuinely hard for you. Takes about 3 minutes. Headphones or quiet room, browser sound must be unmuted.

What is a digits-in-noise test?

A published screening method that presents triplets of spoken digits (0–9) over background noise and measures the point where you can just still follow them — your speech reception threshold (SRT). It was originally developed in Dutch as a telephone screening test, and the method has since been validated methodologically across 17 different languages.

How this works, and how accurate is it really?

The digit-triplet test — three spoken digits presented over steady background noise, with the noise level adapting to each answer — is a real, published screening method. Because it measures a ratio between speech and noise rather than an absolute loudness level, it's less sensitive to playback volume and device differences than a pure-tone test — that's the whole reason it works reasonably well as a self-administered check, and why it holds up without a soundproof booth, unlike classic tonal audiometry.

The original telephone version of the test was validated with a measurement error under 1 dB — genuinely high precision for a test administered over an ordinary phone line. Modern browser- and app-based variants typically take under 5 minutes, and some newer adaptive procedures run in under 70 seconds.

What's different here: published versions use pre-recorded, calibrated speech. This version uses your browser's built-in text-to-speech voice, which varies in clarity across browsers and operating systems and hasn't itself been validated against clinical outcomes. Treat the score as a rough, self-generated signal — useful if it's clearly outside the typical range or if it prompts you to get a real test, not as a clinical result on its own.

What do the results mean?

Lower (more negative) dB SNR is better — it means you could still follow the digits even when the noise was louder than they were. A score clearly outside the range typically reported for normal hearing in the published literature is worth a real evaluation, not self-diagnosis.

Digits-in-noise vs. a pure-tone test

A pure-tone test checks whether specific pitches are audible in quiet — closer to what a clinical audiogram measures. A digits-in-noise test checks something people notice more in daily life: following speech when there's background noise, which pure-tone results alone don't always predict well. The two catch different things, which is why this site offers both.

Connection to tinnitus

Difficulty following speech in noise is a common early sign of the same hearing changes that often accompany tinnitus — see tinnitus and hearing loss: the connection explained for the mechanism, and noise-induced tinnitus if loud environments are part of your history.

When should you see a professional?

A clearly poor score, a sudden change, or difficulty in noise that's affecting work or conversation day to day — any of these are reasons to get a real evaluation, not to keep retesting yourself here.

FAQ

How accurate is a digits-in-noise test?

The original telephone version of this test was validated with a measurement error under 1 dB — genuinely high precision for a self-administered check. It works well even without a sound-treated booth, which is exactly why it's used as a large-scale screening tool, including entirely over the phone.

Why digits instead of a pure tone?

It measures a ratio between speech and noise rather than an absolute loudness level, so it's less sensitive to your specific headphones and playback volume than a pure-tone test — and it reflects something closer to real-world listening: following speech in a noisy room.

Is this the actual validated clinical test?

The method is real and published, validated uniformly across 17 languages. This implementation uses your browser's built-in voice for the spoken digits rather than the pre-recorded, calibrated speech published versions use, so treat your score as a rough signal, not a clinical result.

Prefer a different kind of check?

This tool checks speech understanding in noise. A tone-based hearing test checks whether specific pitches are audible in each ear separately, which is closer to what a standard audiogram measures.