How to use it
- Press Start microphone and allow access when your browser asks.
- Keep Frequency weighting on for an A-weighted reading, which is closer to how the ear hears.
- To estimate sound pressure, play a sound of a known level, enter its level under Level of a known sound and press Calibrate to this level.
- Read Now, Lowest, Highest and Average, or watch the chart. Press Stop meter when you are done.
dBFS, dB SPL and calibration
dBFS measures digital sound against the top of the scale: 0 dBFS is the largest value a digital signal can hold, and every 6 dB below it is about half the amplitude. A sound pressure level, dB SPL, is a level in air, and the page cannot know it from a digital reading alone, because a phone or laptop does not report the gain of its microphone. Calibration fixes that. The difference between a known level and the dBFS reading you get for it is an offset, and the offset is added to every later reading to give an estimate.
estimate (dB SPL) = dBFS reading + offset, where offset = known level − dBFS reading during calibrationA-weighting
The ear is less sensitive to very low and very high frequencies. A-weighting, defined in the IEC 61672 standard, reduces those parts of a sound in the same way, so a reading in dBA follows the loudness people hear. The page uses a digital filter that follows the A-weighting curve within half a decibel from 100 Hz to 4 kHz.
Typical sound levels
The table gives typical average levels in dBA from the NIDCD page on noise-induced hearing loss. They are typical values, not measurements of your room. The NIDCD notes that sounds at or above 85 dBA can cause hearing loss with long exposure, so a reading near that level for a long time deserves care.
Limits
The page turns off the browser's automatic gain control, noise suppression and echo cancellation, because they change the level. The microphone's own sensitivity is still unknown, so compare readings only with the same device, in the same place and at the same distance.
Example: calibration
A reference meter reads 70 dB while the page reads -30 dBFS. The offset is 70 − (−30) = 100 dB. A later reading of -20 dBFS is then an estimate of 100 + (−20) = 80 dB SPL.
Frequently asked questions
Is a phone a good sound level meter?
Not for a binding figure. A phone microphone is not calibrated, and the page turns off its automatic processing, but the result is still an estimate. Use a calibrated meter for work that needs an exact number, such as a workplace noise check.
What does dBA mean?
dBA is a sound level with A-weighting applied, which reduces the very low and very high frequencies to match the ear. Most noise guidance is given in dBA, including the NIDCD figures in the table on this page.
Why does the page show a dash?
A dash means there is no sound yet. Silence has no level in decibels, so the readings stay blank until the microphone picks up something.
What is the average on this page?
The average is an energy average of the readings since the last reset or calibration. One loud moment raises it more than several quiet ones, which is how the Leq measure used for noise exposure works.
Can I calibrate without a reference meter?
Not accurately. Calibration needs a sound of a known level, such as one measured by a calibrated reference meter. Without one, use the dBFS reading to compare sounds with each other rather than reading absolute levels.
Does the page store or send my sound?
No. The sound is analysed on your device, nothing is recorded or stored, and closing the page stops the microphone.
Sources
- Noise-induced hearing loss, NIDCD (National Institutes of Health)
- A-weighting, Wikipedia (IEC 61672 formula)
- Web Audio API, W3C specification (AnalyserNode)
- Media Capture and Streams, W3C specification (getUserMedia)
Figures and text last checked on Tuesday, November 10, 2026.