Using Your Phone as a Decibel Meter

Can a phone measure decibels?

Yes, for everyday questions. Open this site in a current mobile browser, press Start and grant microphone permission — the page becomes a phone decibel meter with live estimated dBA, dBC or dBZ plus Leq, minimum and maximum. It answers practical questions well: how loud is this room, is the street noisy tonight, how strong is the speaker at one metre, how quiet is the bedroom. These are relative and reference checks, exactly what a phone microphone suits. Learn the three-step flow in how to use the meter.

What accuracy to expect

Treat phone readings as rounded estimates, not exact figures. Two different models on the same sound commonly disagree by 5 to 10 dB because capsules, preamps, cases and positions differ. Below about 30 dBA most phone microphones hit their own noise floor and read too high; above roughly 95 to 100 dBA they clip and read too low; deep bass is under-reported because phone capsules favour the speech range. Ignore decimal places and compare averages over 30 to 60 seconds instead of instants. The full picture is in accuracy and limitations.

iPhone and Android differences

Both platforms work, but their audio processing differs. Safari on iPhone tends to keep automatic gain and noise processing active even when a site asks to disable it, so a given iPhone reads consistently yet offset against other devices. Chrome on Android usually honours the request to switch echo cancellation, noise suppression and automatic gain off — this meter requests exactly that and shows which processing stayed enabled in its device panel. Whichever platform you use, one calibration against a known reference removes most of the device-specific offset.

Browser or app?

A browser meter wins for quick checks: no install, no account, no store permissions beyond the microphone, always the current version, and audio that never leaves the device — see privacy. A native app suits offline-first or background-logging workflows. For a one-off measurement on any phone, including somebody elses, the browser page is the faster honest option.

Best practices on a phone

• Find the microphone hole, usually on the bottom edge, and keep it uncovered; check that the case does not cover it.
• Measure at ear height where people actually sit or stand, with the microphone pointed at the source.
• Hold the phone still, stay silent while measuring, and note the distance with every reading.
• Run at least 30 to 60 seconds and use the Leq or average, following how measurement works.
• Calibrate once per phone and keep the same position and settings for comparisons.

Chrome, Edge, Firefox and Safari in practice

All four current browsers run this meter over HTTPS with permission remembered once per origin, capturing mono audio that never reaches your speakers. This page requests echo cancellation, noise suppression and automatic gain off, then reports what stayed active in its device panel. Chrome on Android generally honours the off request, so the trace follows the room; Safari on iPhone often keeps system processing engaged, so the trace looks smoother yet sits offset. Edge on Android behaves close to Chrome, Firefox between the two by version. Expect several decibels of spread before correction, converging once each device carries its own calibration from a known reference at the same spot.

Uncalibrated, every browser shows the same labelled estimate — weighted dBFS plus a nominal +100 dB offset — with raw dBFS visible in Digital Diagnostics. An idle page shows "--", never a simulated room tone. There are no accounts and no downloads; audio stays on the handset.

A five-minute side-by-side you can trust

Set both phones on the same table, bottom-edge microphone holes uncovered and pointed at the same speaker 1 metre from TV dialogue. Choose A-weighting with Slow response on each, press Start together, stay silent for 60 seconds, and compare Leq rather than dancing instants. One handset may read 68 dBA Leq while the other reads 73; that gap is capsule, case and processing, not a broken meter. Save each offset separately and never move one phone's correction onto the other. For export, CSV and JSON carry dB values only when calibrated, and the meter keeps 20 local sessions per browser — see speaker test and dBFS vs dB SPL.

When a phone is not enough

A phone is never evidence. Noise complaints, workplace assessments, legal disputes and regulatory checks need a calibrated Class 1 or Class 2 instrument operated by a qualified person — see the validation protocol and measurement disclaimer. For exposure context, continue with noise exposure and the decibel comparison chart.

Open the online decibel meter → · Microphone troubleshooting →