The instrument in one paragraph — and why repeatability matters
A decibel meter — also called a sound level meter, noise meter, dB meter or sound meter — measures how strong sound is at a point in space and reports it in decibels. A microphone converts pressure into voltage, electronics average that signal over a standardized time and frequency filter, and the display shows the result. Professionals carry palm-sized Class 1 or Class 2 meters meeting IEC 61672, with a lab-measured capsule and a 94 dB pistonphone calibrator; everyone else can now open the same measurement principle in a browser tab on the homepage meter.
The key insight most beginners miss: the meter does not hear like you do. It follows strict filters (A/C/Z weighting) and timings (Fast/Slow/Impulse) so two meters, operated correctly, agree with each other. That repeatability is what separates a meter from a volume knob — and it is why noting weighting, duration and distance matters more than the number alone. “62 dBA Leq, 60 s, 2 m” is data; “loud” is not.
Inside the box: microphone, weighting, time averaging, display
Every sound level meter, hardware or browser, has the same four stages. 1. Microphone: a capsule with known sensitivity converts pascals to millivolts. Lab capsules are flat across frequencies; phone MEMS capsules are optimised for voice and roll off bass. 2. Frequency weighting: A-weighting damps lows and extreme highs to mimic hearing, C-weighting stays flatter for bass-heavy sound, Z-weighting shows raw flat response. 3. Time weighting and integration: Fast (125 ms) follows fluctuations, Slow (1 s) smooths them, Impulse catches brief transients; Leq integrates energy over the whole session for exposure questions. 4. Display and statistics: current level plus minimum, average, maximum, peak and Leq, often with a time-history graph and spectrum.
Our online implementation does all four with the Web Audio API entirely on your device: incoming audio is sliced into short windows, energy per window is computed, weighting is applied per frequency band, the selected time constant smooths the stream, and the result is converted to decibels against a stored calibration offset. Nothing is uploaded or recorded — see methodology for the signal chain and privacy policy for data handling. Because phone capsules differ in sensitivity, uncalibrated absolute readings are estimates; relative comparisons on the same device over 30–60 second windows are genuinely useful.
Reading the screen: every value explained with examples
Current is the live level right now — flickers with every syllable or passing car. Minimum marks the quietest moment — use it for background noise when no events occur (a bedroom minimum of 31 dB with peaks of 58 dB tells you traffic intrudes on an otherwise quiet room). Maximum marks the loudest event — the door slam the average hides. Average is the typical level; Leq is the energy average professionals quote — one loud hour dominates a quiet hour, so quote Leq for any exposure claim and feed it to the exposure calculator. Peak catches the sharpest instantaneous pressure (hammer blow, balloon pop); RMS tracks perceived loudness better than raw peaks and underlies the current reading.
Letters and speeds complete the picture: dBA for hearing and daily life, dBC for bass and engines, dBZ for raw data (see dBA, dBC and Leq tools and dB vs dBA); Fast for most checks, Slow for steady hum, Impulse for claps and impacts. The frequency analyzer answers which pitches dominate — boom near 150–300 Hz, harshness near 2–4 kHz, hiss above 6 kHz — which a single dB number can never tell you.
Measuring right in 60 seconds: the routine professionals recognise
Hold the device at ear height where people actually are — desk, bedside, seat — with the microphone pinhole uncovered and facing the source. Silence yourself (your voice adds 10–20 dB), wait until the reading settles, and record average or Leq over at least 30 seconds rather than a single instant. Write place, distance, weighting and duration every time. Repeat from a second spot and compare averages: agreement within a few dB means you can trust the picture; wild swings mean something nearby pulses and the maximum tells that story.
Calibrate once and your comparisons get far more honest: play a known 94 dB reference tone or compare against a trusted second meter at the same position for 30–60 seconds, enter reference minus reading as the stored offset in the Calibrate dialog, and retest. Full method in the calibration guide and device tricks in how to use an online meter on phone or PC. Log repeats in the Noise Journal with CSV/JSON exports so outliers stay visible.
What an online meter cannot do — honest limits
Consumer microphones compress bass, apply unseen gain control and suppression, and vary several dB between phone models — two phones on the same table commonly disagree by 5–10 dB. Laptop fans add 35–45 dB of background; phone cases covering the pinhole cost up to 10 dB; wind outdoors adds artefacts no filter fully removes. An online meter is therefore a reference and education tool: fine for checking a nursery (aim under 50 dB at the crib for white-noise machines), a classroom (averages above 55–60 dBA hurt concentration), or rehearsal volume — but never evidence for courts, regulators, or workplace compliance. Treat oddities as clues: a library reading 70 dB means a blocked mic or suppression fighting you, not a loud library.
See what affects accuracy for quantified factors and how we validate for test method. The math is defensible; the capsule is the limit — and we say so plainly because AdSense-grade trust requires it.
When you need certified hardware — and how to brief the professional
Reach for an IEC 61672 meter with laboratory calibration whenever money, law, or compliance rides on the number: workplace exposure assessments under OSHA or the EU directive, planning and nuisance disputes, product specifications, classroom or concert-hall commissioning. Expect a competent surveyor to calibrate before and after, measure at defined positions and durations, log weather and operating conditions, and report Leq plus statistical levels (L90, L50, L10) with uncertainty — not a single phone screenshot.
Bring your online readings anyway. A dated log — “living room 63 dBA Leq 15 min, peaks 78 dBA, dBC 81, window closed, 22:10–22:25” — tells the surveyor exactly where and when to deploy expensive kit, cutting survey time and cost. For everything else — curiosity, comfort, sleep, study, rehearsal, travel — the free meter on the homepage is the fastest answer to “how loud is this, really?” Open it, measure a minute, decide with data, and read when loud becomes harmful next.
Reference tool only — not for legal, medical, occupational, or regulatory decisions.