Online Sound Level Meter with dosimeter – measure dB(A)

This free online sound level meter turns your phone or laptop into a decibel meter: it reads the sound pressure level in dB(A), averages the whole measurement into an Leq, and tells you how long you can stay at that level safely using the NIOSH criterion. It runs in the browser — nothing to install, no sign-up — and the audio is analyzed and discarded on the fly: nothing is recorded or uploaded.

The reading is indicative (±10 dB), not lab-grade, because a phone microphone is not calibrated. It still answers the two questions people actually bring: “is this around 60 dB or 90 dB?” and, more importantly, “how long can I stay here without hurting my hearing?”.

How to use the sound level meter, step by step

1. Tap “Enable microphone” and accept the browser prompt. The page must be served over HTTPS (this one is) or the browser will not grant access.
2. Keep the mic clear, pointed at the noise source and roughly at ear height. Do not cover it with a finger or a thick case — that alone costs you several decibels.
3. Let it run for at least 10–15 seconds. The instantaneous number jumps around; what describes a noise is the average.
4. Read the Leq, not the peak. “Now” is for catching individual events; the Leq is what you compare against limits and against the safe-time table.
5. Switch weighting if you need to (A, C or Z). Changing it resets the average and the maximum on purpose: mixing two scales would corrupt both.
6. Tap “Stop and see result”. You get the Leq, the peak, the safe time at that level, the second-by-second timeline, a shareable card and a CSV with the full series (up to 15 minutes).

One practical rule: measure where you actually are — the couch, the bed, your workstation — not next to the source. In a free field the level drops about 6 dB every time you double the distance, so one meter from a speaker and four meters from it are two different stories.

What each number means

A, C or Z weighting: which one to use

The A-curve here is a biquad approximation built on the standardized pole frequencies (20.6 · 107.7 · 737.9 · 12194 Hz) and normalized to 0 dB at 1 kHz — the same shape as IEC 61672 A-weighting, within the accuracy a phone mic can justify.

Calibrating it — the step that makes the number useful

Put a meter you trust next to your device with a steady sound playing, and move the calibration offset (70–130 dB) until both read the same. The offset is stored in your browser, so you only do it once per device. A calibrated reference gets you to roughly ±2 dB in the middle of the range; without one, treat every reading as ±10 dB and use it for comparisons — this room versus that room, window open versus closed — rather than as an absolute number.

How many decibels are dangerous, and for how long

85 dB(A) for 8 hours is the NIOSH recommended exposure limit, and every extra 3 dB halves the safe time. That exchange rate is why loud is worse than it sounds: a bar at 100 dB(A) spends your whole day’s allowance in 15 minutes.

The tool adds this up for you as today’s dose: 100 % means you have used the full daily allowance, whether that took eight hours at 85 dB(A) or a quarter of an hour at 100 dB(A). If you want the reasoning behind the numbers — and why NIOSH and OSHA publish different limits for the same noise — that is a separate read.

What people actually use a phone sound meter for

Neighbor noise, building work and venues

This is the most common reason people open a noise meter, and the useful part is knowing what “too loud” means indoors. Spanish regulation (RD 1367/2007), for instance, sets interior acoustic quality objectives of 40/40/30 dB(A) in bedrooms and 45/45/35 dB(A) in living rooms for the day (7:00–19:00), evening (19:00–23:00) and night (23:00–7:00) periods. Those are low numbers: a fridge two meters away already reads around 35 dB(A). Measure with the windows shut, in the affected room, and log several sessions with timestamps — a pattern is far more persuasive than one dramatic peak.

Concerts, bars, gyms and rehearsal rooms

Here the number that matters is not the level, it is the dose. Start the meter when you walk in and leave it running: three hours in a room at 97 dB(A) is roughly six times your daily allowance. It is also the fastest way to settle an argument about whether the gym playlist or the rehearsal room is “a bit loud” or genuinely harmful.

Classrooms, an unusually demanding case

WHO community noise guidance puts the recommended background level in a classroom at 35 dB(A) during teaching (the US standard ANSI S12.60 lands on the same figure), because a teacher’s voice reaches roughly 55–60 dB(A) at one meter and speech needs about 15 dB above the background to be understood without effort. Put on a screen as a live “noise traffic light”, the meter turns that into something a group can self-regulate against.

Your hi-fi, home cinema or studio

Use it to match speaker levels and to know where you really listen. Play a steady signal — pink noise from the tone generator is the standard choice — set each speaker to the same reading at the listening position, and you have channel balance without a calibration disc. Switch to C weighting when a subwoofer is involved, or the A-curve will hide most of what it is doing.

Headphones: why you cannot measure them this way

You cannot measure headphone level with a phone microphone, and no browser tool can. An honest figure needs an ear simulator or a coupler that reproduces the acoustic load of an ear canal; a mic held against the earcup leaks and reads low. Use the 60/60 rule instead — 60 % of maximum volume, no more than 60 minutes at a stretch — and check how your hearing is holding up with the online hearing test.

Work, workshops and machinery

As a triage tool it is excellent: it tells you which machine, which corner and which shift deserve a proper measurement, and whether hearing protection is optional or obvious. It is not, and cannot be, a compliance measurement — that takes a calibrated meter and a technician.

Is a phone sound level meter accurate?

Expect roughly ±10 dB versus a real meter, with the largest errors at the quiet and loud extremes. That is not a flaw in the software; it is the hardware.

What happens inside a phone microphone

Phone mics are tuned for speech, not for measurement: the response is far from flat below ~100 Hz and above ~10 kHz, and the capsule starts to compress and then clip somewhere around 100–110 dB, so very loud events read lower than they are. Phones also apply automatic gain control and noise suppression by default — this tool asks the browser to switch them off, and it usually obeys, but a browser or an OS can override that. Two phones side by side reading 4–5 dB apart is normal.

What NIOSH measured

The reference here is NIOSH’s own iOS sound level meter app, which they document as accurate within ±2 dBA and as meeting Type 2 (IEC 61672) requirements only with a calibrated external microphone. They also state plainly that verifying an Android app’s accuracy is not currently possible, because the hardware varies too much between manufacturers. If the agency that wrote the exposure limits needs an external calibrated mic to claim Type 2, no web page running on an unknown phone should promise more.

How this tool computes the numbers

Level is the RMS of each analysis window (about 45 ms, a little faster than the 125 ms “Fast” response of a standard meter, which is why the reading is livelier), converted to dBFS and shifted by your calibration offset. The Leq is a true energy average, the dose follows NIOSH (85 dB(A) / 8 h with a 3 dB exchange rate), and the timeline and the CSV store one sample per second. Our measurement method is published in full, including what we do not claim.

Can I use it to report a noisy neighbor?

No — no phone app produces a legally valid noise measurement, this one included. Regulations require instruments under legal metrological control (Class 1 for anything that can end in a sanction or a court case), a calibrator checked before and after the measurement, and a qualified technician following a defined procedure. An uncalibrated phone fails all three.

What it is good for is building the case that gets a real measurement done: export the CSV, keep dated sessions showing when the noise happens and how long it lasts, and take that to your local authority or building manager. “It’s loud at night” is an opinion; “42 dB(A) average between 1:00 and 2:30, six nights out of seven” is a reason for someone to come and measure properly.

A meter tells you how much noise there is; the next question is what it is made of. Run the real-time spectrum analyzer on the same sound and you will see whether the problem is a low-frequency hum, a whine or broadband hiss — which is what decides how you fix it.

Frequently asked questions

Indicative measurement on an uncalibrated microphone (±10 dB). Not valid for legal or workplace noise compliance.