Online Decibel Meter — Free Tool

This online decibel meter shows live microphone levels as a digital reading in dBFS plus a calibrated estimate in dB SPL, with Fast and Slow ballistics, a peak hold and a segment meter. It runs entirely in your browser through the Web Audio API, and no audio is uploaded or recorded.

How to use it

1. Enable the microphone. Press Enable Microphone and allow access. The page requests a raw mono stream with echo cancellation, noise suppression and automatic gain control all switched off, so the reading follows the real sound instead of the browser’s speech processing.
2. Pick the ballistics. Fast (125ms) tracks speech syllables and impacts, reaching 63% of a step change in 125 ms; Slow (1000ms) settles on steady noise such as traffic or an air handler and needs about 3 seconds for the same change. Both are exponential time constants, not moving averages.
3. Set the calibration offset. The slider covers 70 to 120 dB and starts at 94 dB, the sound pressure level of 1 pascal. Move it until the reading agrees with a known reference.
4. Watch the three readouts. Digital Level is the smoothed RMS in dBFS, Peak Hold keeps the highest recent value and falls at 8 dB per second, and Peak SPL applies the current offset to that peak.
5. Use the benchmark table for orientation. It lists 30, 50, 60, 75, 85 and 100+ dB steps with everyday examples and marks the 85 dB row as the workplace exposure area.
6. Press Stop Meter when you finish; the microphone stream closes immediately. Simulate Audio runs a scripted demo level if you want to see the interface without granting permission.

What the numbers mean: dBFS versus dB SPL

A browser microphone never reports pressure directly. It hands the page a stream of samples whose full-scale range depends on the interface, the microphone sensitivity and the preamp gain, so every value here starts as dBFS and becomes an estimated dB SPL only after the offset is added.

• dBFS is relative to digital full scale. 0 dBFS is the largest sample a converter can carry, working signals sit below it, and a drop of 6.02 dB halves the amplitude: 20 × log10(0.5) = −6.02.
• dB SPL is relative to 20 micropascals. One pascal works out to 20 × log10(1 / 0.00002) = 93.98 dB SPL, which is where the 94 dB default offset comes from.
• The offset is linear: estimated SPL = dBFS + offset. Swap the microphone, change the input gain in the operating system, and the same acoustic sound produces a different dBFS number.
• No frequency weighting is applied. The meter reports broadband level, not an A-weighted or C-weighted level, so low-frequency rumble counts at full strength even though hearing discounts it.

Calibrating against a reference

Real calibration needs a known acoustic source. Field calibrators meeting IEC 60942 clip onto the microphone and emit a fixed level, commonly 94.0 dB at 1 kHz on class 1 and class 2 units; adjust the slider until the meter reads that value. Without one, matching the offset against a trusted hardware meter at a steady 70 to 80 dB source gets you within a few dB. A phone microphone’s response is not flat, so any single-number calibration holds only near the level and spectrum you calibrated at.

Distance, reflections and what the math predicts

A point source radiates in all directions, so pressure falls with distance and level falls 6.02 dB per doubling of distance in a free field. The table is the arithmetic behind that rule: 20 × log10(1/r).

Distance from a point source	Level change versus 1 m
1 m	0 dB (reference)
2 m	−6.0 dB
4 m	−12.0 dB
8 m	−18.1 dB
16 m	−24.1 dB

Indoors the free-field rule holds only close to the source. Reflections from walls, floors and desks add energy back, and past the room’s critical distance the reverberant field takes over, so level stops falling with distance. Moving a phone a meter across a live room often changes the reading far less than the table predicts.

Reference levels and exposure

Published SPL figures give a reading context; these values assume a calibrated, A-weighted, time-averaged measurement, which this page does not provide.

Typical level	Example at a normal listening distance	Reading notes
20–30 dB SPL	Quiet bedroom at night, treated studio floor	Near the self-noise of many phone microphones
40–50 dB SPL	Quiet office, gentle rain	Comfortable background range
60–65 dB SPL	Conversation at 1 m	Common quick check for the calibration offset
75–80 dB SPL	Vacuum cleaner, busy restaurant	Speech needs extra effort
85 dB SPL	Lawn mower, printing press, food processor	NIOSH recommended exposure limit: 85 dBA averaged over an 8-hour workday
95–100 dB SPL	Chain saw, nightclub, large sports event	NIOSH notes that speech requires shouting at 3 feet
110 dB SPL and up	Sirens, close-range power tools	Discomfort; hearing protection belongs here

NIOSH sets its recommended exposure limit at 85 A-weighted decibels averaged over eight hours, with the allowed time halved for every 3 dBA increase. For readings with safety, legal or compliance consequences, use a class 1 or class 2 sound level meter conforming to IEC 61672-1 with a current calibration certificate. If loud exposure worries you, see an audiologist rather than relying on a browser measurement.

Accuracy, limits and troubleshooting

• Uncalibrated microphones. Browsers expose no sensitivity or gain metadata. An uncalibrated phone or laptop microphone can sit 10 dB or more away from a reference meter, and its frequency response is not flat, typically rolling off below 100–150 Hz and above 10 kHz.
• No weighting, no certification. IEC 61672-1 defines what a class 1 or class 2 sound level meter must do, including frequency weighting and tolerances. A browser tab cannot substitute for one in any regulated context.
• Operating system processing. Windows audio enhancements and macOS voice isolation can compress, equalize or gate the signal before samples reach the page. Turn them off for more faithful readings.
• Bluetooth input. A headset in call mode often switches to a narrowband mono profile that removes bass content. Use the built-in microphone or a wired interface when it matters.
• Position and handling. Hands, cases and desk reflections change what the capsule hears by several dB. Keep the device clear of your hand and off hard surfaces.
• A pinned 0.0 dBFS means the input is overloading, and the true level is unknown. Move further away or lower the input gain, then read again.
• Simulation is not audio. Simulate Audio is a scripted display for testing the interface, not a measurement.

Related tools

• Loudness Meter — measure integrated, short-term and momentary LUFS of a file with the ITU-R BS.1770 algorithm.
• Audio Spectrum Analyzer — see which frequency bands are producing the level you are reading.
• Hearing Test — a screening staircase that maps heard thresholds from 125 Hz to 8 kHz; informational only, never a diagnosis.