By the Sound Decibel Meter Team

How to Use the Sound Decibel Meter: Complete User Guide

This guide walks through every feature of the browser decibel meter on this site — from a 30‑second first measurement to the advanced settings most users never touch. The meter runs entirely in your browser, requires no installation or account, and uses your device's microphone strictly for the duration of the session. By the end of this page you should know how to read the live display, choose the correct frequency and time weighting for your situation, calibrate your device for accuracy within a few decibels of a Class 2 reference meter, and trust the readings enough to base hearing decisions on them.

If you only have a minute, jump to Quick start. If you want the full picture, read straight through.

Quick start

Open the home page and click Start.
Allow microphone permission when your browser prompts. The meter does not record or upload anything — see the privacy section below for the technical details.
Read the live decibel value at the center of the screen. The meter updates several times per second; the number you see is an A‑weighted, fast‑time‑weighted SPL value.

That is all you need to take a first measurement. The rest of this page explains how to make that measurement more accurate, how to record sustained exposure, and how to interpret the readings against published health and regulatory thresholds.

Permission and privacy

The meter needs getUserMedia permission to read your microphone, and your browser will prompt you the first time you start it. There are two important properties of how this is implemented on our site:

All audio processing happens locally in the browser, in JavaScript and the Web Audio API. No PCM samples, no FFT frames, and no derived numbers are uploaded to any server. The site has no audio analytics endpoints because it has no server‑side audio handling at all.
HTTPS is required. Browsers refuse to grant microphone permission to pages served over plain HTTP. The site uses HTTPS by default and will redirect HTTP requests.

You can revoke microphone access at any time by clicking the lock icon in your address bar (Chrome, Edge, Brave, Firefox), the camera icon (Safari), or by clearing site permissions in browser settings. Closing the tab releases the microphone immediately.

If you are on a corporate network or a managed device that blocks getUserMedia, you will see "no microphone detected" — there is no workaround on our side; you'll need to use a personal device or ask IT for the policy exception.

Reading the display

The meter's main panel shows several distinct values at once. They look like one big number with some statistics below, but each carries different information.

Current level (the big number)

The large central reading is the time‑weighted SPL at this instant. "This instant" is qualified by the time weighting (Fast, 125 ms; Slow, 1 s; Impulse, 35 ms — see Time weighting below) and the frequency weighting (A, C, or Z — see Frequency weighting). The default settings are A‑weighted and Fast, which is the right combination for most environmental and occupational measurements.

The number's color reflects the band it falls into. Green is comfortable, amber is loud‑but‑safe with limited exposure, red is hazardous. The color mapping uses the same boundaries as the reference chart shown beneath the meter.

Min, Avg, Max, Peak

Below the current level, four smaller readings track the session statistics since you last clicked Start (or Reset):

Min — the lowest valid reading observed. The meter waits a short startup window before starting to track Min, so the very first reading doesn't pin Min at the system noise floor.
Avg — the arithmetic average of all readings. Useful as a quick proxy for Leq when the noise is reasonably steady; for a true Leq measurement of variable noise, see the tips page.
Max — the highest reading observed. With Fast time weighting this catches sustained loud events; with Impulse it catches transients (gunshots, hammer strikes, doors slamming).
Peak — the highest unweighted sample value, before time weighting. The meter holds the peak briefly so you can see brief transients that would otherwise vanish.

The four colors update independently to reflect the band each value falls into — so Min might stay green while Peak is briefly red, telling you that the room is mostly quiet but had at least one loud event.

Reference chart

To the right of the live values is a vertical reference chart with bands from "threshold of hearing" up through "fireworks." A check mark next to a band indicates that your current reading falls inside that band. The band's color and example text help you sanity‑check the measurement: if the meter says 30 dB but your environment matches the description for 70 dB, your calibration is probably off (see Calibration).

Banner notification

Above the chart, a contextual banner appears whenever your current reading lands in a meaningful band. It shows the matched range and a short health advisory ("Comfortable for normal activity," "Prolonged exposure may cause fatigue," etc.), with the dynamic values colored to match the band.

Settings

The settings panel (gear icon, top‑right of the meter) groups the parameters into four sections. Defaults are sensible for almost every use; the explanations below cover when to change them.

Frequency weighting (A / C / Z)

A‑weighting de‑emphasizes low frequencies in a way that approximates human hearing at moderate listening levels. It is the standard for occupational and environmental noise (OSHA, NIOSH, WHO, ISO 1996). Choose A unless you have a specific reason not to.

C‑weighting is much flatter — it preserves low‑frequency content, which makes it the right choice for measuring concert sound, subwoofer output, and impulsive noise. If you are measuring a venue where the bass is the main concern, C‑weighting will give a higher and more representative number than A.

Z‑weighting is unweighted (flat) — used for instrument verification and acoustic research. Most users will never need it.

The math behind the weightings is on the science page.

Time weighting (Fast / Slow / Impulse)

Time weighting is an exponential averaging applied before the value is displayed:

Fast (125 ms) — the standard for most environmental and occupational measurements. Captures speech‑rate variation and short events without flickering.
Slow (1 s) — smoother readings for steady ambient noise. Useful when the source is roughly constant and you want a stable number to log.
Impulse (35 ms) — much faster response, designed to catch transients like gunshots, hammer strikes, and explosions. Reads noticeably higher than Fast for short events.

When in doubt, use Fast. Switch to Impulse only when you specifically need to characterize transient events.

Scale (dB SPL vs dB FS)

dB SPL (sound pressure level) is the standard physical scale, calibrated to 20 µPa = 0 dB. This is the scale you want for any reading that should match published references — including the entire comparison chart.

dB FS (full scale) is a digital‑audio scale where 0 dB FS corresponds to the maximum representable digital sample value. dB FS readings are useful for audio engineers comparing levels in a DAW but cannot be directly compared to physical SPL numbers without a reference calibration. Most users should leave the scale on dB SPL.

Calibration offset

A single number (in dB) added to every reading to compensate for your device's hardware bias. Phones and laptop microphones can be off by ±10 dB straight out of the box. The full procedure for setting this correctly is on the calibration page, but the short version is: pair our meter with a known reference (a Class 2 SLM, a calibrator at 94 dB SPL, or a quiet space at a known ambient level), average for 30 seconds, and enter the difference here.

Update interval and peak hold

The update interval controls how often the live number refreshes. The default of 200 ms (Fast) is a good compromise between responsiveness and visual stability; 50 ms (Realtime) is jumpy but precise; 1 s (Stable) smooths variable noise.

Peak hold time controls how long a peak is displayed before it decays. Default is 3 seconds; Infinity holds the peak until you reset.

Graph types

Below the live values, a graph provides a temporal or spectral view of the signal. Four graph types are available:

Line — running time series of the level for the last several seconds. The line and the area beneath it are colored by the per‑sample band (green / amber / red), so you can see at a glance when exposure crossed thresholds.
Bar — a single live bar, colored by the current band. Cleanest view for at‑a‑glance reading.
Visualizer — a frequency‑domain bar chart (FFT spectrum), useful for identifying which frequency bands contribute most to the level. Bars are colored by intensity, low to high.
Spectrum — same FFT data as Visualizer but rendered as a line with a smooth gradient fill. Useful for spotting tonal components.

The line graph is the right default. Switch to visualizer or spectrum when you need to know whether a measurement is dominated by low‑frequency mechanical hum, high‑frequency hiss, or a particular tonal source.

Mobile vs desktop

The meter behaves identically on iOS Safari, Android Chrome, and desktop Chromium / Firefox. A few practical differences:

Microphone gain varies dramatically across devices. A phone that reads 65 dBA in a quiet office may read 75 dBA on the next phone in the same room. This is not a bug; it is a hardware difference. Set the calibration offset per device.
Battery saving and tab backgrounding can throttle audio processing. iOS specifically suspends Web Audio when the tab is backgrounded for more than a few seconds. Keep the tab active for long measurements, or use the screen‑lock prevention some browsers offer for fullscreen mode.
Bluetooth headsets can confuse the input selector. If readings are oddly low or muffled, disconnect the headset and force the meter to use the built‑in microphone.

For long unattended monitoring, a desktop with the laptop lid open is more reliable than a phone in standby.

Troubleshooting

"No microphone detected" on Start

Causes, in order of likelihood:

Permission denied — click the lock icon next to the URL and reset microphone permission, then reload.
No microphone in the device — check that the system audio settings show an input device.
Browser policy on the page — uncommon, but some embedded browsers (in‑app webviews) block getUserMedia regardless of permission.
Hardware mute — many laptops have a physical mic kill switch.

Reading stuck at 0 dB or extremely low

Usually a hardware mute or a wrong input device. Check system sound settings for the active input. On laptops with multiple microphones, some browsers default to the wrong one.

Wildly fluctuating readings

Most often handling noise — the meter is picking up taps and rubbing on the device, not the room. Place the device on a stable surface, on a soft cloth if possible, with the microphone facing the source. Keep wind off the microphone (a foam pop filter or a parked location is enough).

Numbers don't match a reference SLM

Almost always calibration. Follow the calibration procedure once per device and the readings should land within ±2 dB of a Class 2 meter for steady mid‑range sources.