Office Noise Level Checker

When you press Start, the tool opens your microphone with automatic gain control, noise suppression and echo cancellation switched off — those processors silently change the signal level, so leaving them on would make any workplace noise reading meaningless. It then reads the raw waveform many times a second, computes the RMS (root-mean-square) energy of each short frame, and smooths it with a roughly one-second time constant to give a steady level the way a sound-level meter’s “slow” setting does. This approach mirrors established acoustic comfort measurement practice used by workplace designers and WELL Building Standard assessments.

That level is reported in dBFS — decibels relative to digital full scale, where 0 is the loudest the system can capture and everything else is negative. dBFS is relative: it tells you reliably whether the room is louder or quieter than a moment ago, but on its own it is not an environmental decibel (dB SPL) reading, because the browser has no idea how sensitive your specific microphone is. To bridge that gap you can calibrate: while measuring, read the level on a real sound-level meter or a trusted phone app and type that dB(A) in. The tool stores the difference (offset = your reading − current dBFS) and from then on shows an estimated dB SPL = dBFS + offset. The offset is saved under a shared key, so calibrating here also calibrates the other noise tools on this site.

Once an estimated SPL is available, the tool drops it into the productivity guidance chart and shows a verdict band: Optimal below 50 dB, Moderate from 50 to under 70, and Detrimental at 70 and above. (The chart’s band boundaries are A-weighted reference values while this tool’s own reading is unweighted, so the comparison is approximate.) It also tracks Leq (the equivalent continuous level over the whole session, 10·log₁₀ of the mean of the per-sample energies), L90 (the level exceeded 90% of the time over a rolling last-~7.5-minute window, a good proxy for the steady background noise floor beneath conversation and clatter — similar to what open-plan acoustic design standards use to set HVAC masking targets), and the loudest and quietest samples since you started. For long-term noise dose, a separate noise exposure calculator can convert an estimated Leq into an 8-hour TWA. Before you calibrate, every figure is shown honestly as relative dBFS rather than an invented SPL number. Before/after comparisons with the same mic and the shape of the level over time are meaningful even without calibration; only the absolute dB SPL value depends on it.

Noise in an office is rarely a single broadband hum: it is a mixture of speech intelligibility spillover from neighbouring desks, HVAC rumble, keyboard clicks, and intermittent phone calls — each contributing to your cognitive load differently. Open office noise in particular — the combined ambient din of an open-plan floor — has a well-documented link between noise and concentration: sustained levels above 55 dB erode focus and increase reported stress. The productivity bands in this tool are most relevant to this combined ambient level; if you want to understand the frequency shape of your background noise (for example, to see how dominant low-frequency HVAC hum is), the audio spectrum analyzer can show that breakdown in real time.

There are three decibel scales in play when you use a browser-based noise tool, and they are not interchangeable. dBFS (decibels relative to digital full scale) is what an uncalibrated microphone reads: 0 dBFS is the highest level the digital system can capture, and every real-world sound falls somewhere below that, so all readings are zero or negative. It is a relative level — it tells you whether the room just got louder or quieter, but it carries no information about the actual sound pressure in pascals. This is what this tool displays before you calibrate it; think of it as a browser-based office sound meter that gives you relative trends accurately even without a hardware reference. dB SPL (sound pressure level) is referenced to 20 micropascals, the threshold of human hearing, and requires a calibrated measurement chain — typically a Type 1 or Type 2 sound-level meter with a traceable microphone — to produce a meaningful absolute figure. This is why a phone app or browser tool that shows a positive dB number without calibration is, at best, an estimate: it has applied a fixed internal offset to make the number look right, but that offset cannot account for your specific microphone's sensitivity.

dBA is A-weighted dB SPL: it applies a frequency-dependent filter that de-emphasises very low frequencies (below ~500 Hz) and very high frequencies (above ~6 kHz) to reflect how human hearing actually perceives loudness. OSHA, NIOSH, and nearly every national environmental noise limit — including the productivity bands shown in this tool's reference table — specify dBA figures, because a flat broadband reading would overstate the annoyance of HVAC hum and understate the impact of speech-range noise. This tool reads unweighted broadband level (no A-weighting filter is applied), so even after calibration its numbers will diverge from a true dBA reading in rooms dominated by low-frequency noise. For a certified dBA measurement, use a calibrated sound-level meter. To understand how your office noise breaks down across frequency bands, the noise floor analyzer can help; to convert a measured level into an estimated daily noise dose, see the noise exposure calculator.