Learn

Sound is one of the few things that affects nearly everyone’s health and almost nobody can describe in numbers. People know a room is too loud, that a street keeps them awake, that a show left their ears ringing — and then the conversation stops, because the language for saying so precisely belongs to acousticians and regulators.

Learn exists to close that gap. Our aim is to give venue operators, event producers, site managers, neighbours and regulators enough command of the subject to ask good questions, read a monitoring report without a translator, and make decisions about their environment on evidence instead of argument.

It will keep growing in two directions. Guides explaining the standards and the science, every factual claim traced to a primary source you can open and check yourself. And Beacons Labs — findings from our own monitoring: what continuous measurement actually shows about how sound behaves in real venues, on real streets, across a real show night. Field data, published the same way we publish everything else, with its methods stated and its limits admitted. Labs work is de-identified and aggregate, and site-specific data is only ever published with the operator’s permission.

Every noise limit, complaint report and compliance verdict is written in a handful of symbols. None of them are complicated once someone says plainly what they mean. This is the working glossary — the one we wish existed when we started.

Decibels don't add up the way you expect

The decibel scale is logarithmic, so arithmetic intuition fails. Two identical sources together are 3 dB louder than one, not twice as loud — +3 dB means double the sound energy. Subjectively, most people judge something as “about twice as loud” only around +10 dB, which is ten times the energy. This is also why averaging decibels arithmetically is wrong and quietly under-reports: a minute at 90 dB and a minute at 60 dB do not average to 75.

dBA, dBC, dBZ — frequency weighting

A microphone hears low bass perfectly well; human hearing does not, at least not at moderate levels. A-weighting (dBA) discounts low and very high frequencies to approximate how we actually perceive loudness, and is what almost every community and workplace limit is written in. C-weighting (dBC) is much flatter through the bass, so it keeps the low-frequency energy A-weighting throws away. Z-weighting (dBZ) is unweighted — the raw signal. The gap between them is the useful part: a large dBC − dBA difference means the sound is bass-heavy, which is exactly the sound that travels through walls and generates complaints while the dBA reading looks compliant.

Leq — the number most limits are written in

The equivalent continuous sound level: the steady level that would carry the same total sound energy as the real, fluctuating sound over a stated period. It is an energy average, which is why it is written with its duration — L_Aeq,15min is the A-weighted Leq over fifteen minutes. The WHO’s venue standard is set at 100 dB L_Aeq,15min for exactly this reason: brief peaks are expected at a concert, sustained energy is what damages hearing.

Lmax, Lmin — and why Lpeak is not Lmax

Lmax and Lmin are the highest and lowest levels the meter settled on during the period, using a stated time weighting (below). Lpeak is a different quantity: the true instantaneous peak sound pressure, with no time averaging at all. It is usually C-weighted (L_Cpeak) and is always higher than Lmax. Peak limits exist because a single impulse — a snare hit, a slammed door, a firework — can injure hearing in milliseconds, long before any average notices.

L10, L50, L90 — the percentiles

Statistical levels, describing the shape of the noise rather than its total. L10 is the level exceeded 10% of the time — the intrusive peaks, the part neighbours actually notice. L50 is the median. L90 is the level exceeded 90% of the time and is conventionally read as the residual or background level — what is left when the event stops. The spread between L90 and L10 tells you whether a place is steadily noisy or occasionally violent, and those two situations call for completely different remedies.

Ldn and Lden — the 24-hour indices

Long-term exposure metrics that penalise night. Ldn (day–night level) is a 24-hour Leq with 10 dB added to night hours, reflecting that the same sound does more harm when it interrupts sleep. Lden adds a third, evening period and is the metric European noise mapping and the WHO environmental noise guidelines are written in. A night penalty is not a technicality — sleep disturbance is one of the best-evidenced pathways from noise to cardiovascular harm.

Octave and third-octave bands

A single number cannot tell you what is loud. Splitting the spectrum into octave bands (centred 31.5 Hz through 16 kHz) or the finer third-octave bands shows where the energy sits — which is how you tell a bass-heavy PA from a noisy generator from crowd noise, and how per-frequency limits are enforced when a flat dBA cap won’t do the job.

Fast and Slow — time weighting

How quickly the meter responds. Fast uses a 125 ms time constant and follows transients closely; Slow uses 1 s and smooths them out. The same sound measured Fast and Slow gives different Lmax values, which is why a defensible report always states the weighting. When a number is quoted without one, it is not yet a measurement.

One rule underpins all of it: a measurement is only as good as its provenance. A level without its weighting, its duration and its instrument is a number, not evidence — which is why the Portland study measured continuously at 217 sites instead of sampling with a handheld meter, and why the WHO standard asks venues for calibrated equipment and a named person watching it.

Next up: how U.S. noise limits actually work (EPA vs. WHO), nighttime noise and sleep, and the first Beacons Labs write-up on what continuous monitoring shows across a full show night. Questions you’d like a guide to answer? Tell us — the roster is shaped by what people actually ask.