How Loud Is Traffic Noise?

Almost every other page on this site can point at an object and ask how loud it is. Traffic noise is different in kind. It is what you get when many sources of different types pass at different speeds in different numbers, at a distance that varies with where you stand, filtered by whatever is between you and the road.

This is not a technicality. Agencies that manage road noise professionally do not have a number for it either — they have models. If a chart hands you "traffic: 85 dB", it has silently chosen a traffic volume, a vehicle mix, a speed and a distance, and it will not tell you which.

Where hard, checkable numbers do exist, they apply to individual vehicles under standard test conditions — because that is something you can enforce.

US federal regulation caps new heavy trucks over 10,000 lb gross vehicle weight rating at 80 dB(A), measured 50 feet away under the SAE acceleration test procedure, effective from 1988. Separate in-use limits apply to interstate motor carriers: 83 dB(A) at speeds below 35 mph, 87 dB(A) above 35 mph, and 85 dB(A) stationary, all at 50 feet, effective from 1986.

For roads rather than vehicles, US practice uses the Federal Highway Administration's Traffic Noise Model. Its inputs are vehicle-class reference energy mean emission levels, defined at a reference distance of 50 feet (15 metres), which are then combined with traffic volume, speed, road geometry, ground effects and barriers to predict a level at a receiver.

The important structural point is that the emission level is per vehicle class and per speed, and everything else about the situation is applied on top. There is no step in that process where a single number for "traffic" appears.

For land-use decisions, the Federal Highway Administration's noise abatement criteria give thresholds by activity category — for residential-type land use, around 67 dB(A) as an hourly equivalent level. That is a policy threshold for triggering abatement, not a measurement of any road.

Heavy-vehicle share matters more than total volume for many roads, because a truck emits far more acoustic energy than a car at the same speed. Speed matters, and its effect is not linear: at low speeds powertrain noise dominates, while above roughly 50 km/h tyre-road noise takes over, which is why quieter pavements are a serious engineering strategy.

Distance and screening matter, and not in the simple way people expect. A road is closer to a line source than a point source, so level falls off more slowly with distance than the textbook 6 dB per doubling. Barriers, buildings, ground cover, gradients and wind direction all shift the result further.

We could not extract the Traffic Noise Model's per-speed emission values for each vehicle class from the published figures, so we are not quoting them. Nor did we find a field measurement study we could verify giving levels at stated distances from a described road.

We also chose not to do the obvious thing and convert a truck's 80 dB(A) limit at 50 feet into a street level using a propagation formula. The limit applies to one vehicle under a test procedure; a road is a stream of mixed vehicles, and the arithmetic would give a confident answer to a question the source never addressed.

This is the rare topic where measuring it yourself is genuinely the better route, because the thing you want to know is local and no published figure can supply it.

Our browser meter is built for that comparison. Uncalibrated it reports A-weighted digital signal level in dBFS, a relative number rather than sound pressure; after a one-time calibration against a trusted meter it reports an estimated dBA figure. Since traffic noise fluctuates constantly, the meter's equivalent-level function over a period is far more meaningful than any single instantaneous reading — and comparing front room with back room, or window open with window closed, is something it can do honestly.