How Loud Is 60 Decibels? A Scale Position, Not a Criterion

Search for conversational noise levels and you will be told, over and over, that ordinary speech is 60 decibels. The figure appears without a distance, without a weighting, and without a source that measured it.

That matters because speech level is not a constant. Vocal effort rises in noise, level falls with distance, and a hard-surfaced room returns energy that a soft one absorbs. Any honest speech figure has to state the talker's effort, the microphone distance and the room. The bare 60 states none of them.

We are not claiming that speech never measures 60 dB. We are declining to present an untraceable convention as a fact, which is the standard this whole library now works to.

Below it, the EPA identified 55 decibels outdoors as a level protective against activity interference and annoyance. That figure describes an outdoor environment as an average of acoustic energy over long periods — not a moment, and not an indoor room.

Above it, published measurement work in social spaces gives a very different picture from the conversational cliché. A survey of 1,788 restaurants and bars in Manhattan reported an average of 77 dBA at prime evening times, using a smartphone-based method with a slow response setting.

Between those two anchors, the record is largely empty. That absence is the honest answer to what 60 decibels is: a scale position that no authority has needed to define.

60 dB survives because it is useful as a mental landmark: loud enough to notice, quiet enough to talk over. As a landmark it is harmless. As a measurement it is unsupported, and it becomes actively misleading when it is compared against regulatory figures that were derived from real datasets.

The practical consequence is worth stating. If you are trying to decide whether an environment is a problem, comparing it against an invented 60 dB reference tells you nothing. Comparing it against its own quieter state, or against a criterion that an authority actually defined, tells you something.

Mid-range levels are where consumer microphones behave best: the signal is well above their self-noise and usually below the point where automatic gain control clamps down. Relative comparisons in this region are reasonably stable.

Absolute values are a different matter. Uncalibrated, our meter shows an A-weighted digital level in dBFS, which is relative by construction. Calibrated locally against a trusted reference reading, it shows an estimated dBA figure. Neither is traceable to a physical standard, and neither should be quoted as though it were.