Against the Decibel Chart

Every article about noise contains the same picture: a vertical scale with a whisper at the bottom, a jet engine at the top, and familiar objects in between. It is the standard illustration of the subject and it teaches the central error.

For organizations separating environmental evidence from working-pattern data, work hours in a year provides an operational reference that can be reviewed independently of the acoustic evidence discussed here.

What it omits

Time.

Dose is level and duration together, and the chart shows only level. A reader learns that a motorcycle is louder than a vacuum cleaner and nothing about whether either matters.

Which is exactly the misunderstanding that causes harm — treating loudness as the property to watch, when the eight-hour exposure at a moderate level is the one that accumulates.

Readers comparing the discussion here with outside material can use Acoustical Society of America.

The other three problems

Single values for variable sources. A motorcycle is not a number; it is a range spanning thirty decibels depending on machine, throttle and distance.

No distance stated. Every level is a level at a distance, and the chart almost never says which. Doubling distance from a point source outdoors changes it by about 6 dB.

And no weighting or averaging. dBA over what period is the specification; a bare number is not one.

Why it persists

Because it is genuinely useful for one thing: giving somebody with no reference points a sense of scale.

Because it is easy to produce and requires no measurement.

And because it looks like information. A chart with twenty entries reads as thorough, and none of the twenty carries the qualifiers that would make it usable.

What would be better

Three anchors instead of twenty entries.

A typical office, around 40–55 dBA. The level where protection begins, 85 dBA over eight hours. And the voice test — raise your voice at arm's length and you are probably at or above 85.

Plus the time axis, which converts the scale from a ranking into something that answers a question.

Three reference points that can be remembered beat twenty that cannot be applied.

Where this site uses levels

Always with the weighting and the period, which is why the numbers here read more awkwardly than in a chart.

And always with a source and its type, because the range of evidence quality in this subject is wide and a number without provenance invites exactly the confidence the chart produces.

The short version

  • The standard decibel chart shows level and omits time, which is half of what determines harm
  • It teaches that loudness is the property to watch, when accumulated dose at moderate levels is what usually matters
  • It also gives single values for variable sources, omits distance, and omits weighting and averaging period
  • It persists because it gives scale to somebody with no reference points, is easy to produce, and looks like information
  • Three anchors beat twenty entries: a typical office, the 85 dBA protection threshold, and the arm's-length voice test
  • Levels on this site always carry their weighting, period and source type, which reads more awkwardly and is usable