How Loud Is 90 Decibels? A Regulatory Limit, Not a Sound

OSHA's permissible exposure level is 90 dBA over an 8-hour shift. Table G-16 then pairs shorter durations with higher levels, reflecting a 5-decibel exchange rate: each 5-decibel increase halves the permitted duration.

The table also sets a ceiling for impulsive or impact noise, which may not exceed a 140-decibel peak sound pressure level. That is a separate provision from the averaging table, and it is measured as a peak rather than as an average.

Two details are easy to miss and both matter. The level is specified as A-weighted with slow response, so a reading taken with different weighting or a faster time constant is not the quantity the table governs. And every row is an average across the stated duration rather than a level you may reach, so a shift that spends a few minutes far above the row figure can still average below it — which is exactly why assessment uses a dosimeter worn for the whole shift rather than a spot reading.

NIOSH's recommended exposure limit is 85 dBA as an 8-hour time-weighted average, with a 3-decibel exchange rate. The 3-decibel rate follows the physics of equal energy: a doubling of sound energy corresponds to about 3 decibels, so halving the time compensates for a 3-decibel increase in level.

OSHA's 5-decibel rate is more permissive, and its history is regulatory rather than acoustic. The result is that the same working day can be compliant under one framework and above the recommendation of the other.

The EU takes a third position. Directive 2003/10/EC sets a lower exposure action value of 80 dB(A), an upper action value of 85 dB(A) and an exposure limit value of 87 dB(A), with the limit value assessed taking hearing protection into account.

It is not a property of any object. Pages that tell you a particular machine "is 90 dB" have usually dropped the quantity, the distance and the operating condition that would make such a statement checkable.

It is not a personal threshold either. Occupational criteria are built from population datasets and an accepted degree of excess risk across a working lifetime. They describe policy, not the point at which one person's hearing changes.

And it is not comparable to declared appliance or equipment figures, which are usually sound power values filed under labelling rules. Sound power and shift-average sound pressure are different quantities entirely.

Higher levels bring a specific consumer-device failure: automatic gain control. Browsers and phones commonly reduce input gain as the signal rises, which compresses exactly the peaks you are interested in. Where the browser reports that gain control or noise suppression is active, our meter tells you, because in that state the signal is no longer proportional to sound pressure.

Even with processing off, an occupational assessment requires a dosimeter or a calibrated meter and a documented procedure. Comparing rooms or states with a phone is legitimate; documenting compliance with a phone is not.