From acoustic watts to a decibel reading

This is the set for the moment the numbers arrive in the wrong currency. A fan curve, a pump data sheet or an intensity survey hands you watts; the specification, the bylaw and the complaint are all written in decibels. Everything here is the crossing between the two. Sound intensity is the definition — acoustic power divided by the area it passes through — and its most useful direction is backwards: scan a probe over a surface enclosing a machine, multiply the average intensity by the area of that surface, and you have the machine's sound power in watts, which is the one figure that does not depend on where the microphone was standing. Before anything else, kill the confusion this set sits next to. Acoustic power is not electrical power. A fan drawing 500 W from the wall radiates a small fraction of a watt as sound — a thousandth of the input or less is entirely ordinary — and a data sheet printing both numbers on the same page invites exactly that substitution. If a sound figure looks suspiciously like the motor rating, it is the motor rating.

The inverse-square law is that same definition with the area chosen for you: a point source's power spread over an expanding sphere, so intensity falls as 1/r². The two are not one equation written twice, because the areas are different objects — a measurement hull wrapped around the machine in the first, an imaginary sphere out at the listener in the second — and the power the first one recovers is literally the P typed into the second. What comes out is an intensity in W/m² at the distance you care about, and it is still not a decibel.

Decibel sound level is where the arithmetic changes character, and it hides something in plain sight worth slowing down for. β = 10 log₁₀(I/I₀) compares your intensity against I₀ = 10⁻¹² W/m², and that number is a choice, not a constant of nature. It is not a property of air. It does not fall out of any equation. It was picked because it sits at roughly the quietest sound a healthy young ear detects at 1 kHz — so the decibel scale, and every noise regulation, hearing-protection rule and neighbourhood bylaw built on top of it, is anchored to a property of people rather than a property of the medium. Every decibel elsewhere on this site is quietly measured against it. The familiar pressure reference of 20 μPa is the same choice said a second way: 20 μPa squared, divided by the impedance of air at about 415 rayl, comes to 9.6×10⁻¹³ W/m², which is why the intensity form and the pressure form agree in air and only in air. Change the reference and every number moves with it — an underwater level quoted in dB re 1 μPa describes the same physical sound about 62 dB away from its airborne figure, and comparing the two without that correction is a mistake that reaches print regularly.

Combining sound levels closes the set, and this is the one place in the catalog where the rule stops looking strange. Two 60 dB machines make 63 dB, not 120, and everyone is told to memorise that. Coming through this set you can see why instead: you have been carrying intensities the whole way, intensities are energies per unit area and simply add, and the only reason the addition looks exotic is that a logarithm was taken first. Convert back, add, take the log again — which is precisely what the formula does. Two limits to carry out of here. Everything above is a free-field answer that assumes the sound leaves and never returns, so indoors it holds only close to the machine and the reverberant field governs beyond that. And the level this chain produces is unweighted: a limit written in dB(A) wants A-weighting applied band by band before comparison, and a broadband figure computed here is not that number.