What is SPL? Sound pressure level and decibels explained

From pascals to decibels

Sound is a quick, small swing in air pressure above and below the steady pressure of the atmosphere. The quietest sound a young, healthy ear can pick up at mid frequencies moves it by about 20 micropascals, less than a billionth of atmospheric pressure. Even a sound loud enough to hurt, around 130 dB, changes it by less than a tenth of one percent.

SPL turns that pressure into a level: 20 times the base-10 logarithm of the measured pressure divided by 20 micropascals, with the pressure taken as its RMS value. The reference itself is 0 dB SPL, one pascal is 94 dB and 20 pascals is 120 dB. From 0 to 120 dB the pressure grows a million times, and the logarithm turns that range into numbers you can add and subtract.

The 94 dB figure turns up on spec sheets because it equals one pascal. A microphone's sensitivity is its output at that level, and the Austrian Audio OC16 puts out 11 mV per pascal.

Decibel arithmetic

Power and intensity go with the square of pressure, so one step in decibels means a different multiple for each:

For amplifiers, doubling the power buys 3 dB, and ten times the power buys 10 dB, which listeners hear as roughly twice as loud. A move from a 50-watt amplifier to a 100-watt one raises the loudest clean level by 3 dB.

Levels from separate sources don't add like ordinary numbers. Two sources of equal level playing unrelated sound, two fans or two voices, add 3 dB, so 80 dB and 80 dB make 83. A source 10 dB quieter than another raises the total by less than half a decibel, and the loudest source in a room sets the level.

What dB(A) and dB(C) mean

A plain SPL figure counts every frequency equally. Your hearing doesn't: at quiet and moderate levels it's far less sensitive to deep bass than to the midrange. A-weighting builds that into the reading, turning down low frequencies and the very top roughly as the ear does, so a 50 Hz tone reads about 30 dB lower in dB(A) than in plain dB SPL. C-weighting is nearly flat. It suits loud sounds and peaks, where the ear's own response flattens out too.

Noise-exposure limits are usually quoted in dB(A), and so is a microphone's self-noise, which the noise floor explainer covers. A figure given as plain dB leaves the weighting unstated, and on bass-heavy sound the readings can sit far apart.

SPL figures on a spec sheet

Speaker sensitivity is the SPL measured on axis at one meter from one watt, or from 2.83 volts, which is one watt into 8 ohms. Into 4 ohms, 2.83 volts is two watts, so on a 4-ohm speaker the two readings sit 3 dB apart. The Diora Acoustics Chors 5, a 4-ohm speaker, is rated at 90 dB, and which reference that 90 dB uses decides whether it's 87 or 90 dB per watt.

From there it's addition. A speaker rated 90 dB for one watt reaches 110 dB at one meter on 100 watts, 20 dB more for a hundred times the power. Each doubling of distance in open space takes 6 dB off, so at 3 meters it makes about 100 dB, and the room's reflections give some of that loss back. The SPL distance calculator runs the sum with your own figures, and the best speakers guide ranks the speakers reviewed here.

Headphone sensitivity does the same job at the ear, in dB SPL per milliwatt or per volt, and the sensitivity explainer converts between the two. Most headphone reviews here also work out the power a model needs for 110 dB SPL peaks.

A microphone's maximum SPL is the loudest sound it takes before its own distortion passes the limit its maker has set. How much you need depends on what goes in front of it. The OC16's review puts a snare two inches away at a peak around 125 dB and the inside of a kick drum around 140, the jobs a pad switch exists for: it cuts the signal before the microphone's electronics can overload. At 148 dB the OC16 has no pad and doesn't need one. The FIFINE K688 and the t.bone SCT 2000 both top out at 130 dB, and both reviews find that plenty for loud voices.

How long you can listen at a given level

The risk to your hearing grows with both the level and the time you spend at it, so safe-listening limits are written as a dose. The WHO-ITU standard for personal audio devices gives adults a reference of 80 dB for 40 hours a week. The dose counts sound energy, so every 3 dB louder halves the time: 83 dB for 20 hours, 86 dB for 10, and at 100 dB the week's allowance is gone in well under an hour.

Because the dose adds up energy over time, a steady loud level uses it far faster than a few loud moments in otherwise moderate music. Blocking outside noise helps too: with a good seal or noise canceling, you can follow music on a train at a lower level.

The safe-listening calculator estimates the level at your ear from your headphone's sensitivity and impedance, your source's output voltage and the volume setting, and says how long that level is safe each day by a workplace noise limit.