The distance term beats the power term and it is not close
Everything in the level equation is decibels being added and subtracted, but the three terms behave very differently. Power gives you ten times the log of the wattage, so going from 500 watts to 1,000 watts is worth 3 dB — a change most listeners describe as slightly louder. Doubling the number of identical boxes is worth about the same 3 dB, on the assumption they add in power rather than interfering with each other. Distance takes away 6 dB every time it doubles, in free air.
Put numbers on it. A 96 dB box on 500 watts makes about 124 dB at one metre. At 60 feet it is down to around 98. Going to 120 feet costs another 6, and getting that 6 back by power alone means four times the amplifier: 2,000 watts into a box that probably cannot take it. Getting it back by moving the box 60 feet closer costs nothing and is why delay rings exist.
The 6 dB figure is the free-field one — a field, a car park, a roof with nothing near it. Inside a room it is only true close to the box. Past the critical distance, where the reflected energy in the room matches the direct sound from the speaker, the level stops falling with distance at all, and the practical figure people measure in a reverberant hall is often 3 dB per doubling or less. That is why the field on this page is an input rather than a constant. If someone has walked the room with a meter, their number beats the physics textbook.
Crest factor is the number nobody puts on the spec sheet
An amplifier is sized by the peak and the audience hears the average, and the gap between them is the crest factor of the material. Heavily limited playback might run 6 to 10 dB of crest. A live band with an actual dynamic range runs 12 to 18. Uncompressed orchestral material can be more.
This is what makes the target-level section come out so expensive. Asking for a 95 dB average with 12 dB of crest is asking the system to make 107 dB cleanly, and at 60 feet on a 96 dB box that is thousands of watts. The alternatives are the ones that actually get used: more boxes, boxes closer to the people, or a system that is allowed to run out of headroom on peaks and be limited. All three are decisions. Pretending the average is the whole story is not.
What sensitivity figures are worth
Published sensitivity is measured differently by different manufacturers — swept or pink noise, on axis or averaged, full range or band-limited — and the number quoted is sometimes a peak rather than a continuous figure. A one or two decibel difference between two spec sheets is inside the measurement noise. A six decibel difference is real, and worth asking about.
The bigger omission is directivity. A box has a coverage pattern, and the sensitivity figure describes the middle of it. Somebody standing 40 degrees off axis at the same distance is hearing considerably less, and how much less is in the polar data if the manufacturer publishes it. On a wide audience the off-axis seats are the ones that decide whether you needed another box, and no single-number calculation sees them.
Where this stops being arithmetic
Everything above treats the speaker as a point source in open air. Real systems break that assumption in useful ways. A line array is designed so that the sources interfere constructively over a distance, which flattens the falloff to something closer to 3 dB per doubling in its near field — that is the point of it. Horn-loaded boxes concentrate output into a defined pattern rather than radiating everywhere. Both make the simple sum optimistic in some directions and pessimistic in others.
Weather matters outdoors more than people expect. A temperature inversion over a field bends sound back down and the level at the far fence goes up after sunset. Wind does the same thing in one direction and the opposite in the other. If the show is long and the level is set at soundcheck in the afternoon, it is not the same show at ten at night.
When a second speaker position enters the plan, timing becomes the problem rather than level, and the speaker delay time calculator handles it. For what the amplifiers draw once the system is specified, the event power distribution calculator and the AV rack power and heat calculator cover the electrical side.
Questions people ask
How much does sound drop off over distance?
In free air, 6 dB for every doubling of distance — so 20 feet is 6 dB down on 10 feet, 40 feet is 12 dB down, 80 feet is 18 dB down. That comes straight from the energy spreading over a sphere that grows with the square of the radius. Indoors it only holds near the speaker. Once you are far enough into the room that reflected energy matches direct sound, the level flattens out, which is why the back of a reverberant hall can be loud and still hard to understand.
Does doubling the amplifier power make it twice as loud?
No. Doubling electrical power adds about 3 dB, which most listeners describe as a small increase rather than a doubling. The rough rule people use is that something like 10 dB is needed before a listener calls it twice as loud, and 10 dB is ten times the power. That is the arithmetic that sinks the idea of solving coverage with a bigger amplifier: the amplifier runs out long before the distance does, which is why extra boxes closer to the audience are the usual answer.
What does 96 dB 1W/1m actually mean?
It is the sound pressure level measured one metre in front of the speaker with one watt driven into it, and it exists so that boxes can be compared on a common basis. To get from it to a real level, add ten times the log of your actual wattage and subtract the distance loss. It is a single on-axis number, so it says nothing about what happens off to the side, and manufacturers measure it in different ways — treat small differences between spec sheets as noise.
Why does the calculator ask for crest factor?
Because the amplifier has to reproduce the peaks while the audience hears the average, and the gap between them is not small. Material with 12 dB of crest needs a system capable of 107 dB peak to deliver a 95 dB average. Ignoring that is how systems get specified that measure fine on a steady tone and clip on the first snare hit. If you compress the material to shrink the crest factor you get level back, and you pay for it with the dynamics.
How loud is too loud for an event?
That is not a question a calculator should answer, and the numbers involved are set by regulators rather than by physics. Occupational exposure limits, local noise ordinances and venue conditions all vary by jurisdiction and all change. What is worth knowing is the shape of the risk: hearing damage from sustained exposure is cumulative, permanent and painless while it happens, so nobody notices at the time. If the show is loud enough for the question to arise, it warrants a meter and someone who knows what the local rules are.