What Size PA Speaker Do You Need? Size It Backwards From Required SPL

• ▸Measure the actual path from the speaker position to the furthest seat, in metres. Not the room length — the path, which in a balcony room is longer than you think.
• ▸Take the peak SPL that seat needs from the programme table below.
• ▸Add the distance loss, which is 20 times log10 of the distance in metres. At 10 m that is 20 dB, at 15 m it is 23.5, at 20 m it is 26, at 30 m it is 29.5. The sum is the level required at one metre from the box.
• ▸Subtract the cabinet's sensitivity and convert: watts equals 10 raised to the power of (level required at 1 m minus sensitivity), all divided by 10.

Every step is arithmetic you can do on a phone, and what comes out is a wattage rather than a shortlist. That matters, because a wattage can be checked against a spec sheet and a shortlist cannot.

Crest factor is the gap between average and peak, and it is the number that gets skipped. Speech and live acoustic material run 12 dB or more between the two. Heavily compressed dance tracks run closer to 8 or 10. Size a system to the average and nothing else, and it clips on every transient — which is what people mean when they say a rig sounds harsh at volume.

The noise floor column matters as much as the target. Speech intelligibility needs roughly 25 dB of signal above the background noise, so a gym with a 68 dBA floor needs 93 dBA of programme before anyone understands a word, no matter how good the speakers are.

Split that wattage roughly one third to the tops and two thirds to the subwoofers. The ratio surprises people every time, and it explains why a rig built entirely from full-range tops always sounds strained: the tops end up doing bass work they have nothing like the cone area for, and every watt spent down there is a watt not available for the vocal.

Where the headcount row and the room volume row disagree, follow the room. A hundred people in a 1,600 cubic metre hall is a 1,600 cubic metre problem, and the empty air still has to be filled.

PA cabinets are almost always 8 ohms, which is the one corner of loudspeaker specification where the reference conditions happen to agree: 2.83 V into 8 ohms is exactly 1 watt, so both figures give the same number. Where you do meet a 4-ohm PA box, usually a dual-driver subwoofer, a sensitivity quoted at 2.83V/1m was measured at 2 watts and reads 3 dB better than the same box's 1W/1m figure. Three decibels is a doubling of amplifier power, so it is worth checking which condition is printed before you spec the rack.

Weight is the specification nobody reads and everybody regrets. A 15-inch powered top at 28 kg goes on a pole once, by two people, and then stays there. If the system gets set up and struck weekly by volunteers, buy 12-inch tops and accept the 3 dB.

Adding cabinets does not add loudness the way people expect, and the amount you gain depends on whether the boxes couple.

Two tops flown or poled apart do not couple. Across the frequencies they cover, a wavelength is short compared with the spacing between them, so their outputs sum as power rather than as pressure and you gain 3 dB for the second box. Four tops gets you 6 dB over one. That is real, but it is a quarter of the loudness increase most people assume.

Two subwoofers stacked side by side do couple. Below about 100 Hz a wavelength is over 3.4 metres, so two cabinets 60 cm apart are effectively a single source at the same point, their pressures sum directly and you gain 6 dB. Four stacked is 12 dB over one. That is the entire reason sub arrays are always built touching rather than spread across a stage, and splitting a pair of subs to the two corners of a stage throws away 3 of those 6 dB before anyone has played a note.

The practical consequence: when a system is short on level, adding subs is roughly twice as efficient per box as adding tops.

Set it by the top's woofer diameter, not by taste.

An 8 or 10-inch top wants a 110 to 120 Hz crossover. A 12-inch top wants 100 Hz. A 15-inch top wants 80 to 90 Hz. Below those points the top's woofer is running out of excursion long before it runs out of thermal capacity, and every watt it spends there is a watt it does not have for the vocal range, where the audience is listening.

Use 24 dB per octave slopes both ways. A 12 dB slope leaves the top exposed to too much of the sub's band, and in a portable rig that gets loaded in by whoever is available, that exposure is what kills 12-inch drivers.

High-pass the tops even when there is no subwoofer at all. A 12-inch top rolled off at 60 Hz with a 24 dB slope gains meaningful headroom above 100 Hz and loses nothing anybody misses, because there was never usable output down there to begin with.

Two more that matter in a live room. Check subwoofer polarity against the tops at the crossover point, because a 180 degree error produces a deep notch in exactly the octave the kick drum lives in and it sounds like a bad room rather than a wiring fault. And keep the subs on the floor: a subwoofer lifted onto a stage loses the floor boundary that was giving it 3 to 6 dB for free.

The crossover point belongs to the processor either way. Powered boxes usually carry a built-in high pass with a preset for the matching sub; passive systems need a separate processor between desk and amps, and skipping it is the most common reason a passive rig underperforms a cheaper powered one.

• ▸Path distance from the intended speaker position to the furthest seat, in metres.
• ▸Room volume: length times width times average height, in cubic metres. Vaulted ceilings count, and they are why gyms and churches need more system than their floor area suggests.
• ▸Ambient noise floor with the room occupied and the HVAC running, on any sound level meter or a phone app calibrated against one. This sets the floor you have to clear.
• ▸Reverberation time by ear if you have no meter: clap once, hard, and count. Anything over about 1.5 seconds means intelligibility, not level, is your problem, and more watts will make it worse.
• ▸Ceiling height at the speaker position, which decides whether the tops go on poles, on brackets or flown.
• ▸Who is carrying the boxes, and how often.

If the clap test comes back long and ringing, stop the specification there. In a hard room, a bigger PA raises the reverberant field along with the direct sound and the ratio between them does not improve. The fix is directivity — narrower-pattern boxes aimed at the seats and off the walls — or treatment, and neither of those is a wattage problem.