For the most part, the processing of audio signals can be performed with only minuscule power, either input or dissipated. Analog signals pass through the majority of the overall signal path at average levels in the order of 100mV to 1 volt.
Load impedances may be as high as 100kΩ but even if as low as 5kΩ, only 120μW (a hundred and twenty microwatts; or about a tenth of a thousandth of one watt) would be dissipated.
At this rate, it would take about eight million hours or hundreds of years of playing, for the load to absorb or use one unit (1kWh) of electricity! Most loudspeakers used to reproduce audio are highly inefficient.
Typical efficiencies of common direct radiating speakers are 1% to 0.05%. By comparison, the efficiency of an internal combustion engine (considered highly inefficient by ecologists) is between 2500% and 50,000% greater.
A medium sized car uses about 70kW to move 4 people or hundreds of pounds of goods, and its own weight – altogether at least half a tonne, at speeds of say 70mph. In some sound systems, to move just the weight of air molecules to reproduce a bass drum, as much as 7kW of electrical ‘fuel’ can be burned in bursts.
And yet a loudspeaker only needs to convey 1 acoustic watt to the air to recreate music at the highest practical sound levels in a domestic space, i.e. about 120dBSPL. And a tenth of this level (0.1 acoustic watts) will still suit most of the loudest passages in the less extreme forms of music.
If speaker efficiency is taken as 0.1%, and 1/10th of an acoustic watt is enough, then an electrical input power of 1000 times this is needed, i.e. 100 watts.
The highest SPLs in music can be considerably greater than 0.1 acoustic watt. Loudspeaker drive units exist that can handle short term electrical power bursts (the norm in much music) of 5000 watts (5kW) or more. With 2% efficiency, today’s most capable drivers can generate 100 acoustic watts each.
With horn loading, efficiency can be raised to 10% or more, allowing one drive unit to produce 500 acoustic watts for large scale PA. This allows fewer sound sources to be used, improving quality.
When comparing SPL figures it is helpful to remember that at medium SPLs (sound levels) and mid frequencies, a tenfold increase in watts offers only an approximate doubling in loudness to the ear. But at the lower bass frequencies and at higher SPLs, considerably smaller changes in wattage, say just x3 to x5, have the same doubling effect.