Factors Affecting Perceived Sound Level
A good PA system should increase only the level of the sound source, and should do little to change it in any other way. While many would describe this as making it louder (and in one sense it is), it is important to realise that loudness is a word that describes our experience of sound, and that sound pressure level (SPL) is only one aspect of this. We may sometimes experience one sound as being louder than another when there is no difference in SPL (or even, in some cases, experience sound with a lower SPL as louder).
Where SPL is the only difference between them, two otherwise identical sounds will seem to differ in loudness where their measured SPL differs by more than around 3 decibels. While smaller differences (down to 1dB and below) can be perceived and measured, they are less readily noticeable. Put simply, for some people a difference of one decibel or less may not be experienced at all (unless they are looking at a sensitive meter).
Where two sounds are not identical, some of the non-identical aspects interfere with our experience so that our hearing becomes less reliable in determining overall level. The main culprits are:
One of the things that tells our ears that a sound is loud is distortion. When amplifier circuits are driven into distortion, they create high-frequency harmonics of the original sound. The extra high-frequency content makes the sound seem louder (see Frequency Content, below). Also, when an amplifier reaches distortion, its dynamic range is reduced: although the quietest sounds it reproduces are increased in level, the loudest sounds it can reproduce are limited by its output capability. This will also make it seem louder (see Average Level, below). These two factors combined can make an amplifier with limited power driven to distortion seem louder than an amplifier with greater output operated within its limits.
Some inexperienced guitarists will set the level of an ‘overdriven’ channel at a lower level than the ‘clean’ channel. Without any other cues, the distortion of the ‘overdriven’ sound makes it seem louder, where a meter shows it is in fact lower in level.
The ear can deal with a substantial range of sound pressure levels, from near-silence to the threshold of pain (the threshold of pain typically falls somewhere between 100dB and 140dB SPL, depending on frequency and varying between individuals). Where a sound is relatively constant in level, it may seem louder than a sound with greater variation in level, even though the sound with greater variation has a higher peak level. Sound pressure levels ranging between 100 dB and 104 dB (SPL) with an average level of 103 dB will generally seem a lot louder than levels ranging from 80 dB to 110 dB with an average level of 96dB, even though the peak level of the ‘quieter’ sound is a full 6 dB (the equivalent of four times more amplifier power) higher. For this reason compressors can make a mix seem louder - by increasing the level of the quietest sounds (and thereby increasing the average level) - without increasing the peak level. Note, however, that it is average levels which overheat speakers: increasing the average level increases thermal stress on speaker drivers. Compressors don't necessarily make your loudspeaker system safer.
Our ears do not hear all frequencies equally. They are most sensitive at around 3-4kHz, and much less sensitive at the extremes of frequency (in most concert systems the extremes will lie at around 40Hz and 16kHz). At the threshold of hearing the ear is around 60dB less sensitive at 40Hz than it is at 3.15kHz. Although the difference is less at high sound pressure levels (about 35dB at levels of 120dB SPL), it is still substantial. 110dB at 40Hz is loud, but 110dB at 3.15kHz is above the threshold of pain for many people.
A system with limited frequency response that has high output from 2-4kHz can seem uncomfortably loud, where a system with broader frequency response (and substantially higher acoustic energy overall) may not. Similarly, boosting the midrange EQ on a vocal channel can make a voice sound louder than it would by simply increasing its overall level. The loudness in that case would come at the expense of naturalness (and, arguably, pleasantness), although, conversely, it might improve intelligibility.
In contrast, boosting the lower and higher frequencies (or making comparative cuts in the upper midrange) can make a sound seem ‘bigger’ without making it ‘louder’.
High frequencies attenuate in air much faster than low frequencies, and our ears compensate for this (we ‘expect’ a distant sound to have less high-frequency content than a close sound). This is the basis of the ‘presence’ control on older mixer-amplifiers and some current instrument amps and combos: if the upper mid- and high-frequency content is boosted, the sound seems closer, as well as louder.
Reflections from room surfaces increase as the overall sound pressure level rises, and can interfere substantially with intelligibility, as well as changing the overall frequency content (walls reflect some frequencies more than others). While it may seem natural to turn it up if you can't hear the vocals clearly, it can sometimes actually improve things if you turn it down.
Our ears try to protect themselves from very loud sounds (hearing damage caused by loud noises - also known as Permanent Threshold Shift - is cumulative and incurable). There are two principal protective mechanisms: the bones in the ear shift, reducing the mechanical conduction of vibration to the inner ear; and the blood vessels contract, reducing the blood supply to the organ of Corti (the part of the ear that translates physical vibrations into nerve impulses), and protecting the sensitive hairs that deal with the smallest sounds and uppermost frequencies.
These mechanisms reduce the sensitivity of the ear, with more effect at the very highest frequencies. The immediate effect is a reduction in apparent level: what initially seems very loud quickly becomes tolerable. Once the sound has ceased, the ear will gradually return to normal (usually within 24 hours or so, although it can take as long as a week). In the meantime, there is some loss of hearing (also known as Temporary Threshold Shift). Repeated exposure to sounds that cause Temporary Threshold Shift can lead to Permanent Threshold Shift.
Effectively, your own ears will turn down the volume and mute the damaging highest frequencies if the concert is too loud. Do it long enough and/or often enough, however, and you will damage them anyway.
If you think good music needs to be deafening, it will eventually deafen you.
A good sound system should have broad frequency response, good dynamic range, and produce clean (undistorted) sound at high sound pressure levels. Our own systems are built to that standard, and may not seem to be as ‘loud’ at high volumes as inferior systems with measurably lower acoustic output. However, any sound system - from cheap home ‘hi-fi’ to the most expensive state-of-the-art concert loudspeaker system - can damage your hearing if you get your ears close enough and drive it hard enough for long enough.