Andy Honeybone gives you a boost on all frequencies as he puts the clarity back into tone controls — or eq, if you prefer

The very thought of tone controls is enough to give hard-core audiophiles apoplexy. Like the early navigators, these people believe that flat is beautiful and any interference is crass distortion.

Unfortunately, at low levels, the ear does not hear 'flat' and it is not until a sound pressure level of some 90dB is reached (a loud orchestra) that the whole audio range is perceived at one level. When records are played at less than concert volume, it is necessary to boost the bass and treble to recreate the tone balance of the original work. This correction is the work of the loudness button to be found on affordable hi-fi. It's not worth arguing the fors and againsts, but as musicians, we like to call twiddling with tone controls 'creative equalisation' and, as artists, that is our prerogative.

The historically useless knob marked 'tone' found on guitars of all creeds is the starting point of this month's electronic 'That's Life'. The control may also be correctly described as 'continuously variable'. This grand phrase means that it ain't a switch — often, passive treble cut controls might as well be because they have no effect for 95% of their travel and then, suddenly, there is no treble! This isn't much of a problem as guitarists always leave the control turned off and roots-rocker bass players leave the control full on. From an economic point of view, the highly styled knob is far more expensive than the variable resistor and capacitor below the scratch plate.

The components form a low pass filter whose ferocity is adjusted by setting the variable resistor. The circuit dates from valve amp days when the high impedance characteristic of 'bottles' allowed a more acceptable performance. The capacitor is the component which is frequency selective and, for a given value, passes high frequencies and attenuates low frequencies.

If one side of the capacitor (it is a two-legged beast) is connected to signal earth, then it forms a 'bucket' for the disposal of high frequencies. If the capacitance is too large, then the entire signal will be 'shorted' to earth. The resistance, or more correctly, reactance of a capacitor at a known frequency may be easily calculated (if you can remember pi), and the property of frequency dependant 'resistance' is the basis of all analogue filtering.

Much more effective are the familiar hi-fi treble and bass controls which have the advantage of allowing both boost and cut. The circuit may be constructed using passive components (no amplifying stage) at the price of losing a large amount of signal when the controls are set flat. This loss is the amount of available boost. It appears that these controls are mainly used for boosting so the loss of signal is not such a problem after all.

The classic design is hard to improve on and uses a minimum of components in a frequency selective, potential divider configuration. The bass control is separate from the treble and behaves in an entirely different manner. Rotation of the bass control actually changes the frequency at which boosting or cutting occurs. The treble control, however, acts as a gain control for high frequencies. The drawback with the treble section is that in order to boost a signal at 4kHz, it is necessary to add four times that boost at a frequency 2 octaves higher. There is no choice because the control has its maximum effect at the extreme of the audio range.

The next stage of sophistication is the active tone control which entails transistors, power supplies and inevitable hiss. Inclusion of tone forming networks in the feedback loops of amplifying stages gives symmetrical boost and cut (mirror image response curves), no signal loss, and allows the use of smaller value components. Feedback in this sense is not the howl of a microphone too near a P.A. cabinet, but refers to the cure-all effect of routing some of the output of an amplifying stage back to its own input. Feedback may be positive or negative — positive leads to oscillation whereas negative has a stabilising influence and trades gain for lower distortion.

Further, if treble boost is applied in the feedback path, the overall amplifier will end up with a treble cut response. In other words, tinkering with the feedback gives the opposite effect to the output signal. By engineering the frequency dependent networks to swing between the input and feedback paths of an amplifying stage it is possible to arrange both boost and cut at the turn of a variable resistor. The design of the active treble and bass control is credited to Baxandall who published his 1952 design in Wireless World. When the amplifying stage is just one transistor, a tone control can have a distortion of 1% and many involved implementations have been contrived to reduce this figure.

The maximum cut and boost available with this type of circuit is ±20dB or, in English, ten times greater or ten times less at the extremes of the audio spectrum compared to the level given when the controls are centred. Many designs are limited to ±12dB (±4 times) because the hiss of boosted noise would become unbearable if further amplification was available. Beware specifications which say 'tonal variation 12dB' because this is a naughty way of hiding ±6dB boost and cut.

Although the Baxandall design is still going strong after 30 years, there is the need to be able to cut and boost selected parts of the audio spectrum without affecting frequencies outside that band. To achieve this, a set of band pass filters are used with their centre frequencies spaced at octaves or less. These filters may form part of an input or feedback loop depending on the position of a variable resistor. When the band pass filter is in the feedback loop, the response is band reject (notch), and when the filter is in the input line the response is a selective frequency boost. When the variable resistors are slider types, the unit is termed graphic equaliser because the position of the knobs resembles a response against frequency graph.

Inductors are reactive components which are opposites of capacitors, impeding high frequencies and passing low frequencies. Combinations of capacitors and inductors allow resonant networks to be built which block all but a narrow band of frequencies. Band pass filters were traditionally built this way, but due to the cost and bulk of inductors they are now synthesised by a circuit known as a gyrator, or done away with completely by using multiple feedback active filters.

To cover the audio range adequately, ⅓rd-octave spacing is required leading to an unmanageable number of controls. An alternative is the parametric equaliser which is like a single section of a graphic but can be tuned to any frequency. A further refinement is the provision of a Q control (resonance) which varies the width of the band to be affected. A form of this type of equaliser is now available on some multitrackers as it allows a signal to be brightened without highlighting tape noise at higher frequencies.

A parametric equaliser can sound like a horrendous wah-wah or a cheap static phaser if the Q and cut/boost are set too high. On the other hand, parametrics are capable of some very subtle effects but the controls are tricky to master and it is easy to find yourself sweeping through the audio spectrum waiting for something magic to happen.

One of the main uses of an equaliser is to equalise the P.A./venue combination so that the microphones can be wound right up. The official method is to blast pink (bass boosted) noise through the P.A. and then monitor the sound with a microphone hooked up to a spectrum analyser. The analyser splits the noise into bands and gives a sound level for each band of the audio spectrum. Pink noise is an equal mix of all frequencies hence all the levels should read the same for a flat response. If any peaks and troughs appear, the equaliser is tweaked to remove them. Personally, I think the turn-it-up-until-it-whistles technique still has a lot going for it.