ALL ABOUT EQUALISATION

PAUL WHITE looks at the various types of equaliser, and gives some
tips on how EQ should best be applied to your recordings.

Equalisation can come in many forms, from simple treble/bass controls to

multi-band, parametric equalisers, but, however you dress it up, equaliser
is just another word for 'tone control'. The term equalisation came about
because the very first equalisers were developed to help counteract or
'equalise' shortcomings in telephone systems, but today equalisation is
used creatively as well as to fix problems. Equaliser circuits are based on
electronic filters -- hence the term filter, which crops up quite a lot when
you're talking about EQ. Strictly speaking, a filter is a device that removes
something, but in the context of active equaliser circuits filters can boost
frequencies as well as cut them.

Check any textbook on audio, and you'll see the limits of human hearing
quoted as around 50Hz to 20kHz, though those very same books will also
point out that very few individuals, other than young children, can hear
pitches anything like as high as 20kHz. A more realistic figure might be
around 15kHz for an adult, decreasing further as the years pass. What is
really puzzling, though -- and this is scope enough for a feature in its own
right -- is that even if your measured hearing response starts to fall off
well below 20kHz, it is still possible to hear the effect of equalisation
applied at the top end of the spectrum, where you wouldn't expect to be
capable of hearing any change. The plot thickens further when reputable
studio engineers claim to be able to differentiate between two otherwise
identical circuits if one has been modified to handle frequencies up to
50kHz and one handles frequencies only up to 30kHz. In theory, both
limits are well above the limit of human perception, so it seems that what
goes on outside the audible spectrum has a way of influencing what we
perceive within the range of our own hearing systems.

SHELVING EQUALISERS

Though equalisers all do essentially the same job, there's a great deal of
difference between a simple 2-band treble/bass tone control and a multi-
band studio equaliser. The simplest is the shelving equaliser -- a device
that applies cut or boost, rather like a volume control, but only to the
frequencies above or below the cutoff point of the equaliser, depending
on whether the equaliser is based around a high-pass or a low-pass filter.

A low-pass shelving filter, as its name suggests, passes all frequencies

below its cutoff frequency, but attenuates all frequencies above its cutoff
frequency. Similarly, a high-pass filter passes all frequencies above its
cutoff frequency, but affects all frequencies below its cutoff frequency.
Figure 1 shows the frequency response graphs of a typical treble/bass
EQ using high- and low-pass filters. Note that the filter graph shows a
slope at the cutoff point -- it isn't possible, or desirable, to have a filter that
does nothing at one frequency, and then comes in with full effect when
you move up by just 1Hz. Simple shelving filters typically have a 6dB-per-
octave slope, so that their influence is felt more progressively --though it
is possible to make much steeper slopes if required. The gentler the slope
of the filter, the more frequencies outside the range of the filter will be
affected.

BAND-PASS FILTERS

A filter that passes frequencies between two limits is known as a band-

pass filter, and on a mixer with a mid-range control, the Mid knob controls
a band-pass filter. On a typical mixer, the band-pass filter will have
variable cut and boost, and on more flexible mixers it will also be
tuneable, so that its centre frequency can be varied. This is known as a
sweep equaliser, because although the filter frequency can be changed,
the width of the filter cannot be adjusted. Figure 2 shows a typical band-
pass filter response, including a sweep control function. In a typical mixer,
the high and low shelving equalisers are used to control the high and low
end, while a band-pass filter controls the mid range. However, some
mixers use band-pass filters for EQing the low end too. The argument for
doing this is that a shelving high-pass filter will also boost all frequencies
below the ones you want to work on, whereas a band-pass filter
attenuates frequencies both below and above its cutoff points.

PARAMETRIC EQ

A parametric EQ is very similar to a sweep band-pass EQ, except that a

third control is added to allow the width of the filter response to be
adjusted. The width of a filter response is sometimes described as its 'Q'
value, where Q is the filter frequency divided by the number of Hz the
filter affects -- in other words, its bandwidth. The formula for Q is:
Q=centre frequency in Hz

bandwidth

Because the filter response is curved, the actual frequency width is

measured between the points on the graph where the signal level has
fallen by 3dB. A high value for Q corresponds to a very narrow filter; a low
value of Q corresponds to a wide filter. High Q values are useful for
picking out sounds that occupy a very narrow part of the audio spectrum,
whereas lower Qs produce a smoother, more musical sound.

A studio parametric EQ may have several filter sections, enabling three or

four parts of the frequency spectrum to be treated simultaneously.
Parametric EQs can be time-consuming to set up properly, but they are
the most powerful and flexible of the conventional EQ types. Figure 3
(overleaf) shows a typical parametric equaliser response.

GRAPHIC EQUALISER

A graphic equaliser can be recognised by the row of faders across the

front panel, each fader controlling its own narrow section of the audio
spectrum. For example, a 30-band graphic equaliser provides
independent control over 30 different bands spaced one third of an octave
apart.

Other than the highest and lowest faders, which control shelving filters,
each of the filters in a graphic equaliser is a fixed-frequency band-pass
filter, where boost is applied by moving the fader up from its centre
position, and where cut is achieved by moving the fader down. Graphic
equalisers have the advantage of being very easy to set up, but must be
used sparingly, as if they're not very well designed, they can have an
adverse effect on the sound. They are also less flexible than the
parametric EQ, which can be tuned exactly to specific frequencies. On
the graphic equaliser, the range covered by each fader is fixed, and the
width of each individual band of a third-octave equaliser is actually rather
wider than a third of an octave, to allow a smooth overlap between bands.
Figure 4 shows the response of a typical graphic equaliser.

EQ TODAY

While early studio equipment and, to a greater extent, telephone lines

needed a lot of corrective EQ to make them sound natural, modern
recording equipment is capable of storing and reproducing sound that is
virtually identical to the original. Nevertheless, the original sound isn't
always what we want to hear, so EQ has evolved to take more of a
creative role. What's more, the simple bass and treble 'tone' controls
found on early equipment have been replaced by sophisticated multi-
band equalisers which place far more precise control in the hands of the
user.

At heart, no matter how complicated an EQ is, it is still really just a

frequency-selective volume control, but its subjective effect on the sound
is often more profound than this description might leave you to expect. In
February 1996's SOS I wrote an article that explored some of the psycho-
acoustic aspects of EQ, and concluded that the most logical place to start
searching for reasons why EQ has the effect on us that it does was in
nature -- specifically, the Earth's atmosphere. Here, low frequencies travel
slightly faster than high frequencies, so the further away a sound source
is from the listener, the more delay there'll be between the low
frequencies (which arrive first), and the upper harmonics, which follow.
The higher up the spectrum you move, the greater the delay will be, and,
in effect, the harmonic structure of the sound becomes progressively
more smeared as it travels. Air also absorbs high frequencies more
readily than low ones, so the further away a sound is, the less bright it is.

Obviously, we're used to sound behaving like this, so we accept the effect
as natural, which, by definition, it is. Our brains don't recognise the phase
distortion -- they just recognise the sound as being distant. By the same
token, louder sounds that have little or no phase smearing are perceived
as less distant. In recorded sound, however, we're faced with the problem
that all recorded sounds travel the same distance -- if the hi-fi speakers
are three metres from the listener, then that's how far the sound travels.
To fake the illusion of distance in a mix, therefore, we need to recreate
the smeared harmonics and reduced HF response that occur in nature.

LEVEL & PHASE

We've already established that EQ acts as a selective volume control,

targeting only certain parts of the audio spectrum. However, it's well
known to circuit designers that EQ doesn't only change the level of
specific parts of the spectrum, it also changes the phase of the affected
frequencies relative to those that aren't being cut or boosted. In other
words, the equaliser is creating a similar effect to changing the distance
of the sound source -- it's affecting both the frequency response and the
phase relationships of the signal. As I suggested in last year's psycho-
acoustics article, this may be one of the reasons why brightening up a
sound makes it seem closer, and why taking off some high end might
make a sound seem more distant. Of course, every design of EQ affects
the audio spectrum and phase response in a different way, and -- leaving
aside technical criteria such as noise and distortion -- this might explain
why some EQs have a more natural, musical sound than others.

IMPRESSIONS OF LOUDNESS

It's a well-documented fact that the human hearing curve isn't flat, but
instead is more sensitive to mid-range sounds than to frequencies at the
extreme high and low ends of the spectrum. Of course, we don't notice
this, because we've heard sounds this way all our lives. However, as the
level of sound we're listening to increases, the mid boost of the hearing
system becomes less, and the result is that high- and low-frequency
sounds seem proportionally louder. This is yet another of those
interesting physiological facts that can be exploited to fool the ear into
believing something that isn't entirely true. For example, if we know that
extreme high and low frequencies stand out more when we listen to loud
music, we can create the impression of loudness at lower listening levels
by attenuating the mid range and boosting the HF and LF ends of the
spectrum. The loudness button on a stereo system does exactly this, and
if you look at the graphic EQs used in a club or PA system, you'll often
see them set up with a smile-shaped EQ curve to promote the illusion of
loudness and power. Of course, this works just as well in the studio,
although it's always most effective just to treat some of the sounds in a
mix, to maintain a contrast between the different sounds.

OVERVIEW

As a general rule, equalisation should be employed only after all efforts

have been made to obtain the best sound at source. What's more, there's
a huge subjective difference in sound between a budget equaliser and a
top-quality studio equaliser: if you have to work with a budget EQ, or the
EQ section built into your desk, you'll probably have to use it sparingly if
the overall sound isn't to suffer. Though the character of a really nice
equaliser is difficult to quantify, the best equalisers allow you to make
more drastic changes without the sound appearing unnatural.

Most often, a combination of cut and boost is required -- but always use
the Bypass switch to flip back and forth between the equalised and
unequalised sounds, to make sure you really have improved matters.
Equally, if you feel the need to EQ an instrument in isolation, check again
with all tracks playing to make sure that the settings you're using work in
context with the rest of the mix. More often than not, you'll have to make
further adjustments, but it really is worth striving to get your sounds right
at the outset. EQ is an invaluable ally in shaping well-recorded sounds,
but even the best equalisers have their limits when faced with difficult
material.

TO CUT OR BOOST?

In general, the less EQ boost you use, the more natural the final sound
will be. The human ear is far more tolerant of EQ cut than it is of boost,
so, rather than adding lots of top to vulnerable sounds such as vocals in
order to get them to sit at the front of the mix, try applying high-end cut to
other sounds in the mix that are conflicting with the vocal.

Some classical purists might say that you don't need EQ at all, but in the
world of pop recording, where the emphasis is on appropriate rather than
accurate sounds, equalisation has become a way of life. The close miking
of drums was originally devised to cut down on spill from other
instruments, but now it's become the normal pop drum sound. EQ plays a
very large part in creating a modern drum sound, but because no-one is
trying to emulate a natural acoustic drum sound, the EQ is used in a
creative context rather than a corrective one.

SEPARATING SOUNDS

EQ can be used creatively in many ways, but one of the most popular
applications is to separate two similar sounds within a mix where the
degree of overlap is causing the sound to become confused or muddled.
For example, if two sounds are fighting for the same part of the spectrum,
a peaking equaliser can be used to add a degree of bite to one sound at
one frequency, while the other sound can be peaked up at a different
frequency. Similarly, the top or bottom end of a sound can be 'trimmed' to
avoid conflict, a typical example being the acoustic rhythm guitar in a pop
mix.