Don't get mad get equal
Perfecting your sound with a little equalisation
Equalisation can be one of the most useful and creative devices in the studio, and yet can also wreak untold damage in the wrong hands. In our continuing series of studio tutorials, Bob Dormon explains the many types and uses of EQ - use this in conjunction with the accompanying aural tutorial on CD for a fuller understanding of this essential area of music production.
Equalisation is a very personal area. What is music to my ears may be a load of old socks and mackintoshes to yours. At least your home hi-fi offers some 'contour-control', but drum module and synthesiser sounds tend to come in any colour so long as it's grey.
This is where a spot of equalisation (the fancy name for tone controls) can improve the tonal colour of an instrument or sound source, to fit the picture or soundscape you are trying to create. Where you choose to make those changes is not written in stone. It's a totally individual approach to sound, very much like your own handwriting. Pens, like equalisers, are made in their thousands for everyone to use, but it's what you do with them that makes the difference.
Approach a mix of your music as an artist. Think of it as a picture you're trying to paint, with certain tones more attractive to you than others. Consider the sound spectrum of Low, Mid and High frequencies as Red, Yellow and Blue, out of which all colours are made. All you have to do is vary the amount of these fundamental ingredients to produce the tints and shades that suit. Some sounds will already be coloured to certain hues which may be just what you want.
Take a bass for example. It could sound perfect, but when a low register piano part plays along with it, then things become confused. This kind of thing frequently happens with instruments playing together in similar ranges. One likely solution is to reduce the amount of bass in the piano slightly, or the mid-range in the bass. Effectively, this is using EQ to help separate out the sonic interference by adjusting the tonal boundaries of the conflicting instruments. If you think of it in terms of colour, and your bass and low piano figures are virtually the same shade of red, then you won't be able to tell the two apart.
Musical arrangements dictate the range of frequencies that you'll hear. In general, within the passage of a piece of music, there will be a reasonable spread of low, mid, and high frequencies. With orchestral pieces, the composition may well have been written with the merging of instruments in mind, so that it is difficult to identify an individual instrument. Instead, you hear a unique blend of sounds.
Gil Evans, arranger for Miles Davis on many album projects, was particularly adept at this. While you may not have a Miles album in your collection (shame on you) you can hear Gil's work on Sting's version of 'Little Wing' (the Jimi Hendrix song)..
This blending of sounds can create the most original timbres, and EQ can assist in the process by forcing sounds to sit together so that their identity becomes blurred. On the other hand, the clarity of individual sounds is also very desirable. It has been said that you can always hear the instruments that make up a piece by Mozart. And it is often apt to apply this kind of philosophy to modern music.
Drums and percussion are regulars for an EQ paint job, as the components of a drum kit have a slot in each section of the frequency spectrum.
Working from the bottom upwards, you find the bass drum with most of its weight occupying the low frequency area. Floor toms and tom-toms move into the low mid, as does the snare drum that rattles towards the upper mid. Cymbals have quite a range to them, though fashion dictates that we emphasise their upper-mid to high frequency ranges. Hi-hats too, are fashion victims, and how they are EQ'd depends on the style of music. If you're using old loops and adding hi-hats, then a middly EQ will help them sound more authentic. Others prefer a more punishing high frequency dance approach, and if you check out the CD tutorial you can hear how both can be achieved from one hi-hat sound.
EQ can vastly affect the character of a sound. Used subtly, the presence of a sound can be changed in the mix without making very noticeable changes to its tonal quality. A typical scenario is vocals. True to form, the vocalist at the back of the control room insists you turn her voice up, 'cos she can't hear herself. This could be because there's other instruments fighting for recognition in the same frequency band as her vocal, either that or that the vocal itself is pretty dull.
In such a situation, you might find that boosting the presence on the voice at around 5kHz will help it cut through. Alternatively, see if you can sacrifice some of the boost on the instruments at around 2-4kHz. You don't have to be too drastic with your adjustments, as a little bit can go a long way. A non-EQ option is to pan the instruments either side of the centred vocal. While this might not help in mono, it can aid the definition of the sounds in the mix, as it adds polish to your production and improves your awareness of the overall soundscape.
Bear in mind that when you're using EQ, you are altering the output level of the signal you're treating. A lot of people only boost sounds because they think it sounds better. But really it's so they can hear it better, as the signal is getting louder. You must remember to make compensations in level with the faders, for the work you do on your sounds.
Making sounds too bright is a common flaw in amateur mixing, and the result is a mix with no middle to it. A good mix has a place in the frequency spectrum for all the instruments - they slot into each other like clasped hands. Spectrum analysers have been introduced as a feature on high end mixing consoles to aid the identification of suspect frequency areas. SSL were among the first to do this. A switch transforms the desk meters to represent frequency bands and their associated level changes within the stereo mix.
On a decent mixing desk you'll find more than one type of EQ or filter at your disposal. They perform differently and offer distinctive ways in which to manipulate sounds. Here are some of them:
The various nomenclature used to describe different sections of an equaliser can seem a bit daunting at first, especially if you're unfamiliar with the unit. But take a deep breath and look a little closer, as the numbers surrounding the knobs and switches will help you identify what's what. Similarly, the names of filters do describe their function to the letter. A high pass filter lets all the high frequencies pass through it unaffected.
Very often this type of filter appears as a switch that has a predetermined filter or 'rolloff' frequency usually around 50-120Hz. This means that all frequencies below say 120Hz will be attenuated at a particular rate. This rate is called the slope of the filter, and is described in terms of dB per octave. Typically, a filter will attenuate at -6, -12 or -18dB per octave.
So what's an octave got to do with all this? Well, technically speaking an octave is either double or half the fundamental frequency. Going down an octave means that you're halving the frequency. So in terms of filtering, if we start at 120Hz, then an octave below that will be 60Hz. And with a -6dB per octave filter, you'll find that a signal at 60Hz will indeed be 6dB quieter than the signal at 120Hz. The diagram below illustrates this fact, and shows the filter slope.
Those of you with keen eyes will have noticed that the 120Hz point is actually 3dB below 0dB. The reason for that is another technical detail. Frequencies that are attenuated by less than 3dB are said to be inside the passband. Frequencies that are attenuated in excess of 3dB are said to be in the stop band.
The frequency point that is attenuated by exactly 3dB is called the turnover or cut-off frequency, and it is this point that is used to name the filter.
A low pass filter employs exactly the same principles as a high pass filter except that it operates at the other end of the audio spectrum.
A low pass filter lets all the low frequencies pass through, but rolls off the high frequencies.
No doubt the most common form of equaliser is the shelving filter. This appears on radios, record players and portastudios. It's simply the bass and treble variety, and boosts or cuts the frequencies from a particular point onwards. The cut or boost level reaches a plateau or 'shelf' further into that end of the frequency spectrum, and then continues at this level. The diagram below should help clarify the shelving EQ's action.
This type of equalisation allows you to choose the frequency that you want to work on. There are at least two controls: one selects the frequency while the other alters the gain. By boosting the level with the gain control you produce a peak around the frequency area you've chosen. Doing the opposite, and attenuating the level will cause the sound to dip out around that frequency.
This type of equaliser is extremely flexible, and allows you to remove unwanted resonances. Sweeping through the frequency spectrum, you can track down the erroneous frequencies and then pull them out by reducing the gain at that point. And if you need to emphasise an instrument, you can find the frequency that gives it presence and boost it to taste.
This is an additional control that really helps focus on a particular frequency. When you boost a frequency with a parametric equaliser, you're not boosting just that frequency alone. You could be boosting a range from 500Hz to 2000Hz with 1000Hz as the peak.
That range is called the bandwidth. The Q control on an equaliser will vary the bandwidth that is being boosted or cut. Low or broad band Q values (0.3 or less) affect a wider range of frequencies than high, narrow band values (1.0 upwards). The higher the Q, the sharper the peak becomes; narrowing the bandwidth around the centre frequency.
Q helps to contain your EQ adjustments to within a certain bandwidth. Initially, you can use a narrow Q to help you to isolate an area that needs attention, and then back it off a bit to make any severe changes blend in.
A graphic equaliser is usually one of the first serious EQs that people encounter. Widely used by hi-fi buffs, PA systems and deaf people in battered Cortinas with fluffy dice, the graphic equaliser offers instant gratification. All the frequency bands are clearly labelled, with dedicated faders for each stage. The curious choice of tweakable frequencies is actually a standard set by the ISO.
The stages are one third of an octave apart, so when you reach the fourth stage it should be double the first and so on: 25, 31.5, 40, 50, 63, 80, 100, 125, 160, 200Hz and so on. Graphic equalisers should have little or no interaction between adjacent bands, although they have a low Q whose action varies with design.
The individual sliders have a zero-detent centre and typically provide ±12dB of boost/cut. Graphic equalisers are used in studios to help correct any anomalies in the controlroom's monitoring. For PA rigs, they're invaluable to overcome erroneous acoustic resonances and feedback problems. In car systems, they enable you to irritate pedestrians at traffic lights. You can tell I live in South London, can't you?
No, this isn't the sort of EQ that agrees with everything you say, or behaves like a doormat. Passive EQ uses circuitry that doesn't require power, by employing tuned circuits that work in much the same way as tone controls on a guitar or bass. Adjustments ultimately mean a loss of signal gain, and so a compensating amplifier is inserted at the end of the signal chain. This kind of EQ tends to last forever, and models by Lang and Pultec are much sought after, so much so that Langevin (a.k.a. Manley Laboratories) produce an EQ based on the old Pultec.
A lot of digital outboard gear has some form of equaliser within it, reverbs and guitar effects in particular. The types of EQ they offer are generally multi-prod versions of those above, recreated digitally with user-familiar terminology. Digital mixers are now better specified, quieter and more user friendly than before. Yamaha's ProMix 01 digital mixing console is testament to that (see The Mix issue 1, July '94 for details). This mixer allows you to store all your EQ settings, and you can control the EQ parameters in realtime via MIDI, enabling numerous effects, among them filter sweeps and wah-wah effects.
Digital signal processing does introduce fractional delays. With self-contained digital mixers, all the timing errors are compensated for internally. However, digitally connecting a stand-alone digital EQ as an insert for a digital audio workstation may cause phasing/timing problems. That said, there's nothing to stop you using one to EQ a DAT when making a master copy, as you'll only be manipulating the transferred data, not running it against other tracks.
To suggest that a particular EQ setup will guarantee earhole-friendliness would be foolhardy to say the least. Particular musical styles require distinct differences in approach. A pop piano is typically bright, whereas jazz and classical piano recordings tend towards softer, warmer tones.
Cultural considerations are also a factor. Did you ever wonder why those transistor radios imported from Hong Kong in the Seventies sounded so tinny? It's because the orientals actually liked the sound. The sensitivity of their ears is different to ours. Far Eastern cultures are used to listening to speech at a slightly higher frequency than Westerners. Fletcher-Munsen's 'equal loudness contours' have suggested as much. Next time you have the dubious pleasure of hearing some Japanese glam-rock, listen out for the bass... if you can find it!
Personal tastes aside, here are some generalisations you can apply to certain instruments. Microphone techniques deserve a mention too, as they can alter the frequency response significantly. Careful positioning allows you to perform rudimentary equalisation before you've even started fondling your filters.
EQ can succeed with explosive vocal sounds (such as 'p's and 'b's) where mic technique and pop shields have failed. Rolling off frequencies around 200Hz helps, but if you're worried about how it'll affect the warmth of the vocal, then try using a parametric EQ and tune in to the most pronounced 'popping' frequency before pulling it out. Thinning out the bottom end a bit works well on backing vocals, where you're likely to get compounded popping. You can even apply EQ at source, as quality microphones often have a bass roll-off filter typically operating around 70Hz. Some, like the AKG 414 give you a choice of attenuating 75Hz and 150Hz.
Usually, the mid-range takes care of itself on vocals. However, adding some boost at around 5kHz gives a pleasant airy vocal sound and aids intelligibility. When mixing lead vocals against backing vocals, applying identical treatments can exaggerate these frequencies to uncomfortable levels. Trying slightly different frequency areas to work on, will thicken the sound and can enable the two parts to work together, rather than fight each other.
How you EQ these instruments will be dictated by the range they're played in. Obviously, a baritone sax will require different treatments to an alto sax, but what they share is a quality and range similar to that of vocals. However, the reeds used can change the timbre of an instrument significantly, so it's always worth letting the player try alternative reeds and mouthpieces rather than having to fix it in the mix.
Saxes can be quite intense in the mid range, so a little cut around 1-2kHz can offer some relief. Presence is increased from 3-5kHz, but be careful not to allow it to interfere with any vocals. The resonances flutes produce are sometimes more easily controlled by using compression, but if it's a breathy chiffer you're after, then you can roll off the bass from 500Hz and boost the top around 7kHz. There's usually enough power in a flute sound to retain a hint of the lower frequencies, but the EQ helps to lighten the sound.
If it's a solo sound you're working on, then you can fatten it up around 250Hz or improve the scratchiness with a boost around 7-10kHz. Pads need to be watched, as they can cover such a broad range that you might need to contain their sound within a certain bandwidth, to prevent unpleasant resonant beatings. If you're one of those rare individuals that records real instruments, avoid close-miking violins and the like. Around four feet above the instrument allows it to breathe and brings out the bass.
Once again you can increase the presence from 5kHz upwards. If this is a bit too piercing, then 3.5kHz cuts through quite well. For mellow timbres you might want to pay more attention to the bass and add fullness to the sound from 100-300Hz. Pulling out these lower frequencies can transform the players from the Grimethorpe Colliery Band to Tower Of Power in one fell swoop..!
The type of guitar you use and the pick-up it's set to, can do a lot of the work for you. If you're after Nile Rodgers' Chic chordal chop, use a Stratocaster on the back or mid pick-up, wind off the bass from 300Hz and boost the middle and top at around 3.5kHz and 5-6kHz respectively. You might be surprised by just how severe you'll need to be, especially if you're going straight into the mixer.
More mellow bluesy sounds are easily accomplished with front and middle pick-ups and Les Paul guitars. The humbucking (double coil) pick-ups are warmer sounding than the single coil pick-ups used on Strats and Telecasters. You can afford to leave the bass in here, although you might want to boost around 3.5kHz if you need extra clarity.
If it's a picking funk tone you need, then the low mid can help. 1kHz hardens up the picking while 3-5kHz will bring out some of the strings' 'bite'.
Heavy metal merchants will have to rely on their amps to get the initial overdriven sound they're after. Sometimes tranny amps and effects boxes make things painfully bright, so if you prefer a warmer sound you'll have to sacrifice some of the top by winding off a touch from 7kHz upwards. 2kHz can strengthen the sound somewhat, while a dab of boost at 350Hz helps to fatten it up. Overdriven guitars are prone to mains hum, so you might want to consider winding off the bass at 100-50Hz while keeping the 350Hz whammy warmer.
Careful miking, a decent guitar and a new set of strings will save a lot of audio aggro. Moving the microphone off from the soundhole and towards the bridge brightens the sound significantly, and prevents the low frequencies from becoming too predominant. There is a temptation to thin out acoustics too much to compensate for their pronounced bass response. A slight dip at 200-500Hz can work well but more often than not, it's the low mid-range that requires attention. A boost around 1kHz can produce a more orderly tone, while clarity is improved at points around 2.5, 3.5 and 5kHz. The sound thins with the rising frequency, hence string squeaks and breathing muso's are emphasised.
Bass can be problematic to EQ. Basslines kept within a particular register are easier than those leaping several octaves. The reason being that with the former, you can concentrate your efforts over a specific range and get the tone you want. If you're dealing with a wide range of frequencies, you can often find what works well on one area is to the detriment of another.
Let's say you find the bottom string honks a bit and you'd like a warmer tone. Pulling out 700Hz to 1kHz may be just the job... until the higher notes are played. These can then sound dull or just quieter overall because of your tweaking.
With bass, it's often prudent to avoid doing anything too drastic, especially when recording. Sure, you can attempt to introduce sub-bass tones with some a boost at 25-50Hz but check it out on some small speakers too. Even if they're capable of reproducing such tones, they might not be able to handle extreme level increases in this area, and thus the millions eager to buy your work won't be able to enjoy it at its best on their ghetto blasters either.
In general, you can add fullness to a bass between 50-200Hz, take out the woolliness at 500Hz, add sustain at 1kHz or alternatively remove it if it honks too much, and help improve clarity by introducing attack/string noise at around 2.5kHz.
As I said earlier, piano sounds are very particular to musical styles. Sadly, it's probably true to say that the majority of home recording enthusiasts haven't heard a real live unamplified piano since they were at school. Memories fade, and modern synthesizers present an emulation of the piano that flatters the high frequencies and attack timbres, but doesn't really honour the warmth of the original. To some of you, a real piano might sound like crap compared to a digital doppelganger.
Nevertheless, this is progress at a price, and if you want to get more tinkle out your ivories then 2-4kHz is a very effective area to try boosting. If your piano is too muddy or not warm enough, then concentrating around 150-300Hz. It's likely that you won't need to add any more top to the sound, unless you're musing on a real Mc Coy, Steinway or Blüthner...
Mark Parson's excellent feature Rhythm Methods in the August issue of The Mix covers drum recording explicitly. But here's a few points to bear in mind, and you can check out the CD for an audition of these techniques.
All right, so you've managed to hold back on the bottom end for all the above instruments. Now's your chance to get those trousers flapping with some sub-sonic sound shaping. Boosting from 70Hz downwards on a bass drum is bound to make an impression, but try to exercise restraint especially if your monitoring is bass-light, as you may overcompensate to produce a crap mix. Geddit?
You can remove some flab from a bass drum to tighten it up to jazz-rock specifications by pulling out frequencies around 500Hz. You can also emphasise or dull the attack at around 7kHz.
A snare drum can be fattened up from 150-300Hz. Go above that and you can isolate the resonant frequency of the drum. This can be useful for producing distinctive sounds like the snare on 'She Drives Me Crazy'. You can also crispen up the sound with a boost around 5kHz. Toms also benefit from a touch of top, as this helps pronounce their attack, and a low end boost bolsters their natural resonance which varies with tuning.
Cymbals and hi-hats can be more outstanding in a mix if you wind off some of their low frequencies from 300Hz. If you're recording cymbals with overhead mics on a kit, then the right choice of microphones can do some of the work for you using bass rolloff filters (see 'Vocals', above). Using the mid-range between 2-5kHz adds more definition to the attack/stick noise (useful on ride cymbals), while 7kHz upwards sizzles and shimmers. It's important to vary the EQs applied to cymbals: ride, crash, splash and so on. You can position them in the high frequency spectrum so that they don't conflict with tambourines and shakers. Panning is also a very important consideration for enhanced definition and earhole teasing.
Bob Dormon would like to acknowledge Alan Smith of the Nomis Studio Complex for his untiring assistance in the preparation this month's CD tutorial.
On The Re:Mix CD:
12 EQ tutorial
This disk has been archived in full and disk images and further downloads are available at Archive.org - Re:Mix #5.
Feature by Bob Dormon
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!