People often over-use EQ at home, going through their tracks and making each individual track sound as good as they can. It’s a common mistake, but it’s easy to avoid.
The quick version is…
Use EQ sparingly and only when you think you really need to. Never make EQ decisions when the solo button is on. Always decide about EQ in context with the other sounds in your mix.
Match how loud the sound is before and after EQ so you can A/B compare them at equal loudness. You will always like the loudest one.
Cut instead of boosting as a general rule of thumb.
If you have a bandwidth control, called ‘Q’, aim to use gentler Q settings (lower numbers) for the most natural results.
If you’re fixing a problem, you might use the bigger Q numbers (anything greater than 1) for progressively pointier sounds that stick out and become angular in sonic terms.
THE LONG VERSION IS…
…coming up below, and I hope it’s rewarding to read, and there’s some more great tips, notions and facts.
Read on to see when and why you should be using EQ.
GUS DUDGEON SUMMED IT UP
Some words of wisdom from the late, great Gus Dudgeon spring to my mind. Some of you may not know who he is.
He was the engineer on the Zombie’s “She’s Not There” in 1964 and also John Mayall’s Bluesbreakers album that famously featured Eric Clapton (not only playing but also reading the English comic The Beano in the cover photograph).
Gus became a producer in 1969, when he produced David Bowie’s classic “Space Oddity“.
Sadly, Gus Dudgeon died in 2002. I must say that I would have been interesting to have heard his production chops applied to Commander Chris Hadfield’s International Space Station version last month.
Musically, Gus Dudgeon is remembered mostly for his work as Elton John’s producer through all those glorious classics of the 1970’s, starting with the debut single “Your Song”, released in 1970.
He reunited with Elton in later years to produce Songs From The West Coast and Peachtree Road, two of my favourite Elton records. He knew a thing or two about getting sounds.
He once observed that whenever he had trouble getting a mix, it was because he had used too much EQ on too many things.
This was insightful for it’s time, and exercising due caution with EQ is the norm today.
SO WHAT IS EQUALIZATION?
An equalizer (or EQ, for short) is an amplitude controller for frequency. You can turn things up or down in specific frequency areas with them. You can entirely discard frequency “bands”, discrete ranges of frequency between two given points in the frequency spectrum. Alternatively, you can discard or just turn down the extremities of audio, the very lows and/or the very highs.
The spectrum of sound is greater than what we humans can hear, but the frequencies we do hear run from about 20 Hertz to a little under 20 kiloHertz (20 kHz), roughly speaking.
In standard scientific notation, that’s a range of 20 Hz to 20kHz. As we age, we lose some of the high end over time, although this doesn’t really start to happen until we are over the ripe old age of 30 – unless we spent too much time with our head beside a massive bass speaker at a rave, or hours beside a Keith Moon style drummer. Protect your hearing – you’ll be needing it!
Easily remembered as a rule of thumb – a teenager’s fully working hearing is 20-20.
When used as a verb, “to equalize”, or, more commonly, “to EQ”, simply means to adjust the amplitude of one or more frequencies, changing the tone of the audio overall. Amplitude change means getting quieter or louder. This is measured in decibels, or dB.
Decibels are a logarithmic scale, so they are not linear measures. The actual relationship is that any increase of 6dB, no matter what the original signal level was when stated in decibels, is a doubling of how loud the sound is.
Adding 6dB = twice as loud. Removing 6B = half as loud. Precisely, scientifically, accurately, measurably.
This amplitude change might include the wholesale removal of specific frequencies, for instance the ever-popular “telephone voice”, where certain frequencies are set to zero by filtering them out completely – they are no longer present.
In simple terms, you make things darker or brighter. The interesting thing is that these terms are relative. You can only be darker than something else. And vice versa. It takes two to tango. This contrast is critical to how we humans hear things.
This is not an exact science in aesthetic terms, because any change affects how you hear the rest of the audio that you did not process. Turn the bass down on something and it sounds brighter – but you didn’t turn up any high frequencies. This is a mere glimpse of what I mean.
It’s a lifetime’s work for all types of audio engineers to study the interactions that go on during equalization and the techniques and types of approach that do the most good. Keep listening critically and thinking about what you hear.
It’s fun to listen to music. I love doing that, don’t you?
So, equalization is a tool for adjusting the amplitude of frequencies, in clumps of smaller or larger size around a target frequency. It can also just deal with the low end or high end alone, removing the extreme ends of the audible range to tidy up for us.
There are a number of types of EQ, but at home it’s not that important which you are using. It’s more about using filters to clean up your tracks, both low and high pass types, but mostly highpass, and occasionally calling in the big guns.
The big guns are able to EQ anywhere you want in large or small amount, affecting targeted areas. They can be analog hardware EQs, with tubes or solid state transistors, or they can be digital devices, most commonly as plug-ins in your DAW.
Bandwidth is the targeted area you are interested in EQ-ing when you EQ with big guns, which for our purposes are basically any EQs that are more than just low pass or high pass filters. How wide is the band is around the centre of your target?
BANDWIDTH or Q
Let’s think of bandwidth (Q) this way. You’ve got your knob or onscreen control marked Q, hopefully with little number values around it from low, like 0.2, and numbers like 1 or 5 or 7.
1 is a natural sounding Q setting, so that’s kind of a default.
You’ve got your Big Q values and your small Q values.
Technically, higher Q values (Q= >1) are increasingly narrow, selected areas that are highly affected relative to near neighbours, and lower Q values (Q=<1) are much wider areas, more gently “ramping” into the effect.
What does that mean in English?
Big Q is pointy-sounding and perhaps edgy or boomy or cuttingly bright, but small Q sounds gentle and rounded and the EQ is more flattering to the overall track by being gentler and more natural.
Each has it’s use in making a good mix, and you need both to bake a tasty cake, if I may “mix” my metaphors…
BIG Q (narrow bandwidth) AND SMALL Q (wide bandwidth)
Values I might use for Q most of the time range from, say, o.25 all the way to 7 or more. I tend to like to be as low a number as possible unless there’s a really good reason. Things just sound better that way in the long run and less harm is done in any one area.
Big Q = values greater than 1
Small Q = values less than or equal to 1
Big Q is handy for corrective EQ.
Big Q changes a lot less of the neighbouring frequencies because it’s focused and narrow and targeted.
Small Q is handy for invisibly smooth results.
Small Q changes a much larger number of neighbouring frequencies but does so more evenly and gently.
It’s much harder to go wrong if you use Q settings as near to 1 (or less) as possible. Listen and judge for yourself. As you turn down the Q setting towards zero you are increasing the number of affected frequency bands but each of them is less affected than it’s neighbour outwards away from the targeted, so-called “corner” frequency.
FOUR SIMPLE EQ TIPS
TIP #1 LISTEN CRITICALLY
The single most useful thing you can possibly do is to train your ears by listening critically as you experiment, changing settings with an equalizer and noting thoughtfully what has happened in your estimation. Go through your music collection, and build a reference library of music recorded really well. Doesn’t matter if you like the music. The point is to compare commercial quality mixes, then be enthused by the long process of gradually realizing all the areas you can tweak to get on similar ground to the fancy record labels.
It takes time to master, but there’s nothing better than using a reference library of the funky, luxurious bass sound you like on that Gwen Stefani track and the silvery dobro of a Dixie Chicks classic.
See what you can do to mimic the qualities, to match up how loud the bass end of the mix is relative to the high end, and so on. It gets easier, and it’s an incremental journey.
Listen carefully and learn from the masters! There’s always things to notice and enjoy and inspire for your next recording. Try not to get sucked into related enjoyment too often if you are listening for a purpose. Stick with that bass line through the whole song. Next pass, hear the hi-hat parts all the way through. Just the solo, whatever.
Professional engineers all have reference collections of music they know intimately from a critical listening perspective.
The reverb tails on a Peter Gabriel record, the handclaps on an underground dance classic, whatever it may be. It’s about sonic details, and becoming more in tune with picking them out from what you are hearing overall. Just hearing the floor tom and nothing else in the mix, without soloing it. It’s simply a learned skill, though some have more flair than others, like anything in life.
TIP # 2 EQ ME FOR A REASON
The easiest thing you can do is step away from the EQ. Less is more, and the sound will be richer and fuller at the end of the project if you don’t keep EQ-ing everything one track at a time.
TIP # 3 HIGH PASS FILTER ON EVERYTHING
The third tip is one which you will use routinely, and to great effect. Use a high-pass filter on pretty much everything, but DO check it’s settings religiously every time you use it. In other words, tune it. Tune it by setting the frequency on the HP filter to the highest frequency you can get away with before you can hear it is doing something to the sound. Now back off a bit, by lowering the .
It removes or attenuates frequencies and you can’t get them back once they are removed.
Any EQ plug-in that has a high-pass filter is better than none at all, but the higher the quality of the filter, the better the result, so try to use decent plug-ins for this. I know you need a lot for most of the tracks in your mix, but it will make a very clean sound when summed together to stereo.
Think of it as cleaning up the dust, litter and general rumbly nonsense from the ground floor of your audio soundscape.
TIP # 4 CUT ME, NO REALLY, CUT ME
Try to cut frequencies, rather than boost. It’s usually the best thing to do, and easily prevents unintended distortion and clipping. If it’s bass-heavy, cut bass. If it’s not bright enough, cut bass. If it’s too bright, cut treble. Got the idea?
Find a way to do it with a cut – if you need to boost somewhere, try cutting in an opposite part of the spectrum.
USING HIGH PASS FILTERS
The most useful EQ you have is the high-pass filter. This baby goes on more tracks in a mix than you would think.
Often it’s the only thing inserted on a track. It exists to throw away all frequencies below a set frequency, and allow all frequencies above that frequency to pass unchanged. This is beyond helpful, it is a godsend to a professional result.
Lows are removed, or turned down at least, and all highs pass. it’s a high pass filter.
Naturally, it has a counterpart at the high end of the spectrum called a low-pass filter, where highs are removed or turned down and lows pass unchanged.
You can clean up your mixes no end with an HPF and you set it with, basically, three parameters called Input Level, Slope and Frequency. There will usually be an output level and perhaps a ‘Q’ control on the EQ. Filter slopes typically have a fixed Q, or bandwidth, so you don’t need the Q control for high and low pass filters, since the bandwidth is not an adjustable quantity here. You just change the so-called “corner” frequency.
There’s the input level to the plug-in, where you can check it does not distort at the input, and if it does, you can lower this control as necessary. There will also be a Bypass button so you can make sure the plug-in is not bypassed.
If the EQ plug-in is not just a high-pass filter, but does other EQ jobs, you will need to set it to operate as an HP filter. There will be a button or symbol that you can click on to make it behave as an HPF (high-pass filter).
There’s the frequency below which you would like to throw away audio. Above this setting, you’ll keep all audio.
It’s not quite that cut and dried a cut-off point in reality, but you can make it a smooth, slow transition around neighbouring frequencies or a sudden sharp cut-off using the third control, called filter slope or just slope.
With slope control, you can choose how many dB of loss of volume will occur to the chosen frequencies.
A slope of -6dB per octave (half as loud per octave below the chosen frequency) is applied to the unwanted side of the selected frequency and you have “rolled off” the lows, by attenuating them (turning them down). This is a gentle slope, so there are minimal artifacts introduced into the audio (pretty much an unavoidable side-effect of using EQ).
Harsher slopes like -24dB per octave can make audible issues arise in the sound, so keep your ears open and listen critically to what changes when you use a filter. Sticking to -6 or -12 dB per octave is smart in most cases.
A -6dB per octave slope at 100 Hz throws away everything below 100 Hz by turning it all down relative to the untreated signal. That would be a gradual change, not very sudden, where each octave down you go from the frequency, another 6dB of level is lost.
A -6dB filter slope is sometimes called a single-pole slope. A double-pole slope would be 12dB per octave. A three-pole is -18dB per octave four-pole slope is -24dB per octave, which is the highest slope you’ll see on a high-pass filter.
This -24dB slope is rather dramatic about cutting off the frequencies below the chosen point (called the “corner” frequency in the trade), so it really cuts down those lower frequencies. On the other hand, the more dramatically you do this task, the more harm you can do – and might not immediately notice.
Caution pays, because you can’t bring back stuff you’ve removed once it’s been processed. Let’s be careful out there!
Vocals work well with anything below 80Hz thrown away (you don’t need footfalls and traffic rumble, and voices really don’t contain much energy below 100Hz).
For bass sounds like kick and bass guitar or synth, anything below 35 or maybe 40Hz is fair game to lose. In fact, losing some very low mud from bass sounds will make the remaining bass sound clearer, and, counterintuitively, it will seem louder as well. Piano has a left-hand part, a lot of the time, so don’t throw too much out there below 80Hz, unless it’s a “right-hand only” part in the middle to higher registers.
Hi-hats and cymbals, you can typically throw out things below 250Hz with great results. There is low energy there, but it doesn’t usually help your mix to keep it around, since it masks other more important things in the mix, reducing clarity and diminishing other sounds in the mix that are interfered with through “masking” interactions.
Most folks have speakers at home that can’t play back 35Hz anyway, unless they are using a discrete subwoofer. Even then, you rarely need much of this frequency area. What seems like REALLY LOUD bass is usually loud in a higher area, in the lows rather than the very lows.
I’m going to give you specific frequencies to consider removing for most instruments and vocals as I go through the tracking and mixing parts of my blog during early June, but I’ll also talk about…
EQ IN MASTERING
I’ll talk about this aspect in detail when I discuss mastering in late June for a few days.
As a mastering engineer, I have spent decades listening intently to the effects on finished stereo mixes that processing such as equalization can have.
Equalization is an art form in itself. It can be creative (do it for fun) or corrective (fix a problem). It’s there to be used at every step of the process in tracking, mixing and mastering. Mastering is the final step in making a record, the last chance to polish and “quality control” the sound before the final result goes on sale.
My own mastering signal chain typically includes a Manley Massive Passive, a highly regarded analog tube EQ that contains some 27 vacuum tubes (or valves, to Europeans) and adds a certain magical property to the audio simply by being present in the chain. It is not suited to every style of music or musical signal as the tubes do add incredibly subtle yet flattering amounts of noise and distortion to the sound. It sounds so good on so many things that it’s certainly not unusual for audio to be very delicately kissed by the Massive Passive on it’s way through my studio. I also use a Drawmer 1969 tube EQ and the Manley “Enhanced Pultec” midrange EQ on occasion.
Sometimes I need digital “linear-phase” EQs, which do not suffer from the phase artifacts caused by any analog EQ, no matter how good. They do have their own drawbacks too, though, so nothing does everything perfectly.
My main digital equalization comes in the form of DAW plug-ins, many of which model existing analog hardware, by Waves, Izotope, Brainworx, UAD, Sony, Digidesign, Sonalksis and quite a few others.
AIM TO CUT RATHER THAN BOOST
Now we come to the worst problem for the home studio. Most people, especially novices, overuse the EQ by an insane amount, and they are listening in spaces where they can’t really trust the lows and mids are accurate.
Then, they end up with a mix where things sound pointy or boomy, and, weirdly, things jump out at you every so often for a moment.
No need to go mad. The more you EQ, the worse things are likely to sound at the mix, and the harder it will be to control and balance everything in your mix. Believe it. This is the way it is.
Keep your hands where I can see them, and step away from the equalizer.
It’s easy to see how this arises. Frequencies interact with each other in all sorts of ways due to the harmonic series and the complex waves involved in almost all musical sounds. Things can resonate in sympathy with other things, or they can be absored or “masked” so you can’t hear them any moredue to some other sound in the mix.
Most sounds have a very complicated wave structure, and this structure usually extends to frequencies beyond the range of our hearing, sometimes in lows as well as highs. The bits we can’t hear are still there, and still interact with other sounds.
Sounds also interact with their environment, since they are traveling through the medium of air to reach our ears. This affects how you perceive sound too. It’s important to think about controlling room acoustics somewhat in the space you are recording and mixing. A thrift store might supply much of what you need in terms of heavy quilts and blankets and the like, or you can go to the music instrument retail store or online and get RealTraps or Auralex products, or similar.
Make your EQ decisions in the most neutral environment you can. Change the environment as best you can to facilitate this.
Just as the equalizer raises or lowers the amplitude of chosen frequencies, so does the behaviour of the room acoustics where you listen, or the type of speakers you listen on, even if you use headphones that eliminate the room part of the equation. This is a complicated matter, but there is no need to freak out just yet.
It also takes time to train your ears to listen critically enough. The simplest solution, since it is so easy to be misled, is to layback on the EQ use. Be subtle. Try not to use a lot unless it is necessary. If you must use a lot of boost or cut on a frequency, make sure nothing else gets a lot of EQ action so any sonic issues are minimized. Boosting usually causes overloading.
Another matter is making judgments in a “loudness-neutral” way. Whatever is the loudest sound is perceived as better by listeners for various pyscho-acoustical reasons. Loud is always the winner. This caused much of the “loudness wars” of the 2000’s in mastering room around the world.
To make a good judgment, try to make the EQ OUT and EQ IN sound as loud as each other for the purpose of deciding whether the EQ change is an improvement or not. It’s very important to do this, otherwise you will ALWAYS like the loudest sound the best. It’s really true, and true of all humans.
FREQUENCY RESPONSE AND OCTAVE RELATIONSHIPS
For convenience, let’s separate equalization response into the following areas:
VERY LOW = Below 80Hz
LOW = 80Hz to 350Hz
LOW MID = 350Hz to 2kHz
HIGH MID = 2kHz to 6kHz
HIGH = 6Hz to 12kHz
AIR = 12kHz and above
FREQUENCY TO NOTE RELATIONSHIP
Every musical note you tune to with your tuner has a defined, agreed-upon pitch. That pitch is equal to a specific frequency.
I defined low as being below 80Hz because the lowest string on a guitar or bass guitar is 82 Hz, the E note.
For example, here is the second octave of C up the piano keyboard, the lowest bass notes you might actually use in a song and not sound too weird doing it.
Note Freq (Hz)
Eagle-eyed readers will have noticed that the note A in this octave is at 110Hz. The octave above that, the A is 220Hz.
Go up another octave, and A is at 440Hz. This is the standard tuning for concert pitch. A = 440. Now you know what they mean. Every doubling of the octave doubles the frequency. You already realize that 880Hz will be A in the fifth octave of the C major scale on the piano, and that 1.76kHz must be the value of A in the sixth octave. Correct!
The Massive Passive is happy to boost or cut (or both, being a great EQ design) at the really high frequency of 27kHz, which is way, way above human hearing. Yes, you can hear the effect, but obviously you are hearing interactions lower down the spectrum rather than the actual change at 27kHz.
This reinforces my point about interactions being the key to getting used to what an EQ does to the sound when you use it.
LEARN TO LISTEN CRITICALLY
I started by saying this, and it bears repeating.
It takes a lot of practice to get good at this, but it comes with time, and is very well worth acquiring as a skill.
If you do want to engineer your own recordings, it will help you immeasurably to know what you are hearing reliably.
Ear training is really, really useful. Get an old graphic equalizer with, ideally, 31 bands per channel from a yard sale or pawnbroker. Live bands often get rid of them. Cheap gear is fine, although better gear will deliver more honest results.
Play some of your favourite music through it, and see what happens when you turn up just one of the bands, say the 1kHz band. Boost it so you can really hear it’s influence, then cut it the same amount and see what changed about the whole sound. Do that with all the bands on the EQ, one after another, and do it for a number of sessions. Over time, it will be easy to get good at hearing what to change when you feel like it.
That’s it for EQ. The next four posts will be about dynamics processing, which means using compressors and their ilk.
Plenty to talk about there, so I’ll be taking five days to cover it all. Thanks for reading! You rock.