Sometimes I will refer to previous articles since I don't want to repeat myself too much.
I know it's been some time since the last article but I've been very busy. Not the least with recording, mixing and putting online new music, which you can find at my site at mp3.com :-) But I will try to keep this series alive and kicking.
The following articles will probably be shorter than those that are already online. While the previous articles were more or less like a consecutive series and should be read from part I to VI, the following ones will have a certain topic but will be independent from each other. Nonetheless I will keep on numbering them, for convenience.
This article is about equalizers and how you work with them during the mixdown of a song (or other piece of music). Some people say that equalisation is bad and you shouldn't use it but work on making better recordings in the first place so you don't have to EQ your tracks. Of course it is always recommended to make as good recordings as possible, but you can't do all with microphone placement etc. especially not if you are working in a home studio and have neither the best microphones nor a really good sounding recording room available. The latter is probably the biggest problem of the home studio.
And you should always be wary of such dogmatic statements like "never use EQ". In real life, nothing is black and white. Using words like "never" and "always" is of one use: it allows you to stop thinking of certain things. But it also limits your options. People used to say that man would never fly to the moon... even scientists are not immune to the dangerous use of the word "never"... Philosophical mode off now.
Anyway, those "never use eq"-sayers have a point. You can't fix everything with eq. But you can do an astonishing lot of things with it. So you should not abstain from using it. Try to get away with as little EQ as possible, but use as much as needed.
An equaliser (all kinds, analogue, digital, whatever) processes the signal and alters its frequency response. It can boost or cut certain frequency ranges that are already present in the signal. It should not create frequencies in the output signal that were not present in the input signal. If it does, it distorts ("adds harmonics" is a friendly way to put it). Sometimes this is desired - with tube or tube-emulating equipment for example. But the ideal EQ does not add distortion of its own.
A common equaliser is most times divided into several bands. A band is a filter that is defined by a set of parameters:
center (or cutoff) frequency: this is basically the middle of the frequency band that is affected or the frequency above or below the eq works.
bandwidth (or Q): this defines how wide the range of frequencies that are affected is.
gain: if this is positive, the eq boosts the frequencies selected by the center frequency and the bandwidth, if it is negative, the frequencies are cut.
the frequencies below (highpass) or above (lowpass) the cutoff frequency are cut. The order of the filter determines "how fast" the frequencies are cut, how steep the slope of the filter curve is. There are 1st, 2nd, 3rd, 4th or higher order filters. The steepness of the slope is usually measured in db/octave, that means how many dB per octave the signal below or above the cutoff is attenuated.
These filters, if implemented in a synthesizer, often have a "resonance" control. This controls how much the frequencies around the cutoff are boosted before the attenuation starts. A resonant filter is the secret behind the most noticeable synth sounds.
^ gain | /\ <- resonance peak | / \ 0 |------------ \ | \ cool, huh? | \ | \ -----------------------> freq. | cutoff
If you wondered of what relevance this is for guitarists: a wah pedal is either a resonant bandpass or a resonant lowpass filter (bandpass is more common).
Some filters have a resonance control that has such a range that it leads to self-oscillation - the filter circuit starts to emit sine oscillations with the cutoff frequency if the resonance is turned up way high. The Minimoog filter is an example for this.
|Filter order||slope steepness||usage examples for lowpass|
|1st||6 db/oct||passive R/C stages like a guitar tone knob|
|2nd||12 db/oct||some synthesizers|
|3rd||18 db/oct||the famous 303 bass synth has 18dB/oct filters, those give that special "acid" sound|
|4th||24 db/oct||classical synth filter, the legendary Minimoog filter had 24 db/oct. This is a pretty steep filter already.|
Shelving filters have in common with the highpass/lowpass filters that all the frequencies above or below the cutoff frequency are influenced. Only they do not cut all frequencies that are attenuated to zero, but they lower them a adjustable amount. A frequency response curve for a shelving filter would look somewhat like this (only rounder)
gain ^ | /----------- boost | / 0 |-------< | \ | \----------- cut | | -----------------------> freq.
Examples are the tone controls on your home stereo, those are usually shelving filters with a fixed cutoff frequency and steepness, where you can control the gain only. On the most smaller mixing consoles, the bass and treble bands of the channel strip eq are made as shelving filters. On high-class mixers you can sometimes choose between a "normal" eq band and a shelving band for the uppermost and lowermost eq bands. Some software mixers allow this too.
Those affect only a certain frequency range around two frequencies. A bandpass lets through only frequencies inside the band while the bandstop does the opposite, it prevents frequencies inside the band. A really extreme version of the bandstop is the notch filter, which, when done right, completely removes all content on a very small frequency band. Those kinds of filters can be hardly audible, but very valuable when you want to get rid of, for example, 50Hz line hum. Feedback suppressors work with notch filters, too.
Graphic equalisers have a certain number of eq bands, of which only one parameter, the gain, is user-adjustable. The other parameters, like Q or frequency, are fixed. There are some common spacings (that means the distance between two adjactent bands frequency-wise) for graphical eqs, 1/3 octave or 2/3 octave for example. 1/3 octave Eqs usually have 28-31 bands to cover the whole audible spectrum, while 2/3 eqs have 15 bands.
Common uses for graphic eqs are as a master eq in a live PA, to adjust the overall sound, or as a master eq on the whole stereo mix. One does not use a graphic eq (commonly, that is) on a single channel strip. This is a) overkill and b) nowhere near as effective as a parametric eq.
Those are common on small-to-midsize consoles. You have usually two controls for one eq band: frequency and gain. So you can adjust the center frequency of the filter and the gain, but not the bandwidth, which is fixed. How many bands you have depends on the console. Most smaller mixers have shelving filters for treble and bass (as I said above) and one (maybe two) semiparametric band for the mid frequencies.
Those are the most powerful and most difficult to use equalisers. All parameters of an eq band are user-adjustable - that means three controls for each band, gain, frequency and bandwidth (or "Q"). Fully parametric eqs are found in expensive big mixers, as standalone units or in software programs (which is gonna interest us most). The mixer EQs of cubase of samplitude for example are fully parametric.
Because it gives you so much user-control, use of a parametric EQ must be learned. Watching an old pro working with a parametric eq can be amazing. Two quick turns of knobs, and he has found the frequency range that wasn't quite right (too dominant or too weak) and fixed it. And you stand next to him and think "how the heck did he do that?". And he tells you that you have to learn the eq and you have to have an ear for the frequencies. Most of the time I couldn't coax more information out of the people I tried to learn from. They were right: it takes practice to use such an EQ effectively. But you don't have to wait for the magical ears to appear on your head and suddenly you can tell 900Hz from 1kHz...
I will now tell you how I memorized certain frequency ranges.
The basic working process is simple. First get some sounds/music with a wide frequency range. It has to be material that you are familiar with. You should have a memorized idea of how it sounds with no EQ. Get a parametric eq. Set the bandwidth not too narrow. A semi-parametric will do, too. It should be able to sweep through the whole audible range. Set the gain to boost, lets say, 10 dB. Don't change the gain control through the first part of the process. This is a drastic change of sound, don't worry, it's only for educational purposes.
Now get a sheet of paper. First write down four or five basic frequency ranges like:
use more or less just like you desire. This is no rule, it's just a tool to help you learn the EQ. Everyone perceives sound in a different way, so you should use your own definitions.
Now sweep the eq through the frequency range. Note the boundaries of the above frequency ranges as you perceive them (your eq should show you at what frequency you're working). You won't be able to hunt them down to the single Hertz, but you should have a general idea of "this is the bass range" and "these are the lower mids" etc. Write the boundary frequencies between the basic frequency ranges on your paper. The idea behind this is that when you think a piece of sound needs adjustment in the treble range, you already have a rough idea of what frequency range you have to work in. Now you have a reference to look at and don't have to guess the frequencies that much anymore. If you work with an eq on a regular basis, you will memorize these quickly.
For me this list would look something like this (these are my subjective peceptions, yours will differ):
Now this is a pretty rough reference chart. Fine tuning an EQ will take more and better definitions.
The EQ is still set to boost, right? Now on to the "fine associations". Move the EQ frequency through one of the basic ranges. Different frequencies will still sound different. Write down the adjectives that you associate with the sound, together with the frequency the EQ is at. For example, like in case of the low mids, "honky". Or "flabby" for a certain bass range. "Harsh" for a certain treble range and so on. Do this until you think that you have covered most of the audible frequency range. You might try different material on the way. Now you should have a pretty long list of adjectives with the associated frequencies. Now take a break. Listen to some music without eq, to "reset your ears". After you done that and feel rested, set the gain of the EQ to cut, say -10dB. Repeat the "fine association" process with the eq now set to cut. You will notice that you associate quite different adjectives with similar frequencies than during the "boost" part of the process.
After you finished this, you probably have a pretty chaotic bunch of notes (or you are better organized than me :-) Sort your notes and make a clean table that looks like this:
Sounds like: Sounds like: Freq. | when cut | when boosted ------------------------------------------ ... | ... | ... 500Hz | hollow | honky ... | ... | ... etc.
I won't give you my associations because a) you will probably have different ones and b) they're all in german :-)
The idea of all this is to have a reference chart. If you feel that a certain track sounds hollow, it probably misses a bit in the 500Hz range (to use the above example). Now you can boost this range a bit to compensate. After some time you won't have to look this list up everytime. You will probably forget parts of it too, but hopefully you will have memorized how certain frequency ranges sound *to you*. Then you can work quickly and effectively with a parametric EQ.
Another nice tip: if you feel that a track has too much of a certain annoying frequency range, don't try to find it while setting the EQ to cut. Set it to boost instead and sweep through the frequency range until the annoying frequencies get particularly nasty. You have then found the right range to cut. This works surprisingly well.
I hope that with the help of these "association tricks" you will be able to master the usage of a parametric EQ, as it is very important for mixing. The above method may seem a bit too analytic and not very musical. Well, you shouldn't listen to material you actually want to mix with the same method. You should only be able to translate the sonic "feel" of a sound into parameter values you can work with on the EQ. Then use the EQ to obtain the musically most satisfying result.
I think the next article will be about compression, and will be called:
"Compression explained for beginners" or "The compressor is your friend, really" :-)
I now that "compressed" is not a word that's loaded with positive associations in the guitar world, but I hope that I can show you that compression is indeed your best friend in the recording situation (after the EQ, that is).
Have fun and make music
If you want me to write another article on a specific topic, drop me a line. Or if you have a specific question, that is not answered in this series. Or if you want to comment/criticize/whatever. Just give some feedback if you want more info and about what.
Back to Part VI
Back to the index page
All this stuff is (c) 2000-2002 Tammo Trueper.
The trademarks I have mentioned all belong to their respective owners.
If you want to drop me a note, or ask a question, here's my eMail address: