We are presenting a series of three instructional articles in which we look at the most widely used audio processors, discussing some of the techniques commonly used with them. This first article deals with equalization; in the next one we will talk about dynamics processors, and the final special will cover reverb and delay.
Frequency spectrum
On average, the human ear perceives sound from 20 Hz to 20,000 Hz, with these limits varying according to factors such as age, prolonged exposure to high sound pressure levels...
This range of frequencies is called the audible frequency spectrum and is usually represented logarithmically. This means the axis on which frequencies are "drawn" does not progress linearly (1, 2, 3, 4, 5...), but logarithmically. The effect is that it goes from 10 to 100 in steps of 10, from 100 to 1,000 in steps of 100, from 1,000 to 10,000 in steps of 1,000, and so on.
There are two main reasons for using this type of scale. On the one hand, these axes are very useful for representing quantities whose extremes are very far apart. Try drawing a linear scale from 1 to 20,000 on a piece of paper and you will quickly understand...
On the other hand, the human ear perceives changes in frequency in a roughly logarithmic way: subjectively, the distance between A4 (440 Hz) and the next A5 (880 Hz) is the same as between A5 (880 Hz) and A6 (1760 Hz), even though in the first interval the distance is 440 Hz and in the second 880 Hz.
These intervals, in which a frequency doubles, are called octaves, and the audible spectrum is very commonly represented in octave bands:
31 - 62 - 125 - 250 - 500 - 1k - 2k - 4k - 8k - 16k (Hz)
When more detail is required, third-octave representation is used, in which there are two more bands between each of the previous ones (31 - 40 - 50 - 62 - 80 - 100 - 125 ...).
An instrument's spectrum
When equalizing, we must keep in mind that a musical instrument does not sit at a single point in the audible spectrum: it covers a range of frequencies.
So a kick drum may have a fundamental at 70 Hz, and boosting this frequency will give you more power, but the attack of the beater against the head may be around 2-3 kHz, and that can be a good spot to boost if we want the kick to cut through the mix more.
Similarly, the body of a voice may be around 200 Hz, but intelligibility is around 3 kHz, and to reduce sibilance we should look around 6-10 kHz.
A good way to find these frequency points is to apply a generous boost and sweep the center frequency until we hear the sound we are looking for suddenly appear. Once we know which frequency we are after, we can apply the appropriate boost or cut.
You can look up charts showing the fundamental notes and harmonics of various instruments. These frequencies will be approximate and can be taken as reference points from which to start looking for an instrument's sweet spot.
Frequency and subjective sensation
Regardless of the instrument in question, certain areas of the spectrum have subjective sensations associated with them:
A boost around 31-125 Hz gives a sense of strength and power. At 125-400 Hz we find the warmth and body of many instruments. Around 400-600 Hz there is a lot of richness in the sound, but too much of these frequencies can make the mix sound muddy and confused.
1000 Hz gives a nasal sound, while around 2 kHz instruments gain hardness and intensity. The area around 3-6 kHz is very important, since it is where the human ear is most sensitive, and it provides presence, definition and excitement.
Boosting the 8-12 kHz band adds brightness, although too much can exaggerate sibilance. Many people describe the frequencies above 12 kHz as the area of "air" and lightness.
Equalizer parameters
The basic parameters of an equalizer are:
- - Center frequency: as its name suggests, the frequency around which gain is applied.
- - Gain: increase (positive gain) or decrease (negative gain) in the signal level.
- - Quality factor (Q): the relationship between the center frequency and the width of the filter. Low Q values indicate wide filters and high values represent narrow filters.
Equal-loudness contours
Loudness is a subjective measure of how intensely a sound is perceived by the human ear; its unit is the phon. Equal-loudness contours show us the intensity in dB required for a tone to be perceived at the same loudness, as a function of frequency.
To construct them, 1 kHz is usually taken as the reference frequency, where intensity and loudness are made to coincide. That is, for a 1 kHz tone to be perceived at a loudness of 20 phons, an intensity of 20 dB is required. However, if the tone is 200 Hz, we will need almost 40 dB of intensity to achieve the same 20-phon sensation.
Probably the best-known equal-loudness contours are those produced by researchers Fletcher & Munson in the 1930s. Today, the ISO (International Organization for Standardization) has the ISO 226 standard to define these contours.
Analysis of the equal-loudness contours shows that the frequency response of the human ear varies with intensity level, with the variation being especially critical at the lowest frequencies and, to a lesser extent, at the highs.
We need to bear this in mind when applying EQ, since the listening level we work at will make the applied EQ sound one way or another. For this reason it is advisable to check mixes at different levels, to make sure our work "translates" well at all listening levels.
A common practice is to mix at moderate listening levels and turn up the level occasionally when working on the low end (low frequencies). In general, if our lows are balanced, a mix that works at low levels will also work well at high levels.
Masking effect
Masking is a psychoacoustic effect whereby a louder sound (the masking sound) makes a quieter one (the masked sound) barely audible or inaudible. This explains, for example, why we have to shout at a concert to make ourselves heard: our voice is being masked by the music.
Masking is most effective when the two sounds are close in frequency, but it can also occur between frequencies far apart in the spectrum, in which case it is more effective when the masking sound is lower than the masked one.
When it comes to equalization, masking can mean that boosting the low frequencies makes the highs harder to perceive and, conversely, cutting the lows can make highs that were previously masked become audible.
That is why it is a good idea to start EQing in the low frequencies and then move on to the highs, so that work on the low end doesn't alter the rest of the spectrum.
Likewise, masking also explains negative equalization techniques, in which adjacent frequencies are attenuated so that certain frequencies become more audible.
Equalizing
Microphone equalization
Curiously, the most important stage of equalization does not use any gain, frequency or Q control.
We are talking about the choice of microphones and their placement. This is where we should experiment and look for the sound we want to get, whether by trying different mics, playing with their position, or in any other way we can think of. When we press the "rec" button, we should be completely satisfied with the sound we are about to record.
It is a mistake to place the mics somewhere they sound more or less fine, thinking "I'll EQ it later." As with any other processor, the ideal is to use EQ to make an already good sound excellent, not to cover up a bad recording.
If we have a good performance from the musician, the correct tuning of the instrument (obvious as it may seem) and the right choice of mics and their placement will do more for our sound than any EQ applied afterward.
The Big Picture
When applying EQ, it is a good idea to define beforehand where we want to go. We need to assess which instruments are in the mix and identify the main ones, since these will be the ones we make "bigger," while instruments that only appear occasionally or don't play an important role in the song can be "smaller."
Some engineers talk about the "Big Picture," in which they position all the instruments in the spectrum so that each one predominates in certain areas without getting in the way of the others. When equalizing, we should keep this big picture in mind so as not to boost the same frequencies on two different instruments.
Which instruments are important is determined by the song itself, although there are certain common patterns depending on the musical style:
In pop and rock, the most important instrument is generally the vocal, while the kick and bass take up much of the low end, supporting the rest of the instruments. In jazz, the most important instrument is the soloist, whether it is a guitar, piano, double bass, etc. In hip-hop and rap, the kick and bass are usually enormous (often with sub-kicks and sub-basses), and the vocals very present and loud.
In any case, the number of instruments to be mixed determines the EQ applied. In general, the more instruments there are, the more we need to limit the spectrum of each one, and the fewer instruments, the bigger they can be.
EQ holes
To make several instruments sound good at the same time, the EQ holes technique is used, in which we apply subtractive equalization (frequency cuts) to one instrument at a certain point in order to apply additive equalization (frequency boosts) to another instrument at that same point.
In essence, what we are doing is "digging holes" in one instrument to fill them with another, and vice versa, so that the resulting spectrum is more or less even.
Subtractive equalization is a very powerful tool, although we often forget about it. When we don't like how something sounds, the tendency is to mask it by boosting what we find pleasant, instead of unmasking what we like by removing the unnecessary frequencies.
On your next mix, try not applying a single positive dB of EQ and use only subtractive EQ. You'll be surprised how easy the mix becomes afterward.
You will probably have to boost a few spots at the end so that everything sounds the way you are used to, but almost certainly much less than you would have if you hadn't used subtractive EQ.
In general, additive EQ is used to make an instrument sound more exciting or different from how it has sounded so far, and negative EQ is useful for getting several instruments to sound good together.
Another thing to keep in mind is that we shouldn't spend too much time EQing an instrument in "solo," since what sounds good on its own may not work alongside the other instruments. What matters is that the final mix works, not each instrument separately.
As a final tip, trust your ears: if it sounds good, it is good. There are no rules. If you have to use +17 dB to get the sound you're after, go ahead. What matters is how it sounds, not what you did to get there.
