TCRM 14 introduced the various types of audio filters (equalizers) and how they work. Now it’s time to consider when and how to use eq. But before we start, let’s remind ourselves that it is best to capture a sound that is close to the intended ideal when first tracking. Informed use of eq can work wonders, but eq can do nothing to cure a bad performance, remove distortion, or revitalize frequencies that are not there. So make a good recording… and use eq to make it even better.
The primary uses of eq are to:
- adjust the tonal character of individual instruments - carve spectral space within the mix - remove unwanted sounds/frequencies - help master the final mix to make it sound its best
These various uses can serve both technical and creative needs.
Sculpting tonal character
As mentioned in earlier installments of this column, it is important to have a strong concept of the exact nuances of the sounds you’re trying to achieve when recording. If the captured sound does not match that ideal, one way to address the problem is through the use of eq.
For instance, if the acoustic guitar sounds too “boomy,” eq might help to reduce the low-frequency energy. If a hi-hat recording is too “dark” (not enough highs) it might be made brighter by using high-shelf eq.
Sculpting spectral space in the mix
Balance in a mix is achieved through three main methods: controlling relative levels (dynamics), panning, and adjusting frequency space. The first two of these three methods, dynamics control and spatialization, will be covered more in-depth later in this series. For now, let’s concentrate on the use of eq to create balance in a mix.
The main concept behind frequency space in a mix is simple: every sound within the mix must coexist within the realm of 20 – 20,000 Hertz. When the majority of frequency energy of two or more instruments competes for the same part of this range, the sounds can mask each other and make the mix unclear. If you’ve ever heard an orchestra performance where an overly enthusiastic brass section seems to wipe out a full string section, you’ve heard a typical example of masking.
Closer to home, vocals, guitars, keyboards, and even snare drum sounds often battle it out for the crucial 1-5 kHz range. When this happens, adjusting faders and inserting compressors helps very little. To clear things up and let the lyrics shine through, the area must be cleared for the vocals. Notch filters can be used on the other instruments to reduce the energy within this range. The amount of reduction does not always need to be too great; a 3 to 6 dB cut will often suffice. While this technique may require a compromise in the tonal characters of the individual instruments, a cleaner overall mix results.
A word of wisdom: when considering treatments for any particular track, do not lose sight of its relationship to all elements of the overall mix. This sort of over-focus creates extra work, sucking up valuable time and causing high levels of frustration. All elements must serve the mix!
Taking out the trash
Besides modifying aspects of a particular instruments tone, eq can be used to remove unwanted sounds or tonal influences from a recorded track. Factors such as noise, rumble, wind, hum, proximity effect, room resonances, bleed from other mics, and both environmental and mechanical noises can ruin otherwise great takes. The ability to identify and remove these is essential.
After all of the tracks have been edited, treated and mixed, it’s time to create a final stereo (or surround) version. For many, this is the last opportunity to make the tune sound its best before being released for public consumption (in one way or another). Eq has already been added to the individual tracks, but now the overall sound may need some tweaking. This generally includes eq, dynamics, phase and format manipulations. If the individual elements have been treated properly, with the overall mix in mind, any final eq may be minor, or even unnecessary.
A quick note on mastering: Mastering is a precise and highly specialized endeavor. If you have the budget and care highly about the final audio product, hire a mastering engineer. (This coming from a die-hard do-it-yourselfer.) They have the objectivity, specialized equipment, and experience necessary to do the job. † Now, with the basic functions of eq in mind, let’s consider other aspects of its use….
Boost and Cut = Yin and Yang
The use of eq is a balancing act across the frequency spectrum. For every action, there’s an equal and opposite reaction. By boosting higher frequencies, for example, it can appear as if there has been a loss in the lower frequencies. Similarly, by cutting lows, the highs can be accentuated. This seesaw phenomenon is a function of both the relative energy levels across different frequency bands, and the nature of our hearing mechanisms.
So if some tonal adjustments can be made by either boosting or cutting, which do you choose? This depends on considerations such as musical and spectral context, filter design, headroom, and personal preference. While both boosting and cutting approaches are useful (and necessary), it is quite common for recording engineers to fall into the habit of choosing one over the other. For most, this†happens subconsciously.
There are some, however, who are both aware of, and adamant about, their knob-twiddling proclivities. Their reasoning is often either technical or philosophical in nature… sometimes both. Boosting is generally seen as a more active, even aggressive, approach. (Rock on!) Yet some feel that boosting adds unnecessary amounts of noise and distortion as compared to cutting. It can also quickly eat up the headroom of a mix by increasing the signal strength. This leads to one of the commonly held arguments in favor of cutting: that headroom is maintained (or increased) since energy is removed. In practice, this is not entirely true.
An odd eq occurrence: Using eq to cut frequencies from a hot track can actually boost the output level, causing clipping and audible distortion. But how could taking something away from a track cause a boost in level? It has to do with interference. Any given frequency can interfere, both positively and negatively, with any other frequency in the spectrum. By removing a frequency that negatively interferes with another (reducing it’s energy) the level of that second tone is essentially raised.
When using eq on a track that has been digitally “normalized,” or recorded such that its level maxes out at 0 dBFS (just short of clipping), either boosting or cutting can cause digital overs. Yuck. So remember: when using any eq, maintain adequate headroom so that the signal won’t clip. (refer to TCRM # 6 for more on headroom and gain structures)
Eq in practice
Outside of some basic technical considerations, the use of eq is mostly a matter of working style, genre and personal taste. Decisions regarding when and how to use it are based on factors as diverse as music itself, and the people that make it.
With that being said, there are lots of common uses for each basic type of eq. I will outline some below, as gathered from professional engineers and personal experience. (Special thanks to Bernie Mack of Flashpoint in Chicago) First we’ll explore the underlying considerations by filter type, and then by specific instrument. Use these as a reference or catalog of ideas, keeping in mind that they may or may not be the best approach in a particular situation. There are too many variables (such as personal taste, mics, acoustics, recording platform, converters, instruments, and musical style) for these to be absolute. But they are a good starting point.
Highpass filters are a fundamental part of getting a good sounding low-end mix. Muddy, undefined, or overly accentuated low end may be due to simply having too much going on down there. Often, there’s a lot of low-frequency information on tracks that doesn’t need to be there at all. For instance, microphones often pick up low-frequency room resonances, mechanical rumbling, air noises, and handling noise. A lot of this can be removed from the tracks of higher-register instruments by using highpass filters.
It is amazing how much this simple technique can clean up the lows in a mix. It leaves space for the low instruments (such as bass and kick) to come across clearly. In addition, since directional microphones emphasize the lower frequencies as they get closer to the sound source, highpass filters are used to combat this proximity effect where it’s not wanted.
When using a DI to record electronic instruments directly, there is often excessive low and/or high-end information captured. This is because the amp and speaker, which normally impose a restricted frequency response on the signal, are bypassed. A HPF is sometimes used, along with a LPF, to emulate this restricted frequency curve.
When a HPF, with its continually declining slope, is too aggressive a tool, a low shelving filter can be used instead. Since these shelves can also boost the lows, they are especially handy on bass and kick drum sounds.
Lowpass filters are used to reduce higher frequency content. Sometimes, this is in the form of noise, sometimes as excessively bright cymbals or guitars. Lowpass filters are also handy in limiting the range of synths, which can generate both extremely high and low frequencies.
As with LPFs, high shelves can be used to attenuate noise or overly bright sounds, but they do so more evenly than LPFs. Since high shelves can also boost, they are also great for adding high-frequency definition and excitement to cymbals, vocals, and other sounds with high-frequency content.
Parametric notch (peaking) filters are particularly suited to finding and addressing issues relating to specific, limited frequency areas and those within the middle registers. They are invaluable for their ability to help identify problem areas through a common, yet powerful, technique I like to call either the “Search and Destroy” or “Search and Rescue” method (S&D or S&R).
The essence of this technique is simply to boost a specific area, and slowly sweep the eq frequency while listening for a specific element to jump out in the mix. This is especially telling if the bandwidth is kept small (1/3 octave or less). This method is handy for such things as: finding vocal formant frequencies and boosting them, finding and reducing hum or buzz, and finding “pinging” or “ringing” tones in drums to either increase or attenuate them, as personal taste demands.
Eq in action
While overhead microphones are often relied on to capture an overall picture of the drum kit, they are absolutely essential for recording the sound of the cymbals. This being the case, a HPF can be used on the overheads to remove kick and other low-frequency sounds that are being recorded by other mics set up for that purpose. The cutoff frequency can be set around 100 Hz for dealing with the kick, or as high as 350 Hz when isolating even more of the cymbals. A high frequency shelf is also useful for boosting the signal above 8 kHz to get more crispness from the cymbal sound.
Getting a powerful, yet clear, kick drum sound with a distinct attack is one of the most difficult but fundamental aspects of recording drums. Many try to get more gusto from the drum by boosting the low end with a shelf or notch filter. While this can lend more power to the sound, it often makes it very unclear or “muddy.”
Generally, removing the low mids (anywhere from 200-400 Hz) will help clean it up, even in conjunction with an accentuation of low frequencies in the 40-80 Hz range for power. The attack of the beater can be emphasized by a notch filter set to the 4-7 kHz range.
Bringing out the “snap” attack in the snare drum is important as it defines the beat. Notch filters work great for finding and boosting these elements in the 2-5 kHz area. Boosting the 10-12 kHz range can bring out more of the high-frequency noise-burst sound by making the rattle of the wire snares more obvious. Of course, it may sometimes be desirable to cut some of this area, thereby accentuating more of the head/shell components of the sound.
Snare drums can also have very obvious modal resonances that cause “ringing tones.” There are lots of opinions as to whether these are desirable, most of which revolve around musical style. Ska music relies on them heavily, while the sound is extremely rare in hair-band metal. The S&D/R method works wonders to remove or accentuate these tones as desired. Of course, proper snare selection and tuning are best for getting the right tones in the first place.
Sometimes, the kick bleed in a snare track is attenuated with a HPF. Similarly, cutting around the 300-400 Hz area with a notch filter can bring out some of the lower hi-hat modes. Be careful here, as this may remove some of the body of the snare tone.
Lower frequency toms, especially “floor” types, benefit greatly from stressing the lowest tones. This is similar in concept to the kick technique and is often accomplished with a notch (peaking) filter set to boost 60-100 Hz. As a general rule, the lowest modal frequency of the toms increases as their size decreases. Therefore, the center frequency of the notch must also increase.
Also similar to sculpting a kick sound, the toms benefit from reducing specific low-mids, thereby accentuating both the lower body of the tonal character along with the higher attack elements. Reductions in the 300-450 areas can do the trick. If even more attack clarity is called for, try boosting the 4-5 kHz range.
Hi-hat tracks often suffer from both kick and snare bleed. This can be reduced by using a HPF set with a cutoff somewhere from 80 to 200 Hz. Sometimes, especially dependant on mic placement, a hi-hat can have accentuated formants in the low-mids, causing either a gong-like quality or the beautiful sound of banging on a garbage can lid. This effect can be de-emphasized by reducing the 280-400 Hz frequency area. Clarity of the stick attack and excitement in the highs can be brought up by boosting the highs, either by a notch filter from 8-12 kHz or by using a shelf to boost everything above 10 kHz.
Though there is an enormous range of guitar tones possible from the amplifier, there are a few basic ideas that may help electric guitars sit better in the mix. First, though some guitarists like to get a lot of low end out of their rigs, it competes for this property with the bass and kick drum. In the mix, the guitar must sometimes be reminded of its proper place.
If the low end is making the mix unclear or muddy, a HPF or low shelf should be used to remove this. Somewhere around 80 Hz serves as a good cutoff frequency. A boost in the 200-250 range is often heard as increasing the “warmth” of the sound. A boosting peak or shelf up above 7 kHz can sometimes be used to increase the clarity of the instrument. As mentioned earlier, removing some of the 1-5 kHz area can help make the vocals clearer as needed.
Low frequency modes from the body cavity, mixed with proximity effect, make even an acoustic guitar a potential cause of low frequency overpopulation. Can you say HPF? While the area of “warmth” is similar to that of the electric guitar (ca. 200-300 Hz), “clarity” and pick/strumming sounds can be either accented or diminished in the 3-7 kHz range.
Electric bass (with amp)
Use a low shelf to boost the lowest harmonics and the power of the bass. Accentuate below 100 Hz. At the same time, lowering the 100–140 Hz range with a notch can help clear things up a bit. The 200-350 range is, once again, the “warmth” area. For the bass, 3-7 kHz brings out the pick attack and upper harmonics to make the pitch of each note more prominent.
Many factors affect the character of the voice. Age, gender, size, and tessitura (range), are a few of the most influential. Furthermore, unlike most instruments, the influence of formants and their frequencies change with both pitch and phoneme.
This makes specific eq techniques and important frequency content hard to specify. Instead, general concepts and ranges must be given, and informed experimentation must be utilized to determine the proper settings from singer to singer… even from song to song, or mic to mic.
Bringing up the 2-5 kHz range can aid in vocal intelligibility. This is called a presence peak, and is often designed into dedicated vocal microphones. If the 3-5 kHz range is stressed too much though, there can be an over-accentuation of sibilance (“s”, “sh”, and “t” sounds). A de-esser may be needed to get the best overall sound without emphasizing these harsh sounds.
When proximity effect does not lend enough low frequency presence to the sound, 60-90 Hz can be boosted to bring out natural chest resonances. These give us an idea of the size of an individual and their proximity to us.
Sometimes, when a singer’s sound has a very nasal quality, it’s wise to reduce 300-400 Hz. If the sound is a bit dull, and more “brilliance” or “sparkle” is called for, a high shelf can be used to boost upwards of 10 kHz.
To sum up eq use in a single statement: Identify the sound you want to have, compare it to the sound you’re getting, and use eq to make the two match. Next month I’ll outline some alternative uses of filters… as effects.
Thanks again to Bernie Mack who’s a recording and mixing engineer in Chicago as well as on the faculty of Flashpoint. His candid input over the years on specific eq usage has been invaluable.
John Shirley is a recording engineer, composer, programmer and producer. He’s also a Professor in the Sound Recording Technology program at the University of Massachusetts Lowell and chairman of their music department. Check out his wacky electronic music CD, Sonic Ninjutsu, at http://cycling74.com/products/c74music/
†Supplemental Media Examples
The following audio examples demonstrate the eq techniques described in the body of the TCRM 15 article. Drums, guitar, and female vocals are used here, but the workflow and methodology translate to all recording endeavors.
The following is a snare as originally recorded (no eq, dynamics or other effects). How does it sound? How can you describe the technical differences between what it sounds like and how you would prefer it to sound? TCRM15_1.wav
One thing that the original recording has is an abundance of ringing modal tones. While this may be a positive thing in some types of music, here let’s try to reduce them.
We begin the process by identifying the frequencies of the ringing tones. To do this, we first use a notch (peaking) filter to find the ringing by boosting and sweeping with a high q (around 7 to 9). Here, a ringing tone is found at 468 Hz. (see TCRM15_pic1.jpg) Compare this recording to the original. TCRM15_2.wav
As there’s more than one noticeable modal frequency in this recording, we can use the same method to find and reduce these as well. TCRM15_4.wav
Even more highs (3.5 dB) are added using a high shelf at 7.3 kHz. TCRM15_6.wav
Next we reduce the low frequencies by way of a high-pass filter at 135 Hz, to remove some of the bleed from the kick. TCRM15_7.wav
Since the removal of the modes and lows has also reduced some of the sense of lower body of the snare tone, I add a bit back by using a notch filter to search for and accentuate some of this energy. I find it at 202.8 Hz and bring it up a little (3.1 dB) TCRM15_8.wav
Now, the kick drum from the same session as originally recorded: TCRM15_10.wav
This following file works it’s way through all four EQ functions outlined above. TCRM15_14.wav
The original hi-hat recording: TCRM15_15.wav
A notch can be used to reduce some of the “dullness” of the tone in a similar fashion as was done to the kick. Here, it can be found around 374 Hz (Q = 2.63, f = 374.3 Hz, Gain = -11.9 dB) (see TCRM15_pic9) TCRM15_16.wav
The crisp sound of the stick hitting the hi-hat is accentuated. Here, I found two frequencies that added this sense. (Notch1 @ Q = 9.96, f = 7.06 kHz, Gain = 5.1 dB AND Notch2 @ Q = 4.77, f = 9.01 kHz, Gain = 7.2 dB) (see TCRM15_pic10) TCRM15_17.wav
Now let's listen to an example of the toms. Here, the low tom as originally recorded: TCRM15_19.wav
With eq, the exciting lower resonance of the drum can be brought out, both by boosting around 75 Hz and some reduction around 368 Hz (a problem frequency in this particular recording). The attack of the sticks is accentuated by way of a positive notch at 4.2 kHz. (see TCRM15_pic12) TCRM15_20.wav
The original drum overhead recording: TCRM15_21.wav Overheads pick up a number of important sonic elements: the sound of the room (reverb), the cymbals, balance (a more natural sense of the kit than close-miking,) and the stereo image of the kit as set up in the recording space. When eq-ing the overheads, important decisions must be made regarding whether each of things should be accentuated or decreased in importance.
Overhead Approach 1
For the next example, we will stress diffuse field reverb, cymbals and high-frequency stereo imaging. The factors of balance within the kit, and overall imaging are left to the close mic tracks.
To achieve this, the ringing frequencies of the snare are removed using notch filters, the lows are brought out with a HPF set at 448 Hz, and a high shelf is used to accentuate the highs a bit more. (see TCRM15_pic13) TCRM15_22.wav
Overhead Approach 2
Another approach is to create a good overall mix in the overheads. This tries to achieve success in all of the areas mentioned above. Sometimes, such a good sound can be achieved this way that little or no close mics are needed in the mix at all. (see TCRM15_pic14 & pic15) TCRM15_23.wav
Put it all together
When mixing a drum kit, the balance, tone, and location of each instrument must be considered. Making things even more complicated is the fact that there are so many mics, each picking up elements of all of the instruments, regardless of which one(s) they are placed on.
For example, you can spend quite a bit of time getting the sound just right on a soloed snare track only to discovered that the sound is compromised when the hi hat or overhead tracks are brought into the mix. Eq decisions must often be reevaluated once all of the channels are in the mix. It’s just part of the process….
Here’s the mix of the kit with all of the eq taken out: TCRM15_24.wav
Now, a section with low toms. First the non-eq'd version: TCRM15_26.wav
Then with eq: TCRM15_27.wav
Now, let's eq the guitar on the same tune. Here's what the original take sounds like: TCRM15_28.wav
Interestingly, there are troublesome ringing tones on this track similar to what we heard in the snare. We can use the same "search and destroy" approach to take these out. Here's what it sounds like when a notch filter is used to find and accentuate the problem area: (see TCRM15_pic16) TCRM15_29.wav
To add clarity and accentuate the attack and pick sounds, a boost at around 5 kHz does wonders: TCRM15_32.wav
This song also includes a female vocalist. The recording is pretty good, but it does contain minor amounts of some of the most common frequency contour problems for vocals. First, give the original recording a listen: TCRM15_34.wav
There's a slight accentuation around 340 Hz that's making the vocal sound a bit dull and "boomy". This is not bad here, but is a very common problem in close-miked vocal recordings. Here's what happens if we accentuate this. (see TCRM15_pic20) TCRM15_35.wav
Now, we can use a notch to cut this area back a bit. TCRM15_36.wav
We may also want to cut the lows, where a lot of wind, mechanical, and handling noise occurs. Since doing this sometimes removes too much body from the singer, making them a bit thin sounding, we should then add some low mids back in. (see TCRM15_pic22 for final settings) TCRM15_38.wav
In the context of the whole mix, the vocals sound a bit dull without the eq: TCRM15_39.wav
With the eq created here, the vocals sound clearer and more present: TCRM15_40.wav
The End Result
So here's the entire mix, without any of the eq added. TCRM15_41.wav
Now with eq on everything except the vocals. A touch of reverb has also been added. TCRM15_42.wav
Finally, the mix with eq on the drums, guitar and vocals (still with reverb). TCRM15_43.wav
Special thanks to The Bay State for the use of the raw tracks to this demo recording of “Liars.” All material used by permission; all copyrights reserved. Demo recorded by Michael Testa.
The Bay State is Tom Tash, Drew Hooke, Susanne Gerry and Evan James.
Check out their music on iTunes (including the commercial release of “Liars” or visit them on facebook at: http://www.facebook.com/thebaystate