The difference between a good mix and a great mix can often be the vocal sound. Learn how to properly sculpt the high end of vocals so that they’re present but not harsh, airy but smooth, poppy but not sibilant.
By Dan Zorn
A great sounding vocal separates seasoned mixers from beginners, and when mixing vocals, controlling the high end is especially important in achieving a “professional” sound. Doing this correctly can be tricky, as there is an art to getting the vocals present and “pop” in a mix, yet also smooth and caressing to the ears. Moreover, the human ear has evolved an acute perception toward high frequencies, so there’s not much room for error in the high register. The good news is, all of these challenges can be overcome by knowing what to listen for.
The range of human hearing is approximately 20Hz to 20kHz, which fades away as we age, starting with the highest frequencies. A teenager may be able to hear up to 19 kHz, while an elderly person may only hear up to 15khz, or even less. Furthermore, the focal point of our hearing, or the frequency range we are most sensitive to, is 3.5-4khz. The extra sensitivity we experience from this range exists since the length of the ear canal matches the actual wavelengths of these frequencies. As a result, even at very low levels, these are the first frequencies we notice. In order to perceive other frequencies of the hearing spectrum with the same loudness as those in the 3.5-4kHz range, their physical volume must be higher in dB. The Fletcher-Munson Curve shows exactly how much louder:
The Fletcher-Munson Curve confirms that we are most sensitive to the upper mid-range, so any mix adjustments here must be made carefully. For example, on vocals, a 5dB cut at 3.5khz is more noticeable than a 5db cut at 100Hz, so considering the nature of how we experience different frequencies is imperative.
Also, keep in mind that high frequencies travel shorter distances than lower frequencies with longer wavelengths. In turn, the amount of high end in a vocal determines how far away the listener will feel from the vocalist. Vocals with less highs will always sound further away, while vocals with a present high end confront the listener more closely. Hip hop vocals, for example, should be particularly impactful and “in-your-face,” so making sure a rapper has enough top end is crucial to the musical emotion of the genre.
High End can be a broad, loosely-defined term among those involved in music. For example, listeners may claim a vocal is bright, sibilant, or airy, so understanding the attributes and unique language used to describe high end is important to help us decide which specific area to address. Let’s dive in.
Intensity refers to the lowest section of the high-frequency spectrum, approximately 1.5khz to 3khz. A subjective description of the “intensity range” would be piercing, powerful or just downright annoying. However, if you need vocals to cut through a mix volume-wise without physically turning up the overall gain, this could be a good area to boost.
The word harsh typically describes a shrill or cold sound and generally speaking, harshness exists in the range of 3kHz-5kHz. Usually, this is the problematic area bothering listeners when they claim “the vocal sounds too bright,” even though “bright” can also be a term to describe higher, airy bands.
Sibilance describes vocals containing consonants sounding too sharp or “ess-y.” Often times, consonants like the letters T, X or S, or words containing a Sh or Ch can be piercing. These are the main culprits, but other letters can have sibilant qualities too, depending on how they are pronounced. These quick, sibilant “spikes” generally happen in the 4-10kHz range, and can be annoying to the ears, as well as interfere with other instruments. Sibilance must be tamed and controlled for a nice high end.
Some mixers associate the air-bands around 8-9kHz and up with descriptions of “sheen” or “presence.” This area tends to give the listener a sense of “closeness” to the vocalist.
Let’s go over how to polish these high-end areas in a mix. Remember, for a full-sounding vocal, all of these frequency areas should be balanced with one another, and also against the music. Not only will every voice’s high end be different, but so will the accompanying instruments making up the highs of the mix, cymbals, hi-hats, etc. If the instrumental has a lot of high end already, you may not need to add as much high end in the vocal as there simply won't be room for it in the mix.
When mixing high frequencies in vocals, control is the name of the game. Traditionally speaking, controlling dynamics first, prior to treating problematic frequencies with EQ, tends to be an effective approach. The reason being, if the vocals are very dynamic, you may react by using EQ to cut a frequency area that was overbearing in only certain parts of the performance. In turn, the sound will be smooth when the vocals are loud; yet, suddenly too dull, or hollow, in quieter sections.
There are, of course, exceptions to this rule, as many engineers will tell you - “Rules? What rules?” However, you may conclude it’s useful to control dynamics after EQ so that you can hear the build-ups and issues with the high end more easily if you decide to brighten them up first. By doing this, you may be able to dial in a compressor, multiband compressor or de-esser more precisely.
You may even use some dynamic control before and after EQ, which is totally fine if it sounds good. The point is to think critically about what exact problem needs to be addressed, and then creating an effective, efficient chain of tools to get the job done.
Frequency build-ups in the 1.5- 3khz range can have numerous causes, many of which are fixable before mixing: the microphone or its placement, the preamp, the distance of the vocalist from the mic during a loud performance, or even the natural resonance of the vocalist. Regardless, even if all of the above is considered when tracking, dynamic treatment is often needed in mixing.
A great way to tame this range dynamically (and transparently) is with a narrow-band de-esser. I like to use the Waves R-DeEsser in Split mode for this. Using the sidechain function, you can solo in on the frequency range you want to control, then adjust the threshold and reduction accordingly. The sidechain method can work well if you know the exact frequency you want to tame, but sometimes it's easier to pull up a static EQ, like the Waves Q10, and sweep around with a medium Q boost between 1.5khz to 3khz to pinpoint the problem frequency that jumps out the most. Then, you can type that frequency into the DeEsser and voilà, precise control that kicks in only when needed.
Too much reduction in the 3-5kHz range will leave the vocals sounding dull, but overly dynamic information in this range can create an unpleasant listening experience. You have to use your ears and discretion here to try to identify what the real problem area is. Try to listen to the mix at different volumes to gain perspective. Remember, our ears are most sensitive to these frequencies. Consider if there’s just too much of this entire range or if there are smaller areas poking that need more precise control.
A great, all-in-one tool to control this range is the Waves C6, a multiband compressor that allows you 6 bands of individual, customizable control, ideal for maximum flexibility. If this area seems a bit too dynamic and things tend to jump out at you, then you can set one of the C6 bands Q to 3-5khz, then set the threshold to only attenuate when things are poking too much. Furthermore, if you find you’ve tamed the dynamics, but still notice build-up, you can always reduce the gain of the band within the C6 (effectively EQing it). The benefit of this method is the initial compression will lessen the amount of static gain you would otherwise reduce with an EQ. Transparency is the name of the game!
Let’s say there's something still not sitting right. Perhaps, the issue may be a small resonant buildup or “whistle” frequency. C6 can work great for this, as it also offers 2 sidebands that have the option for very narrow Q’s, perfect for small whistle tones.
Similar to harshness, taming sibilance prior to EQ is best, so that you can turn up the high end later without these consonants jumping out.
Sibilance occurs lower in the spectrum for males than it does females, due to the natural differences in octaves that males and females speak in. For males, 4-10khz is a good range to listen, whereas female sibilance may start higher at 5-6khz.
Adding a DeEsser is the classic way to reduce sibilant spikes. In addition to using narrow-band de-essing, the Waves R-DeEsser also allows you to attenuate everything above a set frequency in the form of a shelf. De-essing techniques often work, but sometimes even more control is necessary.
A great way to get rid of sibilant spikes outside of de-essing is to manually lower the volume of each sibilant peak with clip-gain edits or volume automation. This allows you to tailor the level and envelope of each sibilant spike on a word-by-word basis, since parts of a song may need more “esses” to poke through the mix, particularly against parts in instrumentals that are overly busy.
As of late, my preferred approach is to zoom in and chop the sibilant peaks on to a new track. Similar to manually lowering each sibilant peak, using a separate track allows for processing each sibilant separately, aside from just volume control. Treating sibilant peaks on a separate track, versus lowering clip-gain within one track, opens up new worlds for transparent processing: adding deeper EQ dips, de-essing, compressing or limiting harder, or even adding more saturation to smear some of the transient spikes.
The upper “air “band is one of the more gratifying areas to work on. Unlike the other areas of the high-end spectrum, treating this area usually means highlighting, or bringing more to life. You’ll usually find that boosting this range (above 8-9khz) will add pleasant sheen and definition to vocals, especially once you’ve dealt with the previous issues.
While the 9khz range is usually clear of sibilant spikes, certain vocalists will have frequencies that still poke through in this area. You can, of course, go back to any of the tools we mentioned earlier to tame this area, like a de-esser set at a higher frequency, or activating the top band on the C6 to suppress high, transient spikes. However, another way to address this problem is to add a bit of tape saturation to “round out” the spikes of the upper register. Using the Waves Kramer Master Tape, you’ll find that by lowering the Tape speed, raising the flux, adjusting the bias and playing around with various input and output ratios, it can truly help smear some of those unwanted high-end transients. Then when all is said in done, you can add further sheen with an EQ. Some of my go-to plugins for adding a nice top-end sheen to vocals are the Waves API 550 and SSL G-EQ.
Getting the high end of vocals to sound present and pop, yet smooth and soothing can seem daunting at first. However, the real trick here is to not rely on anything but your ears and to ask yourself questions. The more questions you ask yourself, the more you can hone in on what the real problem is and ultimately make smarter choices.
Training your ears to identify specific issues does not come overnight. After all, if treating the high end of vocals was as simple as just “turning the treble down,” anyone would be able to do it.
Want more on mixing vocals? Read 12 essential steps for mixing lead vocals here.
Want to get more tips straight to your inbox? Subscribe to our newsletter here.