The take is perfect. The pitch, the emotion, the timing, all there. Then you push the high end for a little air, and every “s” turns into a snake.
The harder you brighten the vocal, the worse the hiss gets until the esses are the loudest thing in the chorus and you cannot unhear them.
That hiss is sibilance, and the fix is de-essing.
On vocals it is a balancing act, because the same frequencies that carry the harsh esses also carry the air and clarity you want to keep.
Take too much, and the singer sounds like they are lisping; take too little, and the hiss still cuts.
Here are the techniques for de-essing vocals cleanly: where sibilance hides on male and female voices, the methods pros actually use, and how to tame the esses without dulling the performance.
TL;DR
- Sibilance is the harsh “s,” “sh,” and “t” hiss, and on vocals it sits roughly between 4 kHz and 10 kHz depending on the voice.
- Male sibilance tends to sit lower (4 kHz to 7 kHz); female sits higher (6 kHz to 9 kHz) and is often more pronounced.
- A dedicated de-esser is fastest, dynamic EQ is more transparent, and manual volume automation is the cleanest of all.
- Place de-essing after EQ and compression, before reverb and delay.
- Use the least reduction that works, usually 3 dB to 6 dB on the worst moments, so the vocal never lisps or goes dull.
The frequency map comes first, then the six techniques, then the habits that keep all of it invisible.
What Sibilance Is and Why Vocals Get It
Sibilance is the burst of high-frequency noise created when air rushes through the teeth on consonants like S, T, Z, and the “sh” sound.
Every singer produces some, but a few things make it worse: a bright condenser mic, close mic technique, and a voice that naturally has a lot of high-end energy.
The recording stage is where it starts, and the mix is where it usually gets amplified.
That is the catch with vocals.
The EQ boosts and compression that make a vocal sit forward also raise the level of the esses, so a vocal that sounded fine raw can turn harsh once it is mixed.
Those boosts usually come from the air-and-presence pass in how to EQ vocals like a pro; sibilance is the bill for that brightness.
De-essing exists to pull those peaks back down without sacrificing the brightness around them.
Stop Sibilance at the Source
The best de-essing is the de-essing you never have to do.
A few choices at the recording stage cut the problem before it ever reaches your mix, and they cost nothing.
You will rarely eliminate sibilance this way, but starting with less of it makes every technique below gentler and more transparent.
- Angle the mic slightly off-axis so the burst of air from the mouth does not hit the capsule head-on.
- Back the singer off the mic a few inches to reduce the proximity-boosted high end.
- Use a pop filter, which tames some high-frequency air along with plosives.
- Choose a less bright mic for a naturally sibilant singer instead of fixing it later.
Where Vocal Sibilance Lives
Sibilance does not sit at one fixed frequency.
It moves with the voice and the sound, which is why the first step of any de-essing job is to find where it actually is on your singer.
This map gives you the starting ranges to sweep, but always confirm by ear.
Sibilance is also just one band of many; the vocal EQ cheat sheet maps the rest of the voice around it.
| Source | Range | Note |
|---|---|---|
| Male vocal sibilance | 4 kHz to 7 kHz | Sits lower than a female voice |
| Female vocal sibilance | 6 kHz to 9 kHz | Higher and often more pronounced |
| “Sh” and “ch” sounds | 3 kHz to 6 kHz | Broader and lower than a hard “s” |
| Hard “s” and “t” | 7 kHz to 10 kHz | Sharp and narrow |
| Bright or condenser-mic vocals | Shifts upward | Exaggerated highs need careful targeting |
To pin down the exact frequency, drop a narrow EQ band with a big boost into the range from the table and sweep it slowly while the vocal plays.
The spot where the esses become almost unbearable is your sibilant frequency. Note it, remove the boost, and set your de-esser or dynamic EQ to target that point.
This boost-and-sweep trick takes seconds and beats guessing every time.
Techniques for De-Essing Vocals
There is more than one way to tame sibilance, and the best choice depends on how much of it there is and how transparent you need to be.
Here are the six methods, roughly from fastest to most surgical. Most vocals only need one, but a stubborn performance can call for two working together.
Use a Dedicated De-Esser
The fastest route is a dedicated de-esser, a frequency-targeted compressor built for exactly this.
You set the target frequency, lower the threshold until it catches only the esses, and dial in a few dB of reduction.
It is the go-to for everyday vocals because it takes under a minute.
On most voices, target the frequency from your sweep, set the threshold so the meter moves only on the esses, and keep reduction around 4 dB.
How to use a de-esser walks the controls one by one, and best de-esser plugins ranks the current field if you are still choosing one.
Dynamic EQ (the Transparent Option)
A dynamic EQ band does the same job with more precision.
Because you control the exact frequency and bandwidth, it often sounds more transparent than a de-esser.
It ducks the sibilance only when it crosses the threshold and leaves the rest of the high end open. It is the modern pro favorite for clean vocals.
Use a narrow band, a fast attack, and a range of just a few dB so it engages only on the harsh moments and releases instantly.
The walkthrough on dynamic EQ on vocals covers the setup.
Static EQ Band Reduction
The quick-and-dirty method is a narrow static EQ cut in the sibilant range. Sweep a tight bell to find the harsh frequency, then pull it down a couple of dB.
It works in a pinch, but it is the least transparent option, because the cut is always on, even during the words that had no sibilance, which can dull the vocal overall.
Keep the cut to 2 dB or 3 dB with a tight Q, because any more and the dullness becomes obvious on the non-sibilant words.
Multiband Compression
A multiband compressor can be set to compress only the high band where sibilance lives, which makes it a capable de-esser.
Set the crossover just below the sibilant band, then compress that band only, with a fast attack and release and a high ratio.
It is more setup than a dedicated de-esser, but it is useful when you already have one on the chain for other reasons.
Manual Volume Automation
The cleanest method of all is also the most tedious: ride the volume down by hand on each sibilant word using clip gain or volume automation.
Because you are only touching the exact moments that spike, nothing else about the vocal changes.
Pull each spike down 2 dB to 4 dB and shape the automation curve so the dip is inaudible.
It is overkill for a whole song but unbeatable for a handful of stubborn esses a de-esser cannot tame gracefully.
Split the Hard and Soft Esses
On a vocal with wildly uneven sibilance, one de-esser set for the worst esses will over-process the mild ones.
The fix is to split the work: use two gentle de-essers in series, or de-ess in stages before and after compression, so no single processor has to clamp down hard.
Set each one for a gentle 2 dB to 3 dB and let their reductions stack into smooth, even control. Spreading the reduction keeps every “s” natural.
Tips for Transparent Results
The goal of de-essing is to be invisible. Done well, no one notices it; done badly, the vocal lisps or goes lifeless.
These habits keep it on the right side of that line, and they are echoed in Sound on Sound’s rundown of vocal de-essing techniques.
- Judge it in the mix, not soloed. Sibilance that screams in solo often sits fine against the track.
- Find the exact frequency with the de-esser’s Listen mode rather than guessing.
- Less is more. Aim for 3 dB to 6 dB of reduction on the worst moments, no more.
- Two gentle passes beat one heavy one on a stubborn vocal.
- Re-check after you add air or brightness, since those moves bring sibilance back up.
- De-ess the lead and backing vocals separately. Stacked harmonies build sibilance fast, so a gentle de-esser on the backing-vocal bus stops the esses from piling up.
- Bypass often as you work. Your ears adjust to the processed sound, so A/B against the untouched vocal to make sure you have not gone too far.
Common Mistakes
Nearly every bad de-essing job is one of these three. The fix each time is a lighter, tighter touch.
1. Over-de-essing into a lisp. Removing too much of the “s” leaves the listener unable to make out the words, turning “sun” into “thun.”
Use the least reduction that controls the harshness and stop there.
2. Processing the whole vocal for a few bad words. If only a handful of esses spike, a heavy global de-esser dulls everything else.
Reach for dynamic processing or manual automation so you only touch the offenders.
3. Targeting the wrong frequency or too wide a band. Aim too low or too broad, and you duck the body and air of the voice along with the esses.
Sweep to pinpoint the sibilance and keep the band tight.
Frequently Asked Questions
Sibilance questions are mostly where-and-how-much questions. Six short answers, with the sweep trick doing the heavy lifting.
What is de-essing?
De-essing is the process of reducing sibilance, the harsh “s,” “sh,” and “t” sounds, in a vocal.
It works by turning down a narrow band of high frequencies only when those sounds spike, so the harshness is tamed without dulling the rest of the vocal.
You can do it with a de-esser, a dynamic EQ, a static EQ cut, multiband compression, or manual volume automation.
What frequency is sibilance in vocals?
It depends on the voice. Male vocal sibilance usually sits between 4 kHz and 7 kHz, while female sibilance sits higher, around 6 kHz to 9 kHz, and is often more pronounced.
Hard “s” and “t” sounds are sharp and narrow up at 7 kHz to 10 kHz, while “sh” sounds are broader and lower.
Always sweep to find the exact spot on your singer.
Should a de-esser go before or after compression?
Usually after. Compression raises the level of sibilance as it evens out the vocal, so de-essing afterward catches the true, final amount of harshness.
Some engineers add a light de-esser before the compressor too, so the compressor is not triggered by the esses.
Either way, keep de-essing before reverb and delay.
Why do my vocals sound lispy after de-essing?
You are removing too much of the “s.”
When the de-esser clamps down hard or targets too wide a band, it strips the sibilance the listener needs to understand the words, so “sun” starts to sound like “thun.”
Reduce the amount of gain reduction, tighten the band, and aim for the lightest touch that controls the harshness.
Is dynamic EQ better than a de-esser for vocals?
Often, yes, for transparency. A dynamic EQ lets you target the exact frequency and bandwidth so it can duck sibilance more precisely and leave the surrounding air untouched.
A dedicated de-esser is faster and simpler, which makes it the better choice when speed matters.
For the cleanest possible result on an important vocal, dynamic EQ usually wins.
How much de-essing is too much?
If the vocal starts to sound lispy, dull, or thin, you have gone too far.
As a guide, keep gain reduction to about 3 dB to 6 dB on the loudest esses and let quieter ones through untouched.
De-essing should make the harshness disappear while leaving the vocal sounding natural and bright, not processed.
The Bottom Line
De-essing vocals is about restraint.
Find where the sibilance lives on your singer, pick the method that fits the job, and remove only as much as the harshness demands.
A dedicated de-esser handles most vocals in seconds, dynamic EQ gives you the cleanest result, and manual automation rescues the stubborn esses nothing else can.
Keep it subtle and no one will ever know it is there.
De-essing is a narrow job done with tweezers. The complete vocal mixing guide covers the broad-brush work around it, from prep through effects.