To learn how to de-ess a vocal, put a de-esser (a frequency-specific compressor) on the vocal after your main compression, set its detection band over the harsh “S” range — usually somewhere between about 5 kHz and 10 kHz — then lower the threshold until only the loudest esses are reduced by roughly 2–6 dB. Work by ear: you want the sting gone, not the whole word softened into a lisp.
Violet Recording is reader-supported — we may earn a commission from links on this page, at no extra cost to you.
Sibilance is the sharp, hissy energy in “s”, “sh”, “ch”, “t” and “z” sounds. A little is natural and helps a vocal cut through a mix. Too much is fatiguing, and it gets worse the moment you brighten a vocal with EQ or squash it with a compressor — both push those high frequencies forward. De-essing is simply the fix: it turns down that harsh band only when it spikes, and leaves the rest of the performance untouched.
What sibilance is and why it needs taming
Every voice produces sibilance, but where it lives depends on the singer and the mic. Deeper male voices often peak lower, around 4–7 kHz. Brighter voices, and many female vocalists, push their harshest energy higher, roughly 6–10 kHz. Bright condenser mics and close mic placement exaggerate it, and so does the modern habit of adding a high-frequency “air” boost. None of that is a mistake — it just means the esses need managing so they sit at the same level as the rest of the words instead of jumping out and stabbing the listener.
The goal is balance, not removal. If you can no longer hear the “s” in a word, you have gone too far. Good de-essing is invisible: the vocal sounds smooth and expensive, and nobody can point to what you did.
How to de-ess a vocal step by step
Here is the reliable, repeatable process. It works with the stock de-esser in almost every DAW — Logic, Ableton, Reaper, Cubase, Pro Tools and FL all ship one — so you rarely need to buy anything to get started.
1. Get the de-esser in the right place. Insert it after your compressor and any brightening EQ. Compression and high-shelf boosts raise sibilance, so de-essing last means you are catching the problem at its worst, after everything that made it louder.
2. Find the sibilance frequency. Most de-essers have a “listen” or “solo” button that lets you hear only the detection band. Turn it on and sweep the frequency control up and down while the vocal plays. When the soloed sound is almost entirely “sss” and “shh” with very little vowel or body, you have found the spot. That is usually somewhere in the 5–10 kHz window described above, but let your ears set the exact number.
3. Set the threshold. Switch back to normal listening and lower the threshold until the meter shows gain reduction only on the esses — not on sustained vowels or held notes. Aim for about 2–6 dB of reduction on the worst syllables. If the meter is pumping on every word, your band is too wide or your threshold is too low.
4. Check the range or amount. Many de-essers have a “range” or “max reduction” control that caps how much they can duck. Setting a ceiling of around 6–8 dB stops the plugin from over-reacting on one loud line and mangling it.
5. Bypass and compare. Toggle the de-esser on and off at a realistic listening volume. The harshness should ease without the vocal losing its crispness. If words now sound like “th” instead of “s”, back off the threshold or narrow the band.
Wideband vs split-band de-essing
De-essers work in one of two ways, and knowing the difference saves a lot of frustration.
Split-band (or multiband) mode only turns down the narrow high-frequency band where the sibilance lives, leaving the rest of the vocal’s tone alone. This is the most transparent choice for most pop, rap and singer-songwriter vocals, and it is the mode to reach for by default.
Wideband mode turns down the whole vocal signal whenever the detector hears sibilance. It can sound more natural on some sources because it preserves the tonal balance of each word, but it also causes a subtle dip in level every time an “s” hits, which can sound like flutter if pushed hard. Try both on a tricky vocal and keep whichever is less noticeable. There is no universally correct answer — it is genuinely a by-ear call.
Where the de-esser goes in your vocal chain
Signal order matters more than any single setting. A dependable starting chain looks like this: subtractive EQ to clean up mud → compression to control dynamics → de-esser → any additive “air” EQ and effects sends. The logic is that you de-ess after compression (which lifts sibilance) but you have a choice about EQ. If you add a big high-shelf boost, put a second de-esser after it, because that boost will re-introduce the harshness you just tamed.
On busy productions, one de-esser is often not enough — some engineers use a gentle one early to catch the worst offenders and a second, lighter one near the end of the chain to polish. Two subtle passes almost always sound better than one aggressive pass. If you want a refresher on how EQ and compression interact before you de-ess, our guide to EQ and compression fundamentals covers the groundwork, and the broader how to mix vocals walkthrough shows where de-essing fits in the full vocal treatment.
Manual de-essing: the surgical alternative
A de-esser is automatic, but it treats every “s” the same. When only a handful of syllables are painful, manual de-essing gives cleaner results. Zoom into the waveform, find the loud sibilant, and either draw a short volume-automation dip over just that “s” or use clip gain to pull it down 2–4 dB. Because you are only touching the problem spots, the rest of the performance keeps every bit of its natural brightness.
Manual work is slower, so a sensible workflow is to let a de-esser handle the overall consistency and reserve hand-editing for the two or three esses that still poke through. This combination is how a lot of professional vocal edits are actually done.
Common de-essing mistakes to avoid
Over-de-essing into a lisp. This is the number one error. If you remove all the high-frequency energy from “s” sounds, the singer sounds like they have a speech impediment. Always compare against bypass and stop as soon as the harshness is gone.
Setting the band too wide. A wide detection band drags vowels and consonants into the reduction, dulling the whole vocal. Keep the band focused on the sibilant energy you isolated in step two.
De-essing before you compress. If the compressor comes after the de-esser, it re-inflates the sibilance you just controlled. Order matters — de-ess late.
Trying to fix a recording problem in the mix. If sibilance is extreme, revisit the source. Angling the mic slightly off-axis, moving the singer a touch further back, or choosing a less bright mic can cut the problem at the source. Our notes on microphone placement for vocals and recording vocals at home are worth a look before you reach for plugins.
When stock plugins aren’t enough
For the vast majority of vocals, the de-esser built into your DAW does everything you need, and you should not feel pressured to spend money. That said, on genuinely difficult material — a very bright vocal with sibilance smeared across a wide range, or a track where you need dynamic, frequency-aware detection — a dedicated third-party de-esser or a dynamic EQ can be a real upgrade over a basic stock unit, offering multiple bands, cleaner detection and better metering. If you hit that wall, it is worth browsing specialist options at. For most people, though, master the stock plugin first.
It is also smart to catch sibilance before the track leaves your hands. Harsh esses that survive into the master are hard to remove later, and an online mastering service such as LANDR will faithfully preserve whatever you send it — including problems. Fix sibilance at the mix stage and the master will thank you. If you are heading toward that finish line, what is mastering and LUFS explained cover what happens next, and you can check reference loudness figures with our free audio tools.
Frequently asked questions
What frequency should I set my de-esser to?
There is no fixed number — it depends on the voice and mic. As a starting point, deeper voices often sibilate around 4–7 kHz and brighter voices around 6–10 kHz. Use the de-esser’s solo/listen mode and sweep until you hear mostly “sss”, then set the band there. Trust your ears over any preset.
Should I de-ess before or after compression?
After, in almost every case. Compression raises the level of quiet detail, including sibilance, so de-essing afterward catches the esses at their loudest. If you also add a bright high-shelf EQ later in the chain, consider a second light de-esser after that boost.
How do I know if I’ve de-essed too much?
Listen for a lisp. If “s” sounds turn into soft “th” sounds, or the vocal loses its crispness and starts to feel dull, you have overdone it. Raise the threshold so the de-esser only ducks the harshest peaks, and always A/B against the bypassed vocal at a normal volume.
De-essing is one of the highest-value moves in vocal mixing, and it takes minutes to learn. For more techniques, browse the full mixing and mastering hub, and see how reverb interacts with a treated vocal in our guide to using reverb and delay.
