What is a de-esser? It’s a frequency-specific compressor that automatically turns down harsh “S”, “T”, “SH” and “CH” sounds — the piercing high-frequency spikes known as sibilance — without touching the rest of a vocal. When a singer or speaker hits an “S”, energy stacks up in a narrow band (usually somewhere between 5 kHz and 10 kHz). A de-esser listens for that spike and ducks it for the split second it happens, then gets out of the way. The result is a smoother, less fatiguing vocal that still sounds bright and natural.
Violet Recording is reader-supported — we may earn a commission from links on this page, at no extra cost to you.
Sibilance is one of the most common problems in home-studio vocals, and it’s almost impossible to fix with EQ alone. This guide explains what a de-esser does under the hood, walks through setting one up step by step, covers the mistakes that make vocals sound worse, and names a few solid free de-essers you can grab today.
What is a de-esser doing that a normal EQ can’t?
A standard EQ is static. If you pull down 6 kHz to soften a harsh “S”, that cut stays in place for the entire track — including the parts where there’s no sibilance at all. You end up dulling the whole vocal to fix a problem that only appears for a fraction of a second at a time. Bright consonants lose their crispness, air disappears, and the voice sounds muffled and lifeless.
A de-esser is dynamic. Think of it as an EQ cut that only switches on when it’s needed. Internally, it’s a compressor with a filter on its detection circuit (the sidechain). The detector only “hears” the high-frequency band you point it at. When energy in that band crosses a threshold — i.e. when an “S” arrives — the de-esser reduces the level for just that moment. When the sibilance passes, the gain returns to normal and the vocal keeps all its brightness. That’s the whole trick: surgical, momentary, automatic.
How sibilance happens and why it’s worth fixing
Sibilance is a natural part of speech — you physically can’t say “sixteen sisters” without it. It becomes a problem in recordings for a few reasons that stack up. Bright condenser microphones exaggerate high frequencies. Close mic placement (common at home) pushes more high-end energy into the capsule. Compression during tracking or mixing raises quiet parts and makes the “S” peaks stick out even more. And boosting the “air” band or adding a high-shelf for sparkle pushes sibilance further over the edge.
Left untreated, harsh sibilance is fatiguing to listen to, and it gets worse the louder someone turns it up. It can also trigger distortion on cheap earbuds and phone speakers, and it eats into your loudness headroom during mastering. Taming it early keeps the rest of your mixing decisions — EQ, compression, saturation — from amplifying the problem downstream.
The key controls on a de-esser
De-essers vary in layout, but almost all of them share the same handful of parameters. Learn these and you can drive any de-esser plugin.
- Frequency (or centre/target): Where the de-esser listens for sibilance. For most vocals this sits between 5 kHz and 9 kHz. Male voices tend to sibilate lower (around 4–7 kHz); brighter female and child voices sit higher (7–10 kHz).
- Threshold (or amount/range): How loud the sibilance has to get before the de-esser reacts, or how much gain reduction it applies. This is your main “how strong” knob.
- Bandwidth (Q): How wide a slice of frequencies it acts on. A narrow band is surgical; a wider band catches more but risks dulling other consonants.
- Mode — split-band vs wideband: Split-band (or “split”) only turns down the offending high band, leaving the rest of the vocal untouched — usually the cleaner, more transparent choice. Wideband ducks the whole signal when sibilance hits, which can sound more natural on some sources but risks a subtle pumping “lisp” if pushed hard.
- Listen / solo / monitor: A button that lets you hear only what the de-esser is removing. This is the most useful control for dialling one in accurately.
How to use a de-esser: a step-by-step setup
Here’s a reliable workflow that works on any vocal, in any DAW.
1. Place it correctly in the chain. A good default is after your main compression and corrective EQ, but before heavy saturation, reverb and delay. Compression tends to push sibilance up, so de-ess after it. If a single de-esser can’t keep up, split the job — one gentle instance before compression, one after.
2. Find the sibilant frequency. Engage the “listen” mode and sweep the frequency control while the vocal plays. You’re hunting for the setting where the “S” and “SH” sounds are loudest and clearest in the soloed signal, and where you hear the least vowel or body. When the isolated sound is mostly “essss” and not “aaah”, you’ve found the spot.
3. Set the threshold by ear. Come out of listen mode and lower the threshold (or raise the amount) until the harshness softens. Watch the gain-reduction meter — you’re usually aiming for just a few dB of reduction on the loudest “S” sounds, and little or nothing in between. If the meter is pinned down constantly, your frequency is wrong or you’re pushing too hard.
4. Choose split-band first. Start in split-band mode for transparency. Only switch to wideband if the vocal sounds unnatural or if you specifically want that gluey, whole-signal duck.
5. A/B against bypass. Toggle the plugin on and off at a normal listening volume. The vocal should sound smoother with it on, but you shouldn’t be able to say “oh, there’s a de-esser.” If you can hear it working, back it off.
Common de-essing mistakes to avoid
Over-de-essing is the number-one error. Push too hard and every “S” turns into a soft “th” — the singer develops a lisp. Consonants carry a lot of a vocal’s intelligibility, so removing too much makes the words mushy and the delivery lifeless. Less is almost always more; aim to reduce harshness, not eliminate the “S” entirely.
The second mistake is treating the whole vocal with one static setting. Sibilance varies line to line — a shouted chorus sibilates differently from a whispered verse. If one de-esser can’t cope, use two gentle ones, or automate the threshold, or handle a few extreme words with volume automation or clip gain instead of hammering the plugin.
Third, don’t reach for a de-esser to fix a problem that lives elsewhere. If the whole track is harsh — not just the “S” sounds — that’s an EQ or microphone-choice issue, not a de-essing one. And check your mic technique: a slight off-axis angle or a pop filter can reduce sibilance at the source so you need far less processing later. Fixing it while recording always beats fixing it in the mix.
Where to get a de-esser plugin
Nearly every DAW ships with a stock de-esser, and for a lot of vocals it’s all you need — learn the one you already own before spending money. When you want more control, cleaner detection or a smoother split-band engine, dedicated de-esser plugins are inexpensive and worth it. You can browse and compare paid options at a store like Plugin Boutique, which regularly runs sales on vocal-processing tools.
Best free de-esser plugins worth trying
If you’d rather not spend anything, several genuinely good free de-essers exist. These are widely used and well regarded:
- TDR Nova (Tokyo Dawn Records) — a free dynamic EQ. It isn’t labelled a “de-esser”, but you can set a dynamic band in the sibilant range to duck it precisely, which makes it a flexible, transparent de-esser and much more besides.
- TDR Nova — a free dynamic EQ from Tokyo Dawn Records; set a band over the sibilant range (around 5–8 kHz) with dynamic gain reduction and it works as a precise, frequency-targeted de-esser.
- Your DAW’s stock de-esser — most DAWs include a capable de-esser (or a dynamic EQ that does the same job). Learn it first before buying anything, as it will handle the vast majority of vocals.
- Your DAW’s stock de-esser — Logic, Cubase, Studio One, Reaper (via ReaXcomp/ReaEQ dynamic bands) and others all include capable de-essing tools at no extra cost.
Free dynamic EQs like TDR Nova are especially powerful here because de-essing is really just dynamic EQ pointed at one band — once you understand that, you can de-ess with tools you may already have.
Frequently asked questions
Where should a de-esser go in my vocal chain?
A dependable default is after corrective EQ and your main compressor, but before saturation, reverb and delay. Compression tends to raise sibilance, so it helps to de-ess afterward. If sibilance is severe, use two light de-essers — one before compression and one after — rather than a single heavy one.
What frequency should I set a de-esser to?
There’s no single number — it depends on the voice. Most vocals sibilate somewhere between 5 kHz and 9 kHz. Deeper male voices often sit around 4–7 kHz, while brighter voices can reach 8–10 kHz. Use the plugin’s “listen” mode and sweep the frequency until the soloed sound is mostly “S” and least vowel.
Can I de-ess with an EQ instead of a de-esser?
A static EQ cut works only if you dull the vocal permanently, which usually costs you brightness. A dynamic EQ, however, is effectively a de-esser: set a band in the sibilant range that only cuts when that band gets loud. That’s why free dynamic EQs make excellent stand-in de-essers.
Keep learning
De-essing is one piece of a larger vocal-mixing picture. To put it in context, read our guide to mixing vocals and our primer on EQ and compression fundamentals, since a de-esser is really both of those combined. For the effects that come after it in the chain, see how to use reverb and delay, and when you’re ready to finalise a track, our introduction to mastering explains where sibilance control fits into the bigger loudness picture. You’ll find more tutorials in the Mixing & Mastering hub.
