How to De-Ess Vocals With an Audio Equalizer
To de-ess a vocal with an audio equalizer, find the band where the harsh "s" and "sh" sounds live, most often around 5–8 kHz, and turn it down a few dB in Vocal Cut's free online audio equalizer: a gentle cut at 6 kHz, or at the 12 kHz shelf if the sizzle sits higher. It takes a few minutes in your browser, needs no account, and your audio never leaves your device. That handles mild to moderate sibilance; for severe, spitty esses a dedicated de-esser does the job better.
What you'll need
- A vocal recording as MP3, WAV, FLAC or M4A. The more isolated the voice, the better: a cut on a full mix dulls the cymbals too.
- Headphones. Sibilance is easy to miss on laptop speakers and easy to over-correct on harsh earbuds.
- A test passage with plenty of "s", "sh", "t" and "z" sounds, so you can hear every change.
What sibilance is (and why it sits around 5–8 kHz)
Sibilance is the hissy, high-frequency energy of consonants like "s", "sh", "z", "ch" and sometimes "t". Those sounds come from air forced through a narrow gap between tongue and teeth, which makes noise concentrated high in the spectrum. Every voice needs some: without it, words lose their edges.
It becomes a problem when the esses jump out louder than the rest of the voice and sting the ear. Where they sit varies with the voice, mic and distance. It can land anywhere from roughly 4 to 10 kHz, most often around 5–8 kHz, and higher, lighter voices tend toward the top of that range, which is why you find it by ear rather than by a fixed number. What audio frequencies are is a quick primer if the ranges are new to you.
De-essing is any process that turns those sibilant peaks down, either with a static EQ cut (this guide) or a dynamic de-esser (covered below).
Two other problems get confused with sibilance:
- Plosives are low-frequency thumps on "p" and "b", at the opposite end of the spectrum. See how to remove plosives and pops.
- Hiss is steady, broadband noise under the whole recording, even in pauses. For that, use the Noise Reducer and the guide to removing hiss from audio. The Noise Reducer targets steady background noise, so it won't fix esses.
How to de-ess with an audio equalizer, step by step
Vocal Cut's Equalizer is a 6-band graphic EQ with fixed bands: an 80 Hz low shelf, peaking bands at 240 Hz, 750 Hz, 2.2 kHz and 6 kHz, and a 12 kHz high shelf. Each slider runs from −12 to +12 dB in 1 dB steps. For sibilance, only the top two bands matter.
- Load your vocal. Open the Equalizer and drag your file onto the drop zone, or click to browse. The "Graphic EQ" panel appears.
- Press Play. The status line reads "Previewing — move the sliders to hear it change live", so every move is audible straight away.
- Find the problem band by exaggerating it. Briefly push the 6 kHz slider up to around +6 dB. If the esses get much nastier, that's your band. Return it to 0 and try the same on the 12 kHz shelf. Whichever boost makes the "s" sounds stab hardest is where the problem lives. Then set it back to 0: the boost is only a magnifying glass. If you want to see it as well as hear it, the Spectrogram shows each "s" as a bright vertical streak high up the frequency scale.
- Cut gently. As a starting point, pull the 6 kHz band down about 2–4 dB. If the sizzle is a bright, airy "tss" rather than a hard "s", try lowering the 12 kHz shelf 1–3 dB instead, or as well. Going much beyond about −6 dB is rarely needed.
- Listen for the lisp. Play the passage again and listen to clarity, not just the esses. If "s" starts sounding like "th", or the voice turns dull, back off 1 dB at a time.
- A/B against the original. Note your settings, click Flatten to reset every band to 0 dB, and listen. Then put the cut back. If the improvement isn't clear, use a smaller cut.
- Export. Choose WAV (lossless, the default), MP3, or Opus where your browser supports it (Safari may hide it), then click the export button, which reads Export WAV, Export MP3 and so on. Reset File clears the tool for the next take.
Audio equalizer settings for sibilance
There's no universal best setting, but these are sensible starting points on this EQ. Judge every one by ear.
| Symptom | Band to try | Starting move |
|---|---|---|
| Hard, stabbing "s" and "sh" | 6 kHz (peaking) | −2 to −4 dB |
| Thin, high "tss" sizzle | 12 kHz (high shelf) | −1 to −3 dB |
| Both | 6 kHz and 12 kHz | Small cuts on each, not one big one |
The trade-off: the 6 kHz band has a fixed, fairly broad width (a Q of 1), so a cut there also lowers the clarity and presence around it, all the time, not just on the esses. That's why small cuts work better than big ones. For the rest of the voice, How to EQ a Vocal in Your Browser covers all six bands.
On a parametric audio equalizer: boost and sweep
On a parametric audio equalizer you can be more precise. Set one band to a narrow Q, boost it hard, sweep slowly through the upper range until the esses jump out most painfully, then turn that same narrow band into a cut. Because it's narrow, less of the surrounding clarity goes with it. Vocal Cut's EQ has fixed frequencies and no Q control, so step 3 is the graphic-EQ version of the same idea: you compare two fixed regions instead of sweeping. Parametric vs Graphic EQ explains the difference, and what EQ is covers the basics.
EQ cut vs a real de-esser
A de-esser is a compressor that only listens to the sibilant band. When an "s" pushes that band over a threshold, it turns it down for that moment, then lets go. Between esses it does nothing, so the vocal keeps its clarity.
| Static EQ cut | Dynamic de-esser | |
|---|---|---|
| When it acts | All the time | Only when an "s" crosses the threshold |
| Effect on clarity | Lowers the band everywhere | Leaves the rest of the vocal alone |
| Best for | Mild to moderate, consistent sibilance | Severe or uneven, spitty esses |
| On Vocal Cut | Yes, the Equalizer | No dedicated de-esser |
If a gentle EQ cut can't tame your esses without dulling the voice, use a de-esser in a DAW. Many DAWs include one, and free de-esser plugins are easy to find. Neither fixes a badly recorded take completely, but a de-esser has far more room before the voice suffers.
Where compression fits
Compression turns loud moments down and, with makeup gain, brings everything back up. It mostly reacts to the louder vowels, so the quieter esses get less gain reduction and then share in the makeup gain. A vocal that was fine before compression can come out hissy afterwards. So:
- Make your sibilance cut before compression, so the compressor isn't reacting to spiky esses.
- Check the esses again after compressing, and nudge the EQ cut by a dB if they've crept back.
Vocal Cut's Compressor is a single full-band compressor (Threshold, Ratio, Attack, Release, Knee and Makeup gain, with Vocal, Master and Aggressive presets). It has no sidechain filter or frequency-selective mode, so it isn't a de-esser. Its Vocal preset (−24 dB threshold, 3:1, 3 ms attack, 250 ms release, 30 dB knee, +4 dB makeup) is a reasonable next step after your EQ. For the fundamentals, see how to use a compressor and compressor attack and release settings.
Don't overdo it: the lisp problem
It's easy to go too far. Cut too much and "s" turns soft and slushy, "speech" starts to sound like "thpeech", and the voice sounds as if it's behind a blanket. Words get harder to understand, which defeats the point.
Aim to bring the esses level with the rest of the voice, not to make them vanish. If the sting is gone but the voice is dull, reduce the cut until some bite returns. A slightly bright "s" is far less noticeable than a lisp.
Prevent sibilance at the mic
The cheapest de-essing happens before you press record:
- Angle the mic slightly off-axis, a little to the side of or above your mouth rather than straight at it. Aiming the capsule just past the stream of air from your teeth softens the esses.
- Keep some distance. Very close placement can exaggerate both esses and plosives; a few extra centimetres often helps.
- Skip big high-frequency boosts on the way in. Add brightness later, once you can hear what it does to the esses.
Mic Placement for Clean Voice Recordings shows common setups, and how to record clean vocals at home covers the full routine.
Troubleshooting
The esses are still harsh after a 4 dB cut. Check whether the problem is at 12 kHz rather than 6 kHz, and try small cuts on both. If it's still harsh, the sibilance is too uneven for a static cut and a dynamic de-esser is the better tool.
The vocal sounds dull or lispy. You've cut too much. Bring the band back toward 0 a dB at a time until the consonants are crisp.
The noise is there even between words. That's hiss, not sibilance, and the Noise Reducer is the right tool.
Related
Once the esses are under control, the free online compressor evens out the level, and How to EQ a Vocal in Your Browser shapes the rest of the tone. More step-by-step guides are in the audio how-to guides.
Frequently asked questions
What frequency is sibilance?
Sibilance usually sits somewhere between about 4 and 10 kHz, most often around 5–8 kHz, though it varies with the voice, the microphone and how close the singer or speaker was. Higher and lighter voices often sit toward the top of that range or above it. Because the exact spot differs between recordings, find it by ear: briefly boost a high band to exaggerate the esses, note where they sound worst, then cut there instead.
Can an EQ replace a de-esser?
For mild to moderate sibilance, often yes. A gentle EQ cut of a few dB around the sibilant band softens harsh esses. The limitation is that an EQ cut is always active, so it also reduces clarity when nobody is saying "s". A de-esser only turns the band down when a sibilant crosses its threshold, which makes it the better tool for severe or uneven sibilance.
Why does my vocal sound lispy after de-essing?
A lispy, dull vocal means the sibilant band was cut too much. The "s" sounds lose their definition and start sounding like "th", and the voice can feel muffled. Reduce the cut a decibel at a time until the consonants are crisp again. The goal is to bring the esses level with the rest of the voice, not to remove them completely.
Does compression make sibilance worse?
It often can. Compression turns the loud parts of a vocal down, and makeup gain lifts everything back up. Because the esses usually get less gain reduction than the louder vowels, they end up louder relative to the rest of the voice. Make your sibilance EQ cut before compressing, then listen to the esses again afterwards and adjust if they've crept back.