You know the moment. The lyric is perfect, the take is right, and then the S jumps out like a needle.
That bite is sibilance. It is the high frequency energy in consonants like s, sh, ch, t, and f. It is part of real speech, and it is also the fastest way to make a vocal hurt at loud volume.
Too much sibilance is harsh and fatiguing. Too little makes the words blur and the performance feel dull. The goal is not to erase it. The goal is just enough.
Just enough sibilance keeps the lyric alive. Too much makes it hurt.
The de-esser is the classic fix
A de-esser is basically a compressor that listens to a specific frequency range and turns the signal down when that range crosses a threshold. You can run it wide band or split band.
Wide band turns down the whole signal for the moment of the sibilance. Split band only turns down the problem range you set in the plugin. I prefer wide band most of the time because it sounds more natural to me. I find the sibilant frequency, then I just want that whole moment to tuck down together.
Try both. Your voice, mic, and song will tell you which one feels right.
Multi band compression can do the same job
A de-esser is a specialized multi band compressor. You can build the same behavior with a multi band compressor by setting a band in the sibilance area and letting it gain reduce when it crosses a threshold. I often aim somewhere around 4 kHz to 10 kHz, then tune by ear.
I usually reach for a de-esser first because it is fast, but it helps to know how the tool actually works.
Dynamic EQ is a clean, precise alternative
Dynamic EQs like Pro-Q 4 let you add a dynamic band right on the sibilant range. It behaves a lot like a split band de-esser. You decide where it listens. It ducks only that area when the energy pops.
This is great when the sibilance is narrow and you want to keep the rest of the top end open.
Resonance suppression is smart and smooth
Tools like Soothe 2 are a different flavor. You tell it the area you want to control, how hard you want it to work, and the plugin hunts down resonant peaks in real time. It feels like a dynamic EQ with a brain deciding which little spikes to tame.
I do not pretend to know the magic under the hood. I just know it can sound very natural when the vocal gets sharp.
Manual volume rides are still the gold standard
Many engineers will tell you the cleanest de-essing is manual. You zoom in, find the sibilant moments, and clip gain them down. That is the most precise control you can get.
Melodyne can help you see those regions too, which speeds things up. I often do a few manual rides on the worst offenders and then use a de-esser to catch the rest.
Try your de-esser before and after compression
Compression can bring sibilance forward, so placement matters. I often run cleanup EQ first, then de-essing, then compression, and sometimes another de-esser after the compressor if it is still poking out. But there is no rule here. Try it before. Try it after. Try both. Let the track decide.
Split the S from the tone
There are plugins that separate the sibilance from the rest of the vocal and send each to its own fader. Apulsoft SplitS and Wavesfactory Re-Esser do this well. You can then turn the sibilance fader down like a simple volume control. It is fast, and it keeps your automation clean.
Background vocals need less S
Stacks of background vocals can turn into a spray of S sounds. It feels sharp and messy fast. I will de-ess background vocals hard, almost to the point of a lisp when soloed. In the mix, the lead provides the articulation and the backgrounds just carry the width.
If the chorus still feels harsh, that trick usually fixes it immediately.
Hear it in my music
If you want to hear how I keep vocals clear without the bite, follow me on Spotify. These mixes are where I test all of this in the real world.
Follow Braylen Hope on Spotify
Want more vocal chain notes
If you want more practical mixing moves like this, join the newsletter. I will send my best vocal chain notes, plus new posts the moment they drop.