Sibilance is the sharp "s", "sh", "ch" and "t" energy in a vocal. A little is natural and helps clarity. Too much hurts on headphones and gets worse with every compressor and brightening EQ you add. A de-esser turns down just that energy, just when it happens.
- Sibilance usually sits at 5–10 kHz
- Lower the threshold until the s sounds natural, not lisped
- Aim for a few dB of reduction on the loudest s sounds
- Before compression, after compression, or both
- Listen in the mix, not solo
How a de-esser works
A de-esser is a compressor that listens to a narrow, high frequency range. When that range gets louder than the threshold (an "s" sound), it turns it down, either just that band (split-band) or the whole signal (wideband). Between sibilants it does nothing, so the vocal keeps its brightness. That's why it beats a static EQ cut, which dulls everything. The EQ side is in the vocal EQ cheat sheet.
Find the sibilance frequency
- Loop a line with obvious "s" sounds.
- Use the de-esser's listen / monitor mode (or a narrow EQ boost) and sweep between 4 and 12 kHz.
- Stop where the "s" is sharpest and most painful. Lower voices often sit around 5–7 kHz, higher voices 7–10 kHz.
De-esser settings
- Frequency: the spot you just found.
- Threshold: lower it until the loudest sibilants are reduced and normal words aren't touched.
- Range / reduction: a few dB on the loudest "s" is usually enough. Much more starts to sound like a lisp.
- Split-band vs wideband: split-band is the usual choice for vocals; wideband can sound more natural on gentle sources.
Where the de-esser goes in the chain
- Before compression if loud "s" sounds make the compressor duck the whole vocal. Common on bright rap takes.
- After compression and brightening EQ, because both raise sibilance. Common on smooth sung vocals.
- Both: two gentle de-essers usually beat one heavy one.
You can see both approaches in real chains: our Crispy Rap chain de-esses right after pitch correction, while Pop Lead Vocal de-esses near the end. Full breakdown in vocal chain order explained.
Signs you've gone too far
- The singer sounds like they have a lisp.
- The vocal sounds dull or loses its air.
- Words become harder to understand.
If that happens, raise the threshold or reduce the range. Also check the reverb and delay: if they're bright, they can splash sibilance around even when the dry vocal is fine. De-ess the effect send, or roll off the effect's highs.
Fix sibilance at the source
- Angle the mic slightly off-axis from your mouth.
- Some bright condenser mics are very sibilant on some voices; a different mic can help more than any plugin.
- For the worst few words, lower the clip volume of just the "s" by hand.
FAQ
What frequency should I de-ess vocals at?
Usually between 5 and 10 kHz. Find the exact spot by sweeping in the de-esser's listen mode.
Should the de-esser go before or after compression?
Before if sibilance triggers the compressor, after if compression and EQ make it louder. Many chains use one in each place.
Can I de-ess with an EQ?
A dynamic EQ band works well. A static EQ cut also removes brightness from every other word, so it's a last resort.
