How to De-Ess Vocals

To de-ess a vocal, target the harsh s and t spikes with a de-esser or dynamic EQ set to act only when sibilance crosses a threshold, usually somewhere between 5 and 9 kHz. The key idea is that sibilance is a dynamics problem, not a tone problem - you do not want to permanently cut the high end, you want to duck it only in the moments it gets harsh. Static EQ dulls the whole vocal; a de-esser keeps it bright and only tames the stings.

What causes harsh sibilance?

Sibilance is the burst of high-frequency energy on s, t, sh, ch and z sounds. It becomes harsh when it spikes louder than the rest of the vocal, which happens for a few reasons.

• Close mic technique exaggerates high-frequency detail, including sibilance.
• Bright microphones and presence EQ lift the exact range where sibilance lives.
• Compression and saturation push sibilance up along with everything else, so the more you process, the harsher it can get.

This is why de-essing usually comes after EQ and compression in the chain - those steps make sibilance louder, so you tame it once it has been brought up.

How do I set up a de-esser?

A de-esser is a fast, frequency-targeted compressor that only reacts to a specific high band. The setup is simple once you find the right frequency.

• Find the sibilant frequency. Most sibilance lives between 5 and 9 kHz. Sweep the de-esser's listen mode until the s sounds are loudest and clearest - that is your target.
• Set the threshold so the de-esser only engages on the harshest s and t sounds, not on every word. Watch the gain reduction meter - it should move only on sibilants, and only a few dB.
• Aim for 3 to 6 dB of reduction on the worst spikes. More than that and the vocal starts to lisp.
• Choose split-band mode if available, so only the sibilant band ducks and the rest of the vocal is untouched.

If you over-de-ess, the singer starts to sound like they have a lisp, with the s sounds swallowed. Back the threshold off until the sibilance is controlled but the words still sound natural.

When should I use dynamic EQ instead?

A dynamic EQ can be a more transparent de-esser, especially when sibilance sits in a narrow range or you want tighter control.

• Set a dynamic EQ band at the sibilant frequency with a fairly narrow Q.
• Set it to cut only when the band exceeds a threshold, so it behaves like a targeted de-esser.
• Use two bands if the vocal has both a lower sibilance around 5 to 6 kHz and a brighter sting around 8 to 9 kHz - address each separately.

Dynamic EQ shines because you can pinpoint the exact problem frequency and leave the rest of the high end fully intact. For the broader tonal picture, see how to EQ vocals.

How do I de-ess without dulling the vocal?

The whole point is to keep the vocal bright while removing the sting. A few habits keep it transparent.

• Only duck what is harsh. Set the threshold high enough that soft words pass untouched.
• Keep reduction modest - a few dB per spike. If you need heavy reduction everywhere, your EQ is probably too bright, so pull back the presence or air boost instead.
• De-ess in stages if one de-esser is working too hard. Two light de-essers, or a de-esser plus a dynamic EQ, share the load more naturally than one clamping down hard.
• Consider manual clip-gain on the worst single words. Pulling one brutal s down by hand is the most transparent fix of all.

If de-essing is fighting your brightness, the real fix might be upstream. See how to make vocals brighter for how to add air without piling on sibilance.

How do I de-ess without it sounding obvious?

A heavy-handed de-esser is easy to hear - the vocal suddenly loses its top on every s, then snaps back, which sounds like the tone is flickering. A few habits keep it invisible.

• Match the reduction to the problem. If only a handful of words are harsh, a light de-esser plus manual clip-gain on those words beats clamping the whole track.
• Check the s-only signal. Many de-essers let you monitor what is being removed - you should hear s and t sounds, not vowels or body. If you hear vowels, your frequency is too low or your Q too wide.
• Automate the de-esser threshold across sections if a loud chorus is far more sibilant than a soft verse. One fixed setting rarely suits both.
• Fix the room and mic first when tracking. A little off-axis mic technique reduces sibilance at the source so you need less de-essing later.

The best de-essing is the kind nobody notices - the vocal just stops stinging while keeping all its brightness and clarity.

Where does de-essing sit in the chain?

De-ess after EQ and compression, because those raise sibilance, and before or alongside the effects sends so you are not feeding harsh s sounds into reverb and delay tails. A clean order is EQ, compress, de-ess, then space. If you push a lot of presence for a bright vocal, expect to de-ess more, and vice versa. The full sequence is in how to make vocals sound professional.

If you would rather not chase thresholds by hand, Vocal Polish balances brightness and sibilance control in one pass, so the vocal comes out present without the sting. Explore more in the vocal guides.