Mixing and Plugins

What Is a De-Esser and How Do You Use One?

What a de-esser does, how to set its threshold and frequency, where to place it in the chain, and free options. Tame harsh s and sh sounds without a lisp.

On this page
  1. What is sibilance?
  2. Wideband vs split-band
  3. How to use a de-esser
  4. Where it goes in the chain
  5. Manual alternative
  6. Free and built-in options
  7. Beyond vocals
  8. Fix it at the source
  9. Common mistakes

A de-esser is a compressor that reacts only to sibilant frequencies, the hissy “s”, “sh”, “t” and “ch” sounds. It turns those frequencies down only when they get too loud, so the rest of the vocal stays untouched.

What is sibilance?

Sibilance is the sharp hiss that comes with sibilant consonants. It usually sits in the roughly 4 to 10 kHz range, and the exact spot depends on the voice and the microphone. For many male voices it’s around 4 to 7 kHz, and for many female voices around 6 to 9 kHz.

Left alone, sibilance can be fatiguing to listen to, and it can make a vocal stand out from the rest of the mix. Natural esses should stay audible, though. The goal is to tame the harshness, not remove the sound.

Wideband vs split-band

  • Wideband de-essing turns down the whole signal when a sibilant hits. It’s simple, but the rest of the vocal can dip with it.
  • Split-band de-essing reduces only the sibilant band. It’s usually more transparent, so it’s the better choice when the vocal needs to keep its character.

How to use a de-esser

Most de-essers have four core controls:

  • Threshold: the level where the de-esser starts to act.
  • Frequency: the centre of the sibilant range you want to target.
  • Bandwidth: how wide the affected band is. A narrow band targets only the sibilant frequencies. A wide band can also catch other high-frequency content.
  • Reduction (often called Range): how much the sibilant band is turned down.

Many de-essers also offer attack and release controls, and a listen (or sidechain monitor) mode that lets you hear only the frequencies the de-esser is reacting to.

Here’s a practical sequence:

  1. Start with low reduction and a conservative threshold. It’s easier to add processing later than to undo an over-processed vocal.
  2. Turn on listen mode and sweep the frequency until you hear the esses isolated. Note the harshest frequency.
  3. Set the threshold so the de-esser triggers on the esses and not on normal vowels.
  4. Aim for moderate gain reduction, about 3 to 6 dB on the sibilants.
  5. A/B against the original, and check for artifacts like a dull or lisping sound.
  6. Judge in the mix. A vocal that sounds too sharp soloed can sit well in the full mix, and the reverse is also true.

Where it goes in the chain

Place the de-esser after compression. Compression raises the level of quiet sibilants relative to the rest of the vocal, so a de-esser placed after the compressor catches what the compressor brought up.

Two gentle de-essers, one before and one after compression, can work better than one heavy de-esser. Also place the de-esser after any big high-frequency EQ boost, because that boost can bring out sibilance.

For more on compression settings for vocals, see vocal compression for home recording.

Manual alternative

If only a few esses are a problem, you can handle them by hand. Use clip gain or automation to turn down individual esses by about 3 to 6 dB. This takes longer, but it’s precise and it doesn’t affect the rest of the vocal.

Free and built-in options

  • Techivation T-De-Esser: a free de-esser.
  • TDR Nova: a free parametric EQ with a dynamic band, which can be used as a de-esser.
  • Stock DAW de-essers: most DAWs include one, and it’s a good starting point before you spend money.

For more free options, see the best free compressor VST plugins for home mixing.

Beyond vocals

A de-esser isn’t only for vocals. It’s useful on harsh hi-hats, the squeak of an acoustic guitar’s strings, and reverb sends that brighten sibilance. The same rules apply: find the frequency, set a conservative threshold and check in context.

Fix it at the source

Some sibilance comes from the recording. Angling the microphone slightly off-axis from the mouth can reduce it before it reaches the computer. A pop filter is a different tool: it reduces plosives, the “p” and “b” bursts, not sibilance. For plosives already in a recording, see how to remove clicks and pops.

Common mistakes

  • Over-de-essing. If the esses are gone, the vocal can sound lispy. Keep them audible.
  • Using a static EQ cut instead. A fixed EQ cut dulls the voice all the time. A de-esser acts only when the sibilance spikes.
  • Judging only in solo. The vocal has to work with the rest of the mix.

For a broader look at the processing that sits around a vocal, see home mixing basics.

The complete guide to mixing and plugins

Home Mixing Basics and Simple Mixer Software
Read the guide

Keep exploring

Guides from other topics

All topics →