How to Use a De-Esser

To use a de-esser, insert it on the vocal track, set its detection band over the frequency range where the harsh “s” and “sh” sounds live (usually 5–10 kHz), and lower the threshold until the de-esser only clamps down on those sibilant peaks. That is the whole job: knowing how to use a de-esser well is mostly about targeting the right frequencies and reducing gain only when sibilance actually occurs, not all the time.

Sibilance is the sharp, hissy energy on consonants like S, T, Z, and SH. A little is natural; too much is fatiguing and gets worse after compression and EQ boosts. A de-esser is a frequency-specific compressor built to fix exactly that.

What a de-esser actually does

A de-esser is a dynamic processor that only reacts to a narrow band of high frequencies. When energy in that band crosses your threshold, it turns the gain down for a few milliseconds, then lets go. Because it only acts on sibilant moments, the rest of the vocal stays untouched. Think of it as a compressor that listens to one frequency range and ignores everything else.

Where to put it in your chain

Order matters. A solid default is:

1. Corrective EQ (remove rumble, fix obvious problems)
2. De-esser
3. Compression
4. Tone-shaping EQ and effects

Placing the de-esser before heavy compression stops the compressor from pushing sibilance up further. If you brighten the vocal with a high-shelf boost later, you may also want a second light de-esser after that boost. For more on how these processors interact, see our guide to EQ and compression fundamentals.

Step-by-step: dialling it in

1. Insert the de-esser on the vocal track or bus.
2. Find the sibilant band. Most de-essers have a “listen” or “solo” mode that lets you hear only the detected band. Sweep the centre frequency until the S sounds jump out clearly — typically 5–8 kHz for male voices and 6–10 kHz for female or brighter voices.
3. Set the band width. Narrow is more surgical; slightly wider catches a fuller range of sibilance. Start moderately wide and tighten if it sounds unnatural.
4. Lower the threshold until the gain-reduction meter only moves on the harsh S and SH sounds. The vowels and body of the words should show little or no reduction.
5. Aim for 2–6 dB of reduction on the loudest sibilants. More than that often produces a lispy, dull vocal.
6. A/B with bypass. Toggle it on and off. The sibilance should soften, but you should not be able to tell the de-esser is “doing” anything obvious.

Wideband vs split-band modes

Many de-essers offer two modes. Split-band only turns down the sibilant frequencies, leaving the rest of the vocal full — this is usually the most transparent choice. Wideband turns down the whole signal briefly when sibilance is detected, which can sound more “analog” but may dull the vocal if pushed. Start with split-band for clean, modern vocals.

DAW-specific tips

GarageBand: there is no dedicated de-esser, but you can use the built-in Channel EQ to gently dip the harsh band, or stack the Compressor with a narrow EQ in front of it. For finer control, add a free de-esser plugin.

Audacity: Audacity has no real-time de-esser, so the practical approach is manual. Use the Spectrogram view or your ears to find loud S sounds, select each one, and apply a small gain reduction, or use a narrow EQ notch on those selections.

FL Studio: the stock Fruity Limiter has a compressor mode, but the cleaner option is the included Maximus or a dedicated plugin set to a sibilant band. Many producers reach for free de-essers like TDR Nova (used in dynamic-EQ mode) here.

Capable free and paid de-essers include TDR Nova (a dynamic EQ that doubles as a precise de-esser), FabFilter Pro-DS, and the Waves Sibilance and DeEsser plugins.

Common mistakes

• Over-de-essing. If the singer starts to lisp, back off the threshold or reduce the maximum reduction.
• Targeting the wrong band. If the vocal goes dull but the S still cuts through, your detection frequency is too low.
• De-essing before fixing the source. Good microphone placement and the right mic reduce sibilance before you ever open a plugin.
• One static setting for the whole song. Sibilance changes with delivery; on tricky vocals, automate the threshold or use a second instance on the brightest sections.

De-essing is a small step that makes a big difference once vocals are loud in a mix. For the bigger picture, work through our full walkthrough on how to mix vocals, and browse more techniques in the mixing and mastering hub.

Frequently asked questions

What frequency should a de-esser be set to?

Most sibilance lives between 5 kHz and 10 kHz. Use the de-esser’s listen mode to sweep until the S sounds are loudest, then set the detection band there. Lower voices often sit around 5–7 kHz, brighter or female voices around 7–10 kHz.

Should I de-ess before or after compression?

Usually before your main compressor, so the compressor does not push sibilance up further. If you add a bright high-shelf boost after compression, a second light de-esser after that boost can help.

Can I de-ess without a plugin?

Yes. You can manually reduce the gain on individual S sounds with clip gain or volume automation, or apply a narrow EQ notch to just the sibilant selections. It is slower than a de-esser but works in any DAW, including Audacity.