De Esser for Beginners: A Complete Guide

By Katie Anderson 25 November 2025 5 min read

Filed under

Add Music Adviser UK to your preferred sources in Google

De Esser for Beginners: A Complete Guide

A de-esser is a dynamics processor that targets sibilance — the harsh, hissing quality on consonants such as "s", "sh" and "ch" — in vocal recordings. It applies gain reduction only when sibilant energy crosses a threshold in a specified frequency band, leaving the rest of the vocal untouched. Understanding how to configure one correctly is straightforward once the underlying principle is clear.

Why Vocals Need a De-Esser

Close-mic setups, large-diaphragm condenser microphones and digital recording all capture high-frequency detail accurately — including sibilance. Singers whose voices naturally emphasise these frequencies, and recordings made in acoustically reflective spaces, can produce sibilance that is distracting or fatiguing over longer listening periods. Compression applied during tracking or mixing can also increase the prominence of sibilants by reducing the level of the surrounding material, making harsh consonants more audible by comparison.

Detection Frequency and Threshold

Two controls determine how the de-esser behaves: the detection frequency, which sets which part of the frequency spectrum is monitored, and the threshold, which determines how loud the sibilant signal must be before reduction is applied. The detection frequency is usually found by sweeping — start around 6–8 kHz and move up or down until the processing catches most sibilants without affecting other parts of the vocal. The threshold should be set high enough that reduction only kicks in on genuine sibilants, not on every voiced consonant.

Checking the Result

After setting up the de-esser, listen through a section of the vocal that contains several sibilant moments and bypass the processor repeatedly to compare the processed and unprocessed signal. The processed version should sound cleaner without sounding dull or lispy. A gain reduction of 3–6 dB on sibilant peaks is typically sufficient; more than 10 dB of continuous reduction usually indicates the threshold is set too low. For an overview of the available products, the buyer's guide covers the options at each price point.

Frequently Asked Questions

Is a de-esser necessary for every vocal recording?

Not always. Vocals recorded with a warm, dark microphone, at a greater distance from the capsule or in acoustically treated rooms may not exhibit significant sibilance. Add processing only when sibilance is audible and distracting — applying it unnecessarily can dull a vocal's presence and brightness.

Where should it sit in the signal chain?

Most engineers place it after equalisation and before or after compression on the vocal channel. Placing it before compression prevents the compressor from reacting to sibilant peaks. Placing it after compression catches any sibilance that compression has brought forward. Both positions are valid; the correct one depends on how the specific vocal responds.

What is the difference between wideband and split mode?

Wideband mode applies gain reduction to the entire signal when sibilance is detected. Split mode applies reduction only to the targeted frequency band, which produces a more transparent result in most cases. Split mode is generally preferred for vocal work.

How much gain reduction is too much?

If the gain reduction meter is hitting 10 dB or more continuously rather than only on sibilant peaks, the threshold is set too low. Excessive processing produces a lispy quality and removes natural presence from the vocal. Aim for 3–6 dB of peak reduction on genuine sibilants only.

Can I automate it instead of using a processor?

Clip gain or volume automation on sibilant consonants can reduce their level before the signal reaches any dynamics processor. This approach is more labour-intensive but can produce very clean results when the recording quality warrants the additional editing time.