What Is a De-Esser? How to Tame Harsh S Sounds (Settings Included)

The short answer: a de-esser is a compressor that only listens to a narrow high-frequency band, usually somewhere between 4 and 10 kHz, where harsh S, T, SH, and CH sounds (sibilance) live. When a sibilant sound crosses the threshold, it turns that moment down, and leaves the rest of the vocal alone. It's the difference between a vocal that sounds bright and one that stabs your ears on every "s". Aim for about 2 to 4 dB of reduction on the esses only. More than that and the singer starts to lisp.

Every vocal on my releases goes through at least one de-esser, often two light ones. Sibilance is the thing that makes a mix sound cheap on earbuds faster than anything else.

How it works

Where it goes in the chain

After compression, almost always. Compressors bring quiet details up, and sibilance is one of the details they bring up. De-ess before the compressor and the compressor undoes your work. On a hard-hit vocal, two light passes (one before, one after the compressor) sound more natural than one heavy one. The full vocal chain order is here, and bright EQ boosts belong before the de-esser for the same reason: boost the air, then catch what the boost exaggerated.

How to tell you've gone too far

• The lisp: "sister" becomes "thither." You've cut too deep, or the frequency is set too low and it's catching the body of the voice.
• A dull vocal: you fixed the esses and lost the air. Narrow the band or reduce the range.
• Pumping brightness: the top end ducks audibly on every consonant. Faster release, less reduction.

The best check: listen on earbuds and a phone speaker at normal volume. Sibilance that hides on studio monitors shows up instantly there.

The tools

Logic ships DeEsser 2; Ableton has no dedicated one, but its Compressor with the sidechain EQ set to the sibilance band does the same job; FL users typically reach for Maximus or a free third-party de-esser. For free, TDR Nova as a dynamic EQ is one of the best de-essers you can get, and it's the pick in the best vocal plugins list. Paid options like FabFilter Pro-DS buy transparency on heavy workloads, covered in the FabFilter alternatives breakdown.

One honest limit: a de-esser can't fix a sibilance problem baked in at the recording. A bright condenser mic pointed straight at the mouth gives you esses no plugin will tame cleanly. Angling the mic slightly off-axis fixes more than any de-esser, covered in the home recording guide.

Quick answers

What does a de-esser do? It automatically turns down harsh S, T, and SH sounds in a vocal the moment they happen, leaving the rest of the performance untouched.

What frequency should I de-ess at? Usually 5 to 8 kHz. Sweep with the listen function to find where the esses are loudest on that specific voice, since it varies.

Should the de-esser go before or after compression? After, in most cases, because compression brings sibilance up. On heavy vocals, a light pass on both sides works best.

Can I de-ess without a de-esser? Yes. A dynamic EQ band (TDR Nova is free) or manually clip-gaining each S down both work, and the manual method is the most transparent of all.