Every vocal has sibilance. The “S” sounds, the “SH” sounds, the hard “T” at the start of a word. On most recordings they sit quietly in the background. But add a condenser mic, some compression, and a bit of EQ brightness, and those same consonants start cutting through the mix like they’re angry about something.
A de-esser is what you reach for when that happens. This guide covers what a de-esser is, how to find the right frequency for your specific vocal, and how to set one up without turning your singer into a lisping mess.
What Is Sibilance and Why Does It Get Worse?
Sibilance is the harsh, hissy quality that comes from consonant sounds like “S”, “SH”, “T”, “Z”, and sometimes “F”. It’s a normal part of human speech, but it gets amplified in the recording chain in ways that make it unpleasant in a mix.
Condenser microphones, especially ones with a presence boost in the high-mids, pick up these frequencies more aggressively than ribbons or dynamics. Close-mic placement emphasizes them further.
And then compression does the rest: when you compress a vocal, you’re raising the quieter parts of the signal, which includes all the consonants that weren’t loud enough to cause problems before. The result is a vocal that might sound fine in isolation but turns into a hiss-fest on headphones, earbuds, or phone speakers. And the more you’ve processed the vocal, the worse it tends to get.
Sibilance doesn’t live at the same frequency on every voice. A bright female vocal through a large-diaphragm condenser might peak at 7-8kHz. A darker male vocal might be causing problems at 4-5kHz. Before you open a de-esser plugin, it helps to know where you’re actually going to aim it. Use the finder below to get a starting point based on your source.
What a De-Esser Actually Does
A de-esser is a frequency-sensitive compressor. Instead of responding to the overall level of the signal, it listens to a specific frequency range and when the energy in that range exceeds the threshold, it compresses. Everything else in the signal stays untouched.
The simplest way to think about it: a very fast, very targeted compressor sitting on the sibilance frequencies only, leaving everything else alone.
I used a short AI-generated vocal demo to show the difference. The raw recording includes basic compression and EQ, nothing else.
Dry vocal
Raw vocal recording, no processing. Listen for the S sounds.
Same vocal with Pro-DS
FabFilter Pro-DS, Single Vocal mode. Detection at 5.815kHz. The S sounds are still there but no longer cut through.
Wideband vs Split-Band: Two Different Ways to Catch Sibilance
Most de-essers give you two operating modes and picking the right one makes a real difference to how the result sounds.
Wideband
Reduces the entire signal when sibilance is detected. Think of it like a compressor that ducks the full vocal whenever an S triggers it.
The result is smooth and natural because the tonal balance of the signal doesn’t change. Everything just gets quieter for a moment.
On a single vocal track, wideband is almost always the better starting point.
Split-Band
Only attenuates the frequencies above the detection point. The low-mids and body of the vocal stay untouched. Only the high frequency content gets compressed.
This can be more transparent on problematic material, but it can also create a momentary tonal shift if pushed too hard.
The voice suddenly loses its top end right on an S, then gets it back.
I almost always start in wideband on vocals. It’s smoother and easier to dial in without artifacts. If I’m running a de-esser on a full mix bus or a vocal stem with multiple voices, I switch to split-band so the body of the mix doesn’t duck every time a singer lands on an S.
Threshold, Frequency, Range: What Each One Actually Controls
Most people load a de-esser and go straight for the threshold. That’s the wrong order. The frequency control is the most important thing on the plugin and you should set it first. Here’s what each control does and why the sequence matters.
Setting Up a De-Esser: Step by Step
Insert the de-esser after compression
Compression makes sibilance worse by raising quiet consonants along with everything else. Insert the de-esser after your compressor so it’s dealing with the signal at its most compressed and most sibilant state.
Use Audition mode to find the frequency
Every good de-esser has an Audition or Monitor button that solos the detection range so you can hear exactly what the plugin is listening to. Engage it, sweep the frequency control until you hear the harshest S sounds in isolation, then disengage. You’ll find the right frequency in about 10 seconds instead of guessing.
Set the threshold
Lower the threshold until you see gain reduction on the sibilant syllables. Then raise it back until the de-esser only triggers on the moments that actually sound harsh. The gain reduction meter should flicker on S sounds and stay quiet the rest of the time.
Set the range
Start at 3-6dB. That’s enough to handle most sibilance without sounding processed. Go higher only if the problem is severe, and always check for lisping on softer consonants when you push past 8-10dB.
Check in context
Play the vocal alongside the full track. Listen for whether the S sounds still cut through, and whether the de-esser is audibly ducking the vocal on every consonant. Neither is right. Adjust until the sibilance is controlled but the vocal still feels present and natural.
Signal Chain Placement: Before or After Compression?
There’s no single right answer here, but there are strong preferences.
- After compression is the most common placement. Compression raises quiet consonants along with everything else, which makes sibilance worse. A de-esser right after catches the problem at its most compressed and most obvious state.
- After EQ makes sense when you’ve added brightness or presence. You boosted the top end, now you’re reining in the harshest part of that boost. Clean and logical.
- Before everything works when the raw recording is severely sibilant. Catching it early means every plugin downstream is working with a cleaner signal.
- Stacking two is something I reach for on difficult vocals. A gentle wideband de-esser early in the chain, then a more precise split-band after compression to handle whatever the compressor made worse. It sounds like overkill but usually results in something more natural than trying to fix everything with one plugin set aggressively.
De-Essing Isn’t Just for Vocals
The vocal is the obvious use case. But harsh high-frequency transients show up on other sources too, and the fix is the same.
Hi-hats and cymbals
In dense mixes, cymbal harshness can sit in the same frequency range as vocal sibilance and cause similar problems. A de-esser placed on the drum bus, tuned to where the cymbal edge lives, can smooth the top end without dulling the attack of the hits. More transparent than a static high shelf cut.
Acoustic guitar
Recorded with a bright condenser, acoustic guitar sometimes picks up string squeaks and pick attacks that read similarly to vocal sibilance. A de-esser set to a higher frequency range handles this more transparently than a static EQ cut, because it only reduces the level when the problem actually occurs.
Mix bus
A broad sibilance issue across a full mix is hard to fix at the source but responds well to a split-band de-esser on the output. Keep the range subtle at 2-3dB. This is also where split-band mode makes the most sense: you want the body of the mix untouched while only the harshness gets tamed.
Every de-esser worth using has an Audition or Monitor button that lets you hear only the frequencies it’s detecting. I always use it before setting the frequency. Solo the detection signal, sweep until you hear the harshest S sounds in isolation, then set it there. You find the right frequency in about 10 seconds instead of guessing.
When the De-Esser Makes Things Worse
De-Esser FAQs
Should the de-esser be audible?
No. If you can clearly hear the de-esser engaging, the range is too high or the threshold is too low. Good de-essing is something you notice only when you bypass it.
Do I need a de-esser on every vocal?
No. Some vocalists and some microphone combinations don't produce problematic sibilance. Use it when there's a problem, not as a default step in every vocal chain.
Does the de-esser affect the whole mix?
Only if you put it on the mix bus. On an individual vocal track it only processes that track.
Can I use a de-esser on instruments?
Yes. Hi-hats, cymbals, acoustic guitar, and some synth leads respond well to de-essing when they're causing harshness in a mix.
What's the difference between a de-esser and a multiband compressor?
A multiband compressor applies ongoing dynamic control across multiple frequency bands simultaneously. A de-esser is a sidechain-triggered compressor focused on a narrow band, built to react quickly to short transient events like S sounds. They overlap in concept but are built for different jobs.
