De-esser settings explained (threshold, frequency, range)

A de-esser looks simple until you open one and see threshold, frequency, range, and a couple of mode buttons, and you're not sure which does what. The controls are actually straightforward once you know what each one is for — and getting them right is the difference between a vocal that's smooth and one that lisps. Here's every de-esser setting explained, and how to set each one.

Key Takeaways

• Frequency — where the de-esser listens; aim it at the sibilant band (5–10 kHz).
• Threshold — how loud an ess has to be before the de-esser acts.
• Range / amount — how much it turns the ess down once it does.
• Listen / solo mode — lets you hear exactly what it's reacting to, to set frequency.

Frequency (or the split point)

The most important setting: where the de-esser listens. Sibilance lives roughly between 5 and 10 kHz, and you point the frequency control at the spot where a given voice's esses concentrate — lower for deeper voices, higher for brighter ones. Set this wrong and the de-esser is watching the wrong part of the spectrum, so nothing else you do will work. Use the listen mode (below) to find it, rather than guessing a number.

Threshold

How loud an ess has to be before the de-esser kicks in. Lower the threshold and the de-esser catches more esses (including quieter ones); raise it and it only acts on the loudest, harshest ones. The goal is to set it so the harsh esses trigger it and the normal vowels don't. If the whole vocal is being ducked, your threshold is too low; if the harsh esses are still getting through, it's too high.

Range (or amount / reduction)

How much the de-esser turns the ess down once the threshold is crossed. This is the amount of gain reduction. A few decibels on the peaks is typical. Set it too high and you get a lisp — the "s" softens toward a "th." The rule: set the range on the harshest word until it stops biting, then stop. Don't dial it until every ess is gone.

Listen / solo mode

Not a processing control, but the one that makes the others easy. Listen mode (sometimes "solo sibilance" or a monitor button) lets you hear only what the de-esser is reacting to. Engage it, play the vocal, and sweep the frequency until the esses are loudest and clearest in the solo — that's your target frequency. Then turn listen mode off and set threshold and range. It turns setting a de-esser from guesswork into something you can hear.

Attack and release (if present)

Simpler de-essers hide these; some expose them. Attack is how fast the de-esser clamps down on an ess — you want it fast, so it catches the onset. Release is how quickly it lets go — quick, so it's off before the next syllable and doesn't dull what follows. Most de-essers set these sensibly by default; if you have them, fast attack and fast release is the safe starting point for sibilance.

Modes: wideband vs split-band

Many de-essers offer two modes. Wideband turns down the whole signal briefly when an ess hits (can dull the voice slightly). Split-band turns down only the sibilant band and leaves everything else at full level — usually the more transparent choice for vocals. If your de-esser offers it, split-band is the better default. (The deeper version of this — following the ess rather than a fixed band — is what phoneme-aware detection adds.)

Putting it together

A reliable order every time: use listen mode to set frequency, lower the threshold until the harsh esses trigger it, raise the range on the worst word until it stops biting, and confirm the vowels stay open. That's a well-set de-esser.

This is also where the tool matters. A transparent, phoneme-aware de-esser like Sibilance needs less fiddling because it classifies the consonant rather than reacting to level — it finds and follows the ess for you, so the settings do less work and the vowels stay intact. There's a free tier to learn the controls on your own vocal.

The short version

Frequency sets where the de-esser listens (aim at the sibilant band, found with listen mode), threshold sets how loud an ess must be to trigger it, and range sets how much it reduces. Fast attack and release, split-band mode where offered. Set frequency first with listen mode, then threshold, then range on the harshest word — and stop before the vocal lisps.

Sibilance is the single-authority de-esser this comes from — a real free tier, and a 7-day Pro trial with no card.

Related reading