De-Essing: Taming Harsh Sibilance Without a Lisp

A de-esser tames the sharp S and T sounds that make a voice pierce on headphones. It works like a compressor that only reacts to a chosen frequency band, pulling those moments down when they get too loud. Used lightly it is invisible, and used carelessly it leaves the speaker lisping.

Sibilance lives mostly between 5 and 9 kHz, though it varies by voice and microphone. Finding the exact band on your own recording is half the job, and the plugin presets rarely find it for you.

Find the band before you set anything

Sweep a narrow boost across the high end while a harsh line plays, and note where the unpleasant edge peaks. That is your detection range. Set the de-esser to listen there and to act only when the energy crosses a threshold you choose.

Then switch the plugin to a narrow band and listen closely. The goal is to shave the spike, not to dull the whole top end. If the air around the voice disappears, you are processing too wide a range.

A good de-esser is boring. You notice the harsh edges are gone, and you notice nothing else at all.

Static or dynamic

Static de-essing applies a fixed cut to the whole take, which is fast but often audibly dull. Dynamic de-essing triggers only when the harsh sound appears, which preserves brightness on the rest of the words. I reach for dynamic almost every time.

Watch the gain reduction meter while the file plays. If it sits pinned for most of the take, your threshold is too low. The needle should flicker on the hot S sounds and rest between them.

Do not ignore the cause

Harsh sibilance is often a recording problem. An off-axis mic, a pop filter too far away, or a hard reflective wall behind the speaker all push the S sounds forward. A de-esser hides that, and a small change in mic placement might remove it.

I fix placement and still keep a light de-esser on the channel. The two are partners rather than rivals, and together they beat either one pushed to its limit.

Key takeaways

• Find the harsh band by ear on each voice instead of trusting a preset.
• Use dynamic de-essing to keep brightness where the S sounds are not.
• Keep gain reduction small, three to five decibels on the loudest hits.
• Check mic angle and room reflections before you push the plugin harder.