De-Esser Plugins Guide: 5 Things You Must Know

What Does a De-Esser Actually Do?

De-essing is frequency-selective dynamic control. When a vocalist hits an "s" sound, energy spikes between 5kHz and 10kHz. A de-esser watches that frequency range, and the moment it exceeds a set threshold, it pulls down the level. The rest of the vocal stays untouched.

There are two main operating modes you'll see across every plugin in this category. Understanding them saves you from making the wrong choice on a session.

Split-Band Mode

Split-band (sometimes called "frequency-selective") mode only reduces the problematic high-frequency band. The rest of the signal passes through flat. It's the more surgical approach, and it's what most engineers reach for first.

The result sounds natural because you're only touching what's broken. If your vocalist has a lovely top-end air that you want to keep, split-band won't breathe down its neck.

Wideband Mode

Wideband mode turns down the entire signal whenever the sidechain detects a sibilant spike. Think of it as a limiter triggered by high frequencies. It sounds more like a vintage hardware de-esser because that's exactly how older hardware worked.

Wideband is counterintuitive on paper, but in practice it often sounds more transparent on dense mixes. The slight dip in the full signal reads less obviously to the ear than a notch pulled out of a specific band. We've found it works better on vocals that sit in a busy arrangement where the effect of split-band compression shows up as an unnatural thinning.

Where Does a De-Esser Go in Your Signal Chain?

Place your de-esser after your compressor and before your reverb or delay. This is the standard position, and there's a logical reason for it.

Your compressor increases gain during quieter moments, which means sibilance that was barely noticeable in the raw recording gets pushed forward in level. The de-esser catches that boosted sibilance before it hits your time-based effects.

If you put the de-esser before the compressor, you're solving a problem that doesn't fully exist yet. The compressor will create new sibilance problems downstream that you've left unaddressed.

What About Two De-Essers?

Yes, it's a legitimate approach. We've run sessions where the vocal needed a broadband de-esser early in the chain to handle the worst offenders, then a second, more precise unit further down to catch what slipped through. It sounds obsessive until you hear how clean the final vocal sits.

Running two instances isn't a sign you've failed. It's a sign your source material had complex sibilance patterns. Rap vocals with deliberate vocal grit and distorted saturation are the classic case for this approach.

How Do You Find the Problem Frequencies Before Reaching for a Plugin?

This is the step most guides skip. They jump straight to plugin selection when the real skill is knowing what frequency range you're actually fighting.

Load a spectrum analyser on your vocal and solo the worst sibilant phrase. Watch where the spike hits during the "s" sounds. On a female pop vocalist, we've seen this peak as high as 8kHz–10kHz. On a male podcaster with a darker tone, the same perceptual harshness might live closer to 5.5kHz–6.5kHz.

SPAN from Voxengo is free, shows real-time and peak-hold spectra, and takes about 30 seconds to read. Put it before your de-esser, find the spike frequency, then set your de-esser's detection range to match. Don't let the plugin guess when you can measure.

We once spent 20 minutes trying to dial in FabFilter Pro-DS on a falsetto vocalist before realising the problem frequency was at 9.4kHz, which was outside the default detection range we'd loaded. Moving the detection point 1.2kHz higher solved it in under a minute. Frustrating lesson, but it taught us to measure first, every time.

Which De-Esser Plugin Should You Actually Use?

The plugin market for de-essers runs from free to $259. Here's where we actually stand on the main options.

FabFilter Pro-DS ($189)

Pro-DS is the most transparent and controllable dedicated de-esser available ITB. The wideband/split-band toggle is clearly labelled, the detection range is fully adjustable, and the real-time display shows you exactly which frequencies are being caught and by how much.

The "Allround" mode uses a lookahead algorithm that detects sibilance before it peaks, which produces a smoother result than reactive detection. On a bright pop vocal with fast, staccato "s" sounds, this makes a noticeable difference in how natural the reduction sounds.

At $189, it's not cheap. But it earns its place on every vocal session we run. If you're mixing vocals seriously, this is the one to own.

Waves Sibilance (often under $50 on sale)

Waves built Sibilance around what they call "Organic Resynthesis" technology. The idea is that rather than simply compressing the sibilant band, the algorithm analyses the sibilance and reconstructs a smoother version of it.

In practice, it works faster to set up than Pro-DS. There are fewer controls, the detection is largely automatic, and on straightforward pop or country vocals it sounds clean with minimal effort. We've used it on podcast archives where 60 vocal files needed de-essing quickly, and it handled the batch work without fuss.

Where it struggles is on unusual vocal sources. A falsetto with harmonic content riding up near 10kHz confused the detection on our tests. The resynthesis artifact on that material was more obvious than a standard split-band compressor would have produced. Fast for normal sources. Less reliable on edge cases.

Soothe 3 ($259, $55 upgrade from Soothe 2)

Soothe 3 isn't technically a de-esser. It's a dynamic resonance suppressor that happens to work brilliantly on sibilance when you narrow its focus to the upper frequency range.

We love it for problem sources where a standard de-esser leaves a weird scooped quality. Soothe responds to the character of the resonance rather than just its level, so the reduction tracks the voice more organically. On distorted rap vocals or heavily processed sources, it's the tool that makes everything else sound wrong by comparison.

At $259 it's an investment. The $55 upgrade path from Soothe 2 is a different conversation. If you already own Soothe 2, upgrade immediately. If you don't, consider whether Pro-DS at $189 covers your needs first.

sonible smart:deess (€129)

AI-driven detection that analyses the specific vocal and adapts its frequency response accordingly. The hands-off approach is genuinely clever. Load it, press the analyse button, and it maps the sibilant signature of your source.

We'd recommend it for producers who mix occasionally and don't want to spend sessions learning detection curves. For engineers who mix daily and want full manual control, Pro-DS wins on flexibility.

Eiosis e2deesser ($99 via Slate Digital)

Underrated. The e2 uses two detection modes ("sensitivity" and "frequency") that give you precise control over what triggers reduction and by how much. It sounds clean on bright female vocals and handles dynamic vocal performances better than most options at this price point.

TDR Nova (Free)

TDR Nova is a dynamic EQ, not a dedicated de-esser. But used correctly, it functions as one with more precision than most stock de-esser plugins. Set a dynamic band around your problem frequency, dial the threshold to catch only the peaks, and you've built a de-esser that's fully transparent and costs nothing.

The workflow is slower than a dedicated plugin. But for producers on a budget or those who want to understand the mechanics of de-essing before spending money, this is the right starting point.

Do Stock DAW De-Essers Actually Work?

They work for light problems. They fall short on anything complex.

Logic Pro's stock de-esser handles an occasional vocal with mild sibilance. The detection algorithm is basic but functional. Pro Tools' Avid de-esser is similar. Ableton Live doesn't ship with a dedicated de-esser at all, though a dynamic EQ band on the stock EQ Eight covers the same ground manually.

The gaps show up on difficult material. Bright pop vocals with fast sibilance patterns, falsetto voices, spoken word recorded close to the microphone, rap vocals with deliberate vocal grit. Stock de-essers on these sources either over-correct (audible lisping) or under-correct (sibilance still cuts through). The detection algorithms don't adapt to the source the way dedicated plugins do.

Our honest position: use stock tools to understand the concept. Then buy a dedicated plugin when stock tools start producing noticeable artifacts. For most serious vocal work, that crossover happens quickly.

External Sidechain Triggering: The Overlooked Technique

Every guide mentions detection frequency. Almost none cover external sidechain routing for de-essing on distorted or heavily processed sources.

Here's the problem. When you saturate or distort a vocal, the sibilance gets baked into the harmonic content. A standard de-esser tries to detect sibilance from the distorted signal, which is a mess of harmonics that confuses any detection algorithm. The result is over-triggering, under-triggering, or artifacts that make the vocal sound broken.

The fix: route a clean, unprocessed version of the vocal to the de-esser's sidechain input. The detection circuit reads the clean signal, but the gain reduction applies to the distorted one. The de-esser knows exactly when to reduce because it's reading clear sibilance data, not a wall of harmonics.

FabFilter Pro-DS and TDR Nova both support external sidechain inputs. This technique is underused. We started applying it to distorted rap vocal chains after a session where every standard approach produced obvious artifacts, and the difference was immediate. Satisfying when it clicks.

Best de-esser plugins ranked

Summary

De-essers solve a specific problem: sibilance spikes in the 5kHz–10kHz range. Measure the problem frequency first with a spectrum analyser, then match your plugin choice to your source material. FabFilter Pro-DS at $189 is our pick for serious vocal work. Waves Sibilance works for quick fixes on standard sources. TDR Nova covers you for free if you're willing to build the signal chain manually. Use split-band mode for transparency, wideband for dense mixes. Always place the de-esser after compression. For distorted sources, use an external sidechain. Stock DAW tools handle light problems. Anything harder requires a dedicated plugin.

Compression plugins guide