Best Free De-Esser Plugins (2026): 7 Picks and How Each One Works

For one free de-esser and nothing to think about, start with Techivation's T-De-Esser 2: a few knobs, a Diff listen mode, and VST, VST3, AU, AUv3 and AAX builds. If the vocal loses air when it triggers, move to a split-band or dynamic EQ design such as TDR Nova, Ghostnote VEIL or ZL Equalizer 2, which only turn down the band that stings. Several free de-esser lists still recommend 32-bit Windows plugins a current Mac cannot load, so every pick below was checked on its maker's page on September 30, 2026.

Disclosure: we make VEIL. It is one pick of seven and gets a labeled section near the end. No affiliate links.

The quick answer: which free de-esser should you pick?

Pick by type before you pick by name: wideband, split-band or dynamic EQ (explained in the next section). All seven were free, with a working download page, on September 30, 2026.

How does a de-esser work, in 60 seconds?

Every de-esser detects an ess, usually by listening to a filtered copy of the high band, then reduces something. The plugins below differ mostly in what gets turned down.

• Wideband: the detector hears the ess, and the gain of the whole signal drops for a few milliseconds. Easy to set. Push it and the vowel next to the ess dips too, which is where the "lisp" comes from.
• Split-band: the voice is split at a crossover, and only the band above it gets turned down. The body of the word never moves, so the vocal does not dull on every ess, as long as the crossover sits above the voice's own brightness.
• Dynamic EQ: a bell or shelf that only cuts when its band gets loud. You choose center, width and depth, so it can hit one narrow ess frequency and leave the air alone.

That is why the type, more than the brand, decides whether your vocal dulls. Our guide on how to remove sibilance without dulling the vocal covers each approach in depth.

The 7 free de-essers

1. Techivation T-De-Esser 2

What it is: a dedicated de-esser that Techivation keeps to a few knobs and buttons. Its product page says that when the targeted range exceeds the threshold, "the whole signal is compressed": a wideband design.

Key controls: Processing (sensitivity), Intensity (amount of gain reduction), Sharpness (a knee), four Frequency Range buttons, make-up gain, Mix, and oversampling up to 8x.

Formats and OS: VST, VST3, AU, AUv3 and AAX; Windows 7+ (64-bit), macOS 10.11+ on Intel and Apple Silicon, plus iPad. Download after a free registration on the T-De-Esser 2 page.

What to listen for: Diff plays the difference between processed and original. You want short ticks of ess; if you hear the vowel, back off Processing, the control Techivation says Diff helps you fine-tune. Filter plays only the band the detector watches.

Who it suits: a de-esser in two minutes, and Pro Tools users.

2. TDR Nova

What it is: Tokyo Dawn Labs calls Nova a "parallel dynamic equalizer". It looks like a parametric EQ with four bands plus high-pass and low-pass filters, but each band carries its own dynamics section with threshold, ratio, attack and release. Tokyo Dawn lists "a sibilance problem in your perfect take" among its uses on the TDR Nova page. A paid edition, Nova GE, adds more bands.

How detection works: each band listens to its own region, so a bell on the ess frequency only reacts when that region gets loud. A high shelf instead gets you close to a split-band de-esser.

Formats and OS: VST2, VST3, AU and AAX; Windows 7+ and Mac OS X 10.9+, native Apple Silicon since 2.1.4. Current version 2.2.2, with direct installers or the TDR Collector manager.

What to listen for: find the frequency where the esses sting, then set the threshold so only the loudest esses trigger it. The reduction should flicker, not sit down.

Who it suits: producers who think in EQ. It is also the one de-esser our free rap vocal plugin guide names inside a whole chain.

3. Analog Obsession LOADES

What it is: a de-esser that Analog Obsession lists as "Smooth & powerful" on its dynamic processing page. The developer's FAQ states the plugins are "totally free", with optional support through Patreon, where the downloads are posted in the LOADES release post.

Key controls: per that post, THR (a threshold with a 60 dB range), REL (release from 1 to 500 ms), MIX (dry against processed), IN and OUT (plus or minus 30 dB), a BAND button that switches the reduction from shelf type to band type, and a SOFT button for smoother processing. The post also states there is no oversampling.

Formats and OS: VST3, AU and AAX on macOS 10.11 or later (Intel and Apple Silicon), and VST3 and AAX on Windows 10 to 11, per the release post, which dates version 2.0 to September 2, 2024. The Mac build needs a Metal-capable graphics card, the Windows build OpenGL.

What to listen for: start in shelf mode; if the whole top end goes soft when it triggers, try BAND. If full reduction sounds lispy, blend in dry signal with Mix.

Who it suits: fans of the hardware layout: threshold, release, mix.

4. Ghostnote VEIL

What it is: our free split-band de-esser. It splits the voice at a crossover you set between 3 and 10 kHz (5.5 kHz by default) and applies gain reduction only to the band above it.

How detection works: VEIL measures sibilance against the voice's own recent balance rather than a fixed level, so loud and quiet passages of a clean recording are treated alike. On a quiet, hissy take, raise SENSE. 3 ms of lookahead puts the reduction in place as the ess starts.

Key controls: SENSE (0 to 100, default 50), CROSSOVER, RANGE (1 to 24 dB, default 10) and RELEASE (20 to 200 ms, default 60), plus three LISTEN modes: Output, Sibilants and Removed.

Formats and OS: VST3 on macOS and Windows, AU on macOS; macOS 11 or later and Windows 10 or later, 64-bit, no licence key. Enter your email on the VEIL page and the download links appear straight away.

What to listen for: in Removed, you should hear esses and nothing else. If words come through, lower SENSE.

Who it suits: takes that drop to a mumble and come back shouting, where one fixed threshold catches too much in one place and too little in the other.

5. Airwindows DeBess

What it is: Airwindows' improved version of its older DeEss, posted in 2019 on the Airwindows DeBess page. A free download, and Airwindows code is MIT-licensed.

How detection works: no crossover knob. It watches how fast the waveform changes, and how that rate changes, a pattern hissy consonants produce, then pulls the signal toward a low-passed copy of itself. You can read the logic in the DeBess source code.

Key controls: Intense (how hard it triggers), Sharp (the length of the comparison window, which the Airwindows page describes as now being a slider), Depth (the maximum reduction), Filter (how the treble is shaped on the esses) and Sense, which above its midpoint plays only what is being removed.

Formats and OS: the easiest way to get it today is Airwindows Consolidated, a free, open-source build of every Airwindows plugin as CLAP, VST3, AU and LV2 for macOS (universal), Windows x64 and Linux. The build on its release page was dated September 19, 2026.

What to listen for: flip Sense on and raise Intense until the esses appear against silence. Airwindows says it wants very bright esses, is not meant for stage mics, and may not engage on prosumer gear or moving-coil mics.

Who it suits: Linux and open-source setups.

6. ZL Equalizer 2

What it is: a free, open-source equalizer (AGPLv3) with up to 24 bands, each of which can be made dynamic. The ZL Equalizer 2 manual documents threshold, knee, attack and release per band, plus a side-chain filter you shape separately.

How detection works: its Relative mode triggers when the band gets louder than the overall side-chain signal, not a fixed level, so the band reacts to the voice's balance rather than its loudness. Dynamic Learning can set threshold and knee for you.

Formats and OS: VST3 on macOS, Windows and Linux, AU on macOS, and LV2 on Windows and Linux, per ZL Audio's plugin installation help, which also says AAX and CLAP are not included and the installers are not notarized. The free, open-source (AGPLv3) status is on the ZL Equalizer 2 installation page. The latest stable release, 1.4.0, was posted on the GitHub releases page on September 19, 2026, with Apple Silicon and Windows ARM builds.

What to listen for: double-click a band's button to solo it, find the ess, then point the side-chain filter at the same region.

Who it suits: Nova-style precision with a detector that follows the voice's balance.

7. Immanent Audio IA De-Esser

What it is: a free plugin that Immanent Audio pitches as de-essing plus resonance suppression. Its product page describes four filter slots with low-cut, high-cut, depth and Q, and three modes: De-ess, Auto, which chooses depth and Q for you, and Smart, which "hunts for any part of the spectrum" that needs taming.

How detection works: the page does not state the method; it does state zero added DAW latency.

Formats and OS: VST3 and AU, on Windows and macOS, downloaded from its Gumroad listing listed free with an optional pay-what-you-want amount; you choose Windows or macOS at checkout.

What to listen for: Smart mode can reach below the ess range, so bypass-compare at matched level.

Who it suits: a harsh vocal with both sibilance and a ringing resonance.

Free de-essers that are not worth installing in 2026

These still appear on free de-esser lists, but those same lists describe them as 32-bit or Windows-only VST, or the maker's site is offline. A 32-bit plugin does not load natively in a 64-bit DAW, and nothing 32-bit runs on a current Mac.

• Spitfish (Digital Fish Phones): listed as Windows, 32-bit on Best Free Plugins, and described by Musician's HQ as more than 20 years old and not updated since 2005, with reports of it not working for some producers.
• Tonmann DeEsser: 32-bit VST per MIDINation and Best Free Plugins.
• Modern De-Esser (Antress) and VeeDeeS (Viper ITB): both Windows, 32-bit per MIDINation.
• Lisp (Sleepy-Time DSP): Windows VST only per Best Free Plugins, so Mac users cannot run it.
• KG Series (Kage Music): still listed on Free VST Hub, but the maker's site returned "The domain has been disconnected" on September 30, 2026.

Before you install any of these, check what your DAW already ships. Many include a stock de-esser or a dynamic EQ that can do the job; in FL Studio, Image-Line's manual describes its Maximus plugin (Producer Edition and up) as a compressor, limiter and de-esser in one.

How to test a free de-esser on your own vocal

Compare three of these in ten minutes, on your voice rather than anyone's list, ours included.

1. Pick the ess-heavy line. Loop the bar with the most s, sh and ts sounds.
2. Listen to what is removed. Use the plugin's difference or removed mode: Diff on T-De-Esser 2, band solo on ZL Equalizer 2, Sense on DeBess, Removed on VEIL. Hearing vowels means it is taking the word.
3. Watch the meter. Gain reduction should move on consonants only. If it stays down through the vowels, the plugin is compressing the vocal, not de-essing it.
4. Compare at matched level. Quieter sounds duller, so match loudness before you judge tone.
5. Check the quiet part. A threshold set on the loud verse often misses esses in the soft one.

For starting frequencies, reduction targets and where the plugin goes in the chain, use our de-esser settings guide; the numbers are the same whichever free plugin you land on.

From Our Own Line

VEIL is our de-esser, and it is free. Gain reduction only lands above the crossover, the detector follows the voice's own balance, and the Removed mode plays only what it takes out. Version 1.2 went live on September 24, 2026.

It is not a dynamic EQ: to draw one narrow band, use TDR Nova or ZL Equalizer 2. Download VEIL, or see the rest of the line on the plugins page.

Frequently Asked Questions

What is the best free de-esser plugin?

It depends on the type you need. Techivation T-De-Esser 2 is the quickest to set and ships as VST, VST3, AU, AUv3 and AAX. If a wideband de-esser dulls your vocal, try a split-band design like VEIL or a dynamic EQ like TDR Nova.

Is TDR Nova a de-esser?

Not by name, but it works as one. Nova is a free dynamic EQ whose four bands each have threshold, ratio, attack and release. Put one bell or high shelf on the ess frequency and it behaves like a de-esser.

What is the difference between split-band and wideband de-essing?

A wideband de-esser turns the whole signal down when it detects an ess, so the vowel around it can dip too. A split-band de-esser only turns down the frequencies above a crossover, which leaves the body of the voice untouched.

Does FL Studio have a de-esser?

Image-Line's manual lists de-essing among the jobs of its Maximus plugin (Producer Edition and up), and any VST3 de-esser on this list also loads in FL Studio. Our de-esser settings guide covers the frequencies and reduction targets to use, and our guide to de-essing in FL Studio walks through each method step by step.

What is a good free de-esser for Mac?

Every pick on this list installs on macOS as an AU. T-De-Esser 2, TDR Nova, LOADES, VEIL, ZL Equalizer 2 and Airwindows Consolidated list Apple Silicon support on their download pages. Skip Spitfish, Tonmann and Lisp, which are listed as Windows-only.

Where does a de-esser go in the vocal chain?

After subtractive EQ and before your first compressor, so the compressor reacts to the voice rather than the esses. Our rap vocal chain guide shows the full order, including a second light de-esser before saturation.