Can’t Afford Pro De-Essers? Achieve Pro Results with Stock & Free Plugins (Part 3)

Sibilance is a common issue in vocal recordings. Not every musician has the means to reduce it at the source or the budget for premium, specialized plugins like FabFilter Pro-DS or oeksound soothe2. That’s the reason for this tutorial series: “Can’t Afford Pro De-Essers? Achieve Pro Results with Stock & Free Plugins!” This series provides free, low-cost methods using only your DAW’s stock plugins and the best freeware to effectively de-ess vocals.

In the first two tutorials, I covered methods using compressor and dynamic EQ combos, standalone dynamic EQ, and multiband compression. These can handle most common sibilance issues. But in the real world of mixing, we encounter trickier scenarios—like sibilance that ‘wanders’ across the frequency spectrum with no fixed center, or situations where a detail-obsessed mixer needs a more transparent, surgical touch.

This third installment dives headfirst into the ‘technical’ side of de-essing. I’ll explore the near-surgical precision of FFT dynamic equalization and spectral editing, unlock advanced de-essing techniques inside the free TDR Nova dynamic equalizer, and introduce a clever sidechain-powered method: using a Noise Gate to control parallel compression.

As always, I will use REAPER 7.39 for the demonstrations, but these concepts are universal and can be adapted to any DAW and plugin suite.

In this series:

 

• • Part 1: Dynamic EQ & Compression Basics
  • Part 2: Multiband & Parallel De-Essing
  • [You are here] Part 3: Advanced Surgical Techniques & Free Plugins

 

---

 

Surgical Precision #1: FFT Dynamic EQ with ReaFIR

When sibilance is elusive—appearing at different frequencies across different words—and you have time for meticulous processing, FFT dynamic EQ is your ultimate tool. This method offers a “surgical strike” on sibilant frequencies, maximally preserving the natural body and texture of the vocal. It’s a lifesaver for dynamic, unpredictable performances, such as fast-paced rap, where “s,” “sh,” and “ch” sounds can vary significantly in frequency distribution.

REAPER’s stock ReaFIR plugin is a powerful FFT-based equalizer. It uses real-time FFT analysis to isolate and process specific frequencies, making it a perfect nemesis for wandering sibilance.

Note: For users of other DAWs, you can install the standalone VST version of ReaFIR from the free ReaPlugs VST FX Suite by Cockos.

 

How It Works

ReaFIR’s core de-essing principle is real-time FFT subtraction. In simple terms, it analyzes the spectral content of a pure sibilant sound to create a “noise profile.” Then, as it processes the full vocal, it dynamically attenuates only the frequencies matching this profile. The human voice’s fundamental body frequencies are left almost completely untouched.

This method offers vastly higher frequency resolution than traditional dynamic EQ or compression, resulting in a more transparent sound.

 

Step-by-Step Walkthrough

Let’s go through the process using a vocal clip as an example.

(Listen to the unprocessed audio: 3.1_Original Vocal with Wandering Sibilance.flac)

(Lyrics: “Won’t fall again, I’m breaking through to live the life I want, step by step. No more silence, starting now, my future’s in my own hands.”)

 

Step 1: Create a Parallel Processing Track

We need two vocal tracks. Label the original track “Vocal De-ess DRY” and create a duplicate directly below it, labeling it “Vocal De-ess WET.” We will insert ReaFIR on the WET track. Because ReaFIR currently lacks a built-in dry/wet mix knob, this parallel routing is essential for blending the processed and unprocessed signals.

 

Step 2: Load ReaFIR and Set the Mode

Insert the ReaFir (FFT EQ+Dynamics Processor) plugin on the “Vocal De-ess WET” track. The interface shows areas for mode selection, a spectral display, and parameter controls.

 

Figure 1: ReaFIR’s initial interface.

From the Mode drop-down menu, choose Subtract. This is the crucial FFT subtraction mode.

Now, check the box labeled “Automatically build noise profile” that appears next to the menu. This tells the plugin to listen to a selected clip and create a spectral profile from it.

 

Figure 2: The interface after selecting “Subtract” mode.

 

Step 3: Select and Play a Pure Sibilance Clip

In your vocal media item, find a very short section containing only a clean sibilant sound—an isolated “s” or “sh” with no underlying vocal tone. Precisely make a time selection around it and hit play.

While playing, ReaFIR will automatically analyze the spectrum and draw the noise profile, usually in red, over the frequency display. You’ll see the sibilant frequency peaks clearly tagged.

 

Figure 3: ReaFIR’s noise profile displayed on the spectrum graph.

Important: If you accidentally included part of the vocal in your selection, click the Reset button and repeat this step with a cleaner sibilance section.

 

Step 4: Adjust the Wet/Dry Mix

Once the profile is perfect, uncheck the “Automatically build noise profile” box to lock it in and prevent it from being updated during playback.

Now, start playback of the entire vocal. Initially, pull the fader of the “Vocal De-ess DRY” track all the way down. Slowly push it up while listening, blending it with the “Vocal De-ess WET” signal. The lower the DRY track, the stronger the de-essing effect. Find the perfect balance point where the sibilance is tamed but the vocal remains completely natural and unaffected.

 

The Result

After ReaFIR processing, the harsh sibilance is effectively suppressed with virtually no negative impact on the vocal’s primary tone. Listen to the before-and-after comparison.

(Listen to the processed audio: 3.2_Vocal After ReaFIR Processing.flac)

 

Pro-Tips & Warnings

 

• • The quality of the noise profile is everything. A ‘contaminated’ profile will yield poor, possibly damaging results. Be meticulous in your selection.
  • If the vocal sounds hollow or muffled, the WET signal is likely too strong, or the profile includes too many non-sibilant vocal frequencies. Re-adjust your fader balance or re-capture a purer noise profile.
  • For extremely short, transient-heavy sibilants, you may need to experiment with several profile captures or manually edit the attenuation curve directly in the spectral display for ultimate precision.

 

---

 

Surgical Precision #2: Spectral Editing

For a more “visual and manual” approach to eradicating wandering sibilance, spectral editing is as precise as it gets. This method lets you see sibilance peaks directly on a spectrogram and manually pull them down, achieving pixel-perfect, point-to-point control.

Many DAWs include spectral editing features, like Apple Logic Pro’s Flex Pitch, Steinberg Cubase’s SpectraLayers, and REAPER’s built-in Spectral Edit.

 

Best Used For

This is ideal for short, critical vocal material where absolute clarity is paramount—think voiceovers, podcast intros, ADR, and commercial narration (ads, documentaries, game voice-overs). These projects often involve short, highly polished clips.

For full-song vocal tracks, the sheer volume and speed of sibilance make this method extremely time-consuming and is generally not necessary.

 

Walkthrough: REAPER’s Spectral Edit

I’ll use a short, self-recorded ad voiceover clip to demonstrate.

(Listen to the unprocessed clip: 3.3_Original Ad Voiceover.flac)

(Script: “Qing Hua Lang, Chishui River Left Bank Estate Liquor”)

 

Step 1: Enable the Spectrogram View

Right-click the vocal item, hover over Spectral edits, and in the submenu, enable Always show spectrogram.

 

Figure 4: Highlighted menu item to enable the spectrogram view.

The item will now show a waveform and a frequency-over-time spectrogram, where brighter colors indicate greater intensity. Time is on the horizontal axis and frequency on the vertical axis.

 

Figure 5: The spectral editing view in REAPER.

 

Step 2: Locate the Sibilance

Play the clip while watching the spectrogram. Sibilance typically resides in the 4–8 kHz range and will appear as bright, short-duration “spikes.” Zoom in for a closer look:

 

• • Scroll your mouse wheel to zoom in/out on the timeline.
  • Vertically resize the track lane or click the + icon at the bottom of the track’s vertical scrollbar to zoom in on the frequency axis.

Tip: For finer navigation, consider temporarily disabling your DAW’s ‘Snap to Grid’ function (the magnet icon in REAPER).

 

Step 3: Manually Reduce Sibilance Spikes

Place your edit cursor over a sibilant spike. Right-click the item and go to Spectral edits -> Add spectral edits to item. A white box with control handles—the spectral edit window—will appear, targeting the frequencies beneath it.

The basic controls are intuitive:

 

• • Drag the borders or corners to resize the box. A wider box affects more time; a taller box covers a wider frequency range.
  • Drag the center of the box to reposition it.
  • The control knob at the top-right adjusts the gain for the selected region.

Adjust the box to frame only the brightest part of the spike. Then, turn the top-right gain knob to reduce the level. Cut of -3 to -10 dB is typical. Use your ears as you fine-tune.

 

Figure 6: Gain adjustment for a single sibilant peak.

Add multiple spectral edits to process every offending spike individually. Review your work by playing the whole clip and making any final micro-adjustments.

 

Figure 7: All manual spectral edits on the example clip.

 

The Result

The transformative power of careful spectral editing makes the vocal remarkably smoother and more polished.

(Listen to the processed clip: 3.4_Vocals After Spectral Editing.flac)

 

In Other Software

 

• • Logic Pro (Flex Pitch): In the Flex Pitch view, select the spectral view to manually attenuate peaks.
  • Cubase (SpectraLayers): This offers even deeper control, allowing you to separate the sibilance into a distinct spectral layer for isolated, non-destructive processing.

 

Pro-Tips & Warnings

 

• • Efficiency is key. For long-form narration or podcasts, this level of detail may be too costly in terms of time. Dynamic EQ (covered in Part 1) is often more practical.
  • Over-processing is a real danger and can cause a ‘lispy,’ dull, or even low-bitrate, degraded quality. Small, cautious cuts are the rule.
  • Frequently A/B compare your edited version with the original to ensure the result is both effective and natural-sounding.

 

---

 

Unlocking TDR Nova’s Advanced De-Essing Potential

TDR Nova is a phenomenal freeware parallel dynamic equalizer with great de-essing prowess. In Part 1, I covered the basics. Now, let’s unlock its more advanced, high-potential techniques for a perfectly natural result.

 

Technique 1: Fine-Tuned A/B Comparison

Minute parameter changes can have a huge impact. Nova’s A/B comparison function allows you to quickly flip between two settings to find the most transparent option.

Your goal is to suppress the sibilance while maintaining a clear, natural “S” sound, avoiding the dreaded lisp.

Let’s use a pop vocal clip with some clear sibilance.

(Listen to the unprocessed clip: 3.5_Original Pop Vocal with Sibilance.flac)

Workflow:

 

1. 1. Initial Setup: Load TDR Nova on your vocal track. Using the sweep method (from Part 1), find the center frequency of the sibilance (say, 5 kHz). Set an appropriate Q to isolate the sibilant region on one band.
   2. Create A/B Settings: Click the A B button at the top of the plugin. If it’s greyed out, first click A > B to copy settings from A to B and activate the function.
      • Setting A (Gentler): Ratio 3:1, Gain -4 dB.
      • Setting B (More Aggressive): Ratio 5:1, Gain -6 dB.
      • Keep the Threshold identical for both, starting around -20 dB and adjusting until the dynamic EQ is reliably triggered by the sibilance.

      
      Figure 8: The interface buttons for switching between and copying the A/B parameter sets.
      Figure 9: Visual comparison of the A and B parameter configurations.

   3. Loop and Compare: Select a loop around a sibilant phrase. Play it continuously and rapidly click the A B button to switch settings. Focus intently on the severity of the sibilance and the clarity of the “S” sound.
   4. Select and Micro-Adjust: Pick the more natural-sounding starting point. If you chose B (more aggressive) and detect a slight lisp, nudge the Ratio down to 4:1 or the Gain to -5 dB, then re-evaluate.

 

The Result

The final setting after careful A/B comparison effectively controls the sibilance while preserving natural articulation and clarity.

(Listen to the processed clip: 3.6_Vocals After TDR Nova Fine-Tuning.flac)

 

Technique 2: Multi-Phrase Check & Automation

A singer’s sibilance intensity fluctuates with emotion and energy. A single static setting rarely works for an entire song. A multi-phrase check and automation are critical.

Workflow:

 

1. 1. Full-Song Review: Listen to the entire vocal track from start to finish. Note the sections where sibilance suddenly jumps out.
   2. Global Adjustment (if needed): If most of the track seems under-processed, make a small, global tweak, like lowering the Threshold by 1 dB. Be conservative to avoid over-processing the quieter parts.
   3. Localized Automation: This is the key step. Instead of a global fix, automate a specific parameter just for the problematic lines.
      • • Method 1: Item Splitting. Split the problematic phrase onto its own item and load a second, more aggressive instance of TDR Nova just on that item.
        • Method 2: Parameter Automation. Create an automation lane for the Threshold or Gain parameter. On the harsh phrase, use automation to lower the Threshold (to trigger more compression) or make the Gain more negative (to add more attenuation). A 2 dB automation move is often perfect.

Expected Outcome: This delivers a consistent, polished de-essing across the entire vocal performance without killing the life of the non-sibilant parts.

 

Technique 3: Multi-Band Targeting

Free TDR Nova supports up to 4 dynamic bands. When sibilance spans two distinct frequency areas (e.g., a sharp peak at 4 kHz and another at 8 kHz), using two separate dynamic bands is far more transparent and precise than one wide one.

Concept: Simply set up two independent bands with different center frequencies and Q values. This allows you to apply more gentle, specific cuts to each sibilance hotspot rather than one huge, life-sucking scoop. Each band’s operation follows the same method from Part 1. Use the internal spectrogram view to visually verify your bands are correctly targeting the spikes.

Note: If the two sibilance frequencies are very close, they may interfere. Use higher Q values (narrower bandwidths) on each band to minimize overlap.

 

Technique 4: Upward Expansion for Air

This is a unique and powerful hidden trick in TDR Nova. By setting the Ratio to less than 1:1 (e.g., 0.8:1), the band acts as an upward expander. We can use this to brighten a dull vocal between sibilant sounds without making the sibilance worse.

How It Works: A Ratio below 1:1 means the signal is boosted when it falls below the threshold. By setting the Threshold above the sibilance peaks, the expander does nothing during the sibilance but lifts high-frequency “air” in the gaps, increasing overall brightness and liveliness.

Setup:

 

1. 1. Select a high-frequency band (e.g., 8-12 kHz for air, or use the default High Shelf for Band 4).
   2. Set the Threshold high, just above the peaks of your sibilance. The idea is that sibilance doesn’t trigger it.
   3. Set the Ratio to something like 0.8:1. This controls the boost amount below the threshold.
   4. Set a slightly longer Attack time (e.g., 50 ms) to avoid a jittery sound.
      Figure 10: Parameter settings for upward expansion.

 

Technique 5: Sidechain De-Essing

For the ultimate in precision, TDR Nova’s external sidechain input allows for a classic split-band de-esser design. The detection signal that triggers the processing is a heavily EQ’d version of the vocal, containing only sibilance.

 

Figure 11: TDR Nova’s I/O routing panel, highlighting the sidechain inputs.

Workflow:

 

1. 1. Create a Sidechain Detection Track: Duplicate your main vocal track. Name it “Sidechain – Sibilance.” Insert an EQ with a drastic band-pass filter, isolating only the sibilance region (e.g., 4-8 kHz). No other frequencies should be audible from this track.
   2. Set Up Routing: In your DAW, route this detection track’s output to TDR Nova’s sidechain input. In REAPER, this involves setting the main vocal track’s channel count to 4 and routing the sidechain track’s output to channels 3/4.
   3. Configure TDR Nova: On the main vocal track’s TDR Nova instance, switch the SC Source from Int SC (internal) to Ext SC (external).
   4. Set Key Parameters: Now, TDR Nova will only react when the detection track has signal—pure, isolated sibilance. Set your dynamic EQ band’s parameters. You can use a very fast Attack and release, as you’re now working from a perfectly clean trigger signal.
   5. Critical Step: The “Sidechain – Sibilance” track must be muted or have its master send disabled. It’s a control signal and should never be heard in the final mix.

The Result: This method ensures the compressor fires only on sibilance with unrivaled accuracy, leaving the rest of the vocal 100% untouched.

 

---

 

Smart Control: Using a Sidechain Gate for Parallel Compression

This builds on the parallel multiband compression technique from Part 2. By adding a noise gate with a sidechain input on the parallel “Sibilance” track, we create an intelligent system that only activates the parallel processing during sibilant moments.

Crucial Caveat: This technique only works when sibilance’s instantaneous level is significantly louder than the rest of the vocal. If the level difference is small, this method will cause the sound to jitter and is not suitable.

 

The Concept

The sidechain gate acts as a smart switch. The original “Body” vocal track triggers it.

 

• • Sibilance Present (Loud Signal): The gate on the “Sibilance” track opens, allowing the heavily compressed signal to blend in and tame the harshness.
  • No Sibilance: The gate is fully closed, completely muting the parallel “Sibilance” track. Only the pristine “Body” track is heard.

This prevents non-sibilant parts from being unnecessarily colored or dulled by the parallel compression.

 

Step-by-Step Walkthrough

Let’s hear the raw vocal first.

(Listen to the unprocessed clip: 3.7_Raw Vocal for Sidechain Gate Control.flac)

 

1. 1. Set Up Parallel Tracks: As in Part 2, create your “Body” track (dry vocal) and “Sibilance” track (with a multiband compressor targeting 4-8 kHz, Ratio 4:1 to 6:1, etc.).
   2. Insert a Sidechain Gate: Insert a gate with sidechain capability (like REAPER’s ReaGate) on the “Sibilance” track, after the compressor.
   3. Configure the Sidechain Input: Route the signal from your “Body” track to the gate’s sidechain input. For even more precise triggering, you can create an “Sib Enhanced” track with EQ that exaggerates the sibilance, and use that as the sidechain source instead.
      Remember: Regardless of which track you use as the sidechain source, its master send must be disabled.
   4. Dial in the Gate Settings:
      • • Detector Input (ReaGate): Set this to Auxiliary Inputs so it listens to the external sidechain signal.
        • Threshold: Set it so the gate opens firmly with every sibilant peak and fully closes otherwise. Watch the green meter versus the threshold slider.
        • Attack: Instantaneous, 0-5 ms.
        • Release: Fast, typically 20–50 ms, so the gate snaps shut as soon as the sibilance ends.
        • Range/Hold: Set Range to -∞ to completely silence the track when the gate is closed. If you hear jitter, slightly increase the Hold time.
          Figure 16: Sidechain gate parameter configuration.
   5. Balance the Mix: With the gate now active, play the vocal. Set the “Body” track fader first, then slowly bring up the “Sibilance” track fader until the harshness subsides. Because the gate mutes the parallel track between sibilants, this blend is incredibly clean.

 

The Result

This approach treats only the sibilance with heavy processing, leaving zero artifacts or dulling on the rest of the vocal signal. It’s incredibly transparent.

(Listen to the processed clip: 3.8_Vocals After Sidechain Gate Parallel Compression.flac)

 

Pro-Tips

 

• • Soloing the “Sibilance” track is the ultimate quality check. You should hear a series of clean, chopped-up “ch,” “sh,” and “s” sounds with total silence in between. Any background leakage means your threshold is too low or release is too long.

 

---

 

Summary & Final Thoughts

In this third part, we’ve expanded your free de-essing toolkit with some truly advanced, surgical methods. There is no single “best” method, as every vocal, genre, and mixer’s style is different. The most professional approach is to listen, diagnose the specific type of sibilance problem, and select the tool that offers the most transparent solution for that task.

Here’s a final recap of the advanced techniques:

 

• • FFT Dynamic EQ & Spectral Editing: Ideal for unpredictable, ‘wandering’ sibilance where maximum surgical precision is needed to preserve the source. These are time-intensive but yield incredibly pure results.
  • TDR Nova’s Advanced Techniques: A powerhouse free plugin with A/B comparison, intelligent automation, multi-band targeting, upward expansion, and external sidechaining. This offers a “Swiss Army Knife” solution for highly complex and dynamic sibilance challenges.
  • Sidechain-Gated Parallel Compression: A smart evolution of the parallel technique for material with very prominent sibilance peaks, adding processing only when needed for ultimate transparency.

No matter which technique you use, always abide by the golden rule of de-essing: “less is more.” Over-processing is the quickest way to ruin a vocal’s natural bite and intelligibility. Meticulous, repeated listening on different systems (studio monitors, headphones, earbuds) is the only path to a flawless, professional-grade result.

This concludes our series on affordable de-essing. I hope these tutorials empower you to solve even the most stubborn sibilance problems and create polished, radio-ready vocal mixes, all without breaking the bank.