How to Remove Breathing and Mouth Noises From Vocal Recordings

Vocal recordings often capture unwanted sounds such as breathing, lip smacks, tongue clicks, and saliva crackles. These noises can distract listeners, undermine professionalism, and reduce the clarity of your message—whether you are producing a podcast, voice-over, audiobook, or music track. Eliminating these artifacts is a critical skill for any audio editor or content creator. In this expanded guide, we will explore the nature of these noises, strategies to prevent them during recording, and a comprehensive set of post-production techniques to remove them without damaging the natural quality of your voice.

Understanding the Nature of Breathing and Mouth Noises

Breathing sounds occur naturally when a speaker inhales and exhales. In a close-miked recording, these breaths can be disproportionately loud, especially during quiet passages or when the speaker has a deep or rapid breathing pattern. Mouth noises—such as clicks, pops, smacks, and wet sounds—are produced by the movement of the tongue, lips, and saliva. These sounds are most noticeable in the silences between words or phrases, and they become more prominent when the recording is compressed or normalized.

Recognizing the different types of mouth noise is the first step toward effective treatment. Common categories include:

Plosives – explosive sounds from “p,” “t,” “k” bursts hitting the microphone diaphragm. Often mitigated with a pop filter but can still leave residual energy.
Lip smacks – sharp, transient clicks created when the lips part after being closed.
Tongue clicks – percussive sounds from the tongue detaching from the palate.
Saliva crackles – low-level, granular noises caused by moisture in the mouth.
Breath noises – broadband, whoosh-like sounds from inhalation and exhalation.

Understanding these distinctions helps you choose the right removal tool and avoid over-processing that can introduce artifacts or dull the vocal tone.

Pre-Recording Prevention: Minimizing Noises Before They Happen

The most efficient way to deal with mouth and breath noise is to prevent it at the source. A few adjustments in recording technique and vocal preparation can dramatically reduce post-production workload.

Microphone Technique

Maintain optimal distance. Keep the microphone 6–12 inches from the mouth, slightly off-axis to reduce plosive force and sibilance.
Use a pop filter. A multi-layer mesh or foam windscreen dissipates the air blast from plosive consonants and also helps catch small saliva droplets that create clicks.
Angle the microphone. Position it just above or below the line of direct airflow to minimize breath hits without losing clarity.
Monitor with headphones. The speaker can hear exactly what the microphone picks up and adjust their delivery accordingly.

Vocal Preparation and Hydration

Drink room-temperature water. Staying hydrated reduces sticky saliva that causes clicks. Avoid dairy, sugary drinks, or caffeine before recording, as these can thicken mucus or dehydrate the mouth.
Mouth exercises. Gentle tongue and lip stretches, tongue rolls, and humming can loosen tight muscles and reduce involuntary clicks.
Snack on an apple. Some voice actors eat a green apple or drink apple juice because the natural acidity helps cut through saliva film and reduce mouth noises.
Avoid dry mouth. If your mouth feels dry, sip water or use a sugarless lozenge that promotes saliva flow.

Recording Environment and Habits

Record in a quiet, treated space. Background noise encourages you to gate aggressively, which can exaggerate residual mouth clicks. A dead room reduces the need for heavy processing.
Coach the talent. Instruct speakers to take natural, softer breaths away from the mic axis, or to hold their breath during very short phrases. In scripted narration, mark breathing points to keep breaths consistent.
Use a short silence before speaking. Many mouth clicks happen when the speaker starts talking. Having the talent wait a beat and open their mouth silently before the first word can prevent an initial click.

Post-Production Techniques for Removing Noises

Even with careful recording, some unwanted sounds will remain. Modern digital audio workstations (DAWs) and specialized plugins offer several effective methods to clean vocal tracks.

1. Manual Editing in Your DAW

The most time-consuming but precise method is manual editing. Zoom in to the waveform and visually identify breath or click artifacts. In most DAWs (such as Audacity, Adobe Audition, or Reaper), you can select a portion of audio and either silence it, lower its gain, or replace it with a tiny crossfade from the surrounding waveform.

Silencing breaths. Select the breath region and apply silence (or reduce gain by –12 to –20 dB). For naturalness, leave a very quiet room tone or fade the edges.
De-clicking clicks manually. Zoom in to sample level to see a sharp spike for a lip smack. Select just that spike (often 1–5 ms) and remove it or reduce its amplitude by 6–10 dB.
Spectral editing. Some DAWs (like Audition’s Spectral Frequency Display) let you draw over unwanted noises in the frequency domain. Breaths often appear as low-frequency broadband energy; click noises as vertical spikes across many frequencies. Use the spot healing brush or lasso tool to remove them without affecting the voice spectrum.

2. Noise Gates and Expanders

A noise gate mutes the signal when the volume falls below a set threshold, effectively silencing breaths and clicks during pauses. For more subtle control, use an expander (downward expander) that reduces gain instead of muting completely.

Set the threshold just above the noise floor but below the quietest spoken passages. Use a fast attack (0–5 ms) to avoid clipping the start of words, and a release time that matches the natural decay of room tone (50–150 ms).
Lookahead feature (if available) lets the gate preview the signal and open before the word begins, preventing unnatural chopping of consonants.
Sidechain EQ. Some advanced noise gate plugins allow you to filter the trigger signal so that only certain frequencies open the gate. This can help prevent stray mouth clicks from keeping the gate open.

3. De-Click and De-Noise Plugins

Specialized plugins automatically detect and suppress clicks, pops, and broadband noise while preserving the vocal timber. These are essential for batch processing or when manual editing is impractical.

iZotope RX is the industry standard. Its Mouth De-click module analyzes transient artifacts and separates them from the voice. You can adjust sensitivity, focus on different noise types (clicks, smacks, crackles), and preview the result. The Breath Control module reduces breath volume while preserving the natural ambience.
Waves X-Click and X-Crackle work well on older recordings or those with heavy tape noise. They are simple to use: adjust the threshold to catch only the transient spikes.
Accusonus ERA Noise Remover uses a one-knob interface to reduce a wide range of noises, including breaths and clicks, but can be less precise.
Acon Digital DeClick and DeNoise are affordable alternatives with clean algorithms.

When using these plugins, always process at the highest bit depth and listen critically to avoid over-processing that creates “pumping” artifacts or an unnatural lisp.

4. Spectral De-Noising

For stubborn broadband noises (like heavy breathing) that overlap with the voice frequency range, spectral editing tools offer pixel-level precision. In addition to iZotope RX’s Spectral De-noise, standalone tools like Adobe Audition’s Spectral Frequency Display or Audacity’s Spectrogram allow you to paint over breath energy. This should be your last resort because it can leave audible artifacts if not done with care.

Advanced Workflow: Step by Step

To achieve professional results without destroying natural vocal character, follow this recommended workflow:

Trim and normalize. Remove absolute silence at the start and end. Normalize peak level to –1 dB or –3 dB to give headroom for processing.
Run a de-click plugin first. Transient clicks and smacks are easiest to remove before any compression or reverb is applied. Set the plugin to a moderate threshold and listen for artifacts on sibilants.
Manual silence or reduce obvious breaths. Use your DAW’s gain envelope to lower breath volume by –10 to –20 dB. For a natural feel, leave some residual breath—complete silence can sound robotic.
Apply a noise gate or expander. Use a gate with a lookahead and a release that doesn’t clip word endings. For dialogue, a downward expander (2:1 ratio) often sounds more natural than a hard gate.
Use spectral editing for remaining issues. Zoom into the spectrogram and use a brush tool to remove crackles that survived the de-click pass.
Compress and EQ carefully. Compression can bring up residual mouth noises, so apply gentle compression (2:1 to 4:1 ratio) and avoid boosting frequencies above 8 kHz where crackles live. A subtle high-shelf cut can smooth out remaining noise.
Mask with ambience. If you silenced breaths completely, add a whisper-thin room tone (recorded during the session) under the dips to prevent audible silences.

Tools and Plugins Comparison

Choosing the right tool depends on your budget, workflow, and the severity of the noise. Here is a quick reference:

Tool	Best For	Price Range	Key Features
iZotope RX 10/11	Professional de-click, de-noise, breath control	$$$	Machine learning, spectral editing, match EQ
Waves X-Click / X-Crackle	Transient click removal on older recordings	$$	Simple controls, low latency
Accusonus ERA Noise Remover	Quick one-knob noise reduction	$$	Easy to use, real-time
Adobe Audition	Integrated spectral editing and adaptive noise reduction	$$	DAW + advanced tools
Audacity (free)	Manual editing and basic gate/expander	Free	Cross-platform, spectral selection
Reaper + ReaPlugs	Flexible gate/expander and free de-click JSFX	$	Customizable, community scripts

Common Pitfalls to Avoid

Over-gating. A gate set too aggressively will cut off the natural decay of words, creating “clipped” transitions. Use the expander first, then a gate only if necessary.
Removing all breaths. Small, natural breaths help the performance feel alive and human. Removing every breath sounds unnatural and can make the speaker sound robotic or breathless.
Processing too early. De-click before you apply compression or reverb. Compression brings up the noise floor, making clicks harder to detect. Reverb masks clicks but also smears them, making removal more difficult.
Ignoring monitoring environment. If you cannot hear subtle mouth noises on your headphones, you might miss them. Use closed-back, high-isolation headphones and listen at moderate volumes.

Conclusion

Removing breathing and mouth noises from vocal recordings is a blend of good discipline during the session and careful post-production editing. By understanding the nature of each noise type, investing in prevention techniques, and mastering manual and automated editing tools, you can produce clean, professional vocal tracks that keep listeners focused on the content. Whether you use a free tool like Audacity or a high-end suite like iZotope RX, the key is to work in stages—de-click first, then manage breaths, and always listen for the subtle balance between clarity and naturalness. With practice, these techniques will become second nature, and your recordings will sound polished and engaging.