Vocal recordings are the foundation of countless music productions, podcasts, audiobooks, and voiceovers. But even the most skilled vocalist produces unwanted artifacts: subtle clicks, pops, lip smacks, and audible breaths. These micro-noises accumulate, distracting listeners and undermining the polished, professional sound audiences expect. Removing or reducing them is a critical step in vocal processing, yet many engineers struggle to balance cleanliness with naturalness. This comprehensive guide explores manual editing, spectral tools, intelligent plugins, and preventative measures to help you achieve pristine vocal tracks without sacrificing performance energy.

Understanding Mouth Noises and Breaths

Mouth noises stem from saliva, tongue movement, and sudden changes in air pressure inside the vocal tract. Common types include:

  • Clicks and pops: Short, sharp transients often caused by the tongue separating from the palate or teeth.
  • Lip smacks: Low-frequency, sticky sounds that occur when lips part after being closed.
  • Throat clears and swallows: Longer, guttural noises that can be especially problematic.
  • Breath sounds: Inhales and exhales vary widely — from gentle, unobtrusive sips of air to loud gasps that compete with the vocal line.

While some breath sounds add realism and emotion (especially in intimate vocal styles), excessive mouth noise is almost never desirable. The goal is not total silence but a reduction of distracting elements to a level where the listener focuses entirely on the performance.

Prevention: The First Line of Defense

Before reaching for the mouse or a plugin, consider prevention. Simple pre‑recording practices drastically reduce post‑production work:

  • Hydration and diet: Encourage vocalists to drink room‑temperature water and avoid dairy, caffeine, or sugary drinks before recording. A dry mouth creates more clicks than a well‑lubricated one.
  • Microphone technique: Place the microphone slightly off‑axis and at a distance that captures the voice without accentuating mouth physics (12–18 inches is typical for spoken word). Use a pop filter to reduce plosives, which can mask or exacerbate clicks.
  • Vocal warm‑ups: A relaxed, warmed‑up vocalist produces fewer involuntary noises. Encourage gentle humming and tongue twisters before the session.
  • Recording multiple takes: If a particular section is plagued by noises, do another take. A clean performance is always easier to edit.

Even with the best preparation, some editing will likely be needed. The next sections cover the most effective correction techniques.

Manual Editing in Your DAW

The oldest and most precise method remains manual editing. Modern digital audio workstations (DAWs) like Logic Pro, Pro Tools, Cubase, and Ableton Live make it straightforward. Here’s how to do it efficiently:

Zoom and Identify

Zoom into the waveform until you see individual transients. Mouth clicks appear as sharp vertical spikes, often followed by a shorter tail. Breaths have a more rounded, lower‑amplitude shape. Listen at a moderate level to distinguish them from the vocal.

Clip Gain or Fading

For mouth clicks, the fastest technique is to draw a short fade (2–5 ms) over the spike. In most DAWs you can use the fade tool, or cut the region and apply a crossfade of minimal length. For louder clicks, reduce the clip gain by 6–12 dB. For breaths, select the region containing the breath and either delete it (if completely unwanted) or reduce its volume by 6–12 dB with a gentle fade‑in and fade‑out to avoid abrupt silence.

Automation Lanes

Volume automation is ideal for breaths that you want to keep but lower. Draw a dip that starts just before the breath and returns to full volume after. This preserves the natural attack of the following note or word. For repetitive breaths (e.g., between every phrase), creating an automation lane takes a few seconds and yields a much more natural result than a global gate.

Strip Silence and Batch Editing

Many DAWs offer a “strip silence” function that detects regions below a threshold. Adjust the threshold to isolate silences and breaths, then apply a uniform gain reduction to those regions. This is faster than manual editing but less precise — always check the results carefully.

Manual editing remains the gold standard for quality because it gives you 100% control over what stays and what goes. Its drawback is time: a three‑minute vocal can take 30–60 minutes to clean thoroughly. That’s where plugins come in.

Plugin‑Based Solutions

Several companies develop dedicated plugins for removing mouth noises and breaths. They save hours of work while producing results that often rival manual editing. Here are the most trusted options:

  • iZotope RX (De‑click, De‑mouth, De‑breath): RX is the industry standard for audio repair. Its De‑click module targets clicks and pops, while the dedicated De‑mouth module (available in RX Advanced) specifically attenuates lip smacks and saliva sounds. The De‑breath module can automatically detect and attenuate or delete breaths. RX’s spectral tools also allow visual cleanup. Learn more about iZotope RX.
  • Waves DeBreath: A simple, elegant plugin that automatically finds breaths and lets you set a target level. It analyzes the vocal track and applies real‑time reduction. It’s fast and effective for spoken word. Details on Waves DeBreath.
  • Accusonus ERA Mouth De‑click: Part of the ERA bundle, this plugin uses machine learning to detect and remove mouth clicks with a single knob. It works well as a first pass but may need manual follow‑up for heavy noise.
  • Waves NS1 / WLM Plus: While primarily noise reduction tools, they can be applied to breath reduction when set conservatively. They’re less precise than dedicated breath tools.

When using plugins, always work on a copied track or use “Apply” (rendering) rather than real‑time processing to avoid artifacts. Over‑processing can cause metallic artifacts, smearing of sibilants, and a “sucking” sound.

Spectral Editing: See the Noise, Kill the Noise

Spectral editing takes visual editing to another level. Instead of viewing a waveform, you see a spectrogram — a time‑frequency heatmap where mouth clicks appear as vertical lines (broad‑band transients) and breaths as diffuse clouds of energy, often concentrated in the 200 Hz–2 kHz range. Tools like iZotope RX Spectral Editor, Adobe Audition’s Spectral Frequency Display, or the built‑in spectrogram in Pro Tools allow you to:

  • Select and delete: Draw a rectangle around a click or a breath region and silence it. Because you can see exactly which frequencies are affected, you avoid cutting into the voice.
  • Attenuate instead of delete: Most spectral editors let you lower the gain of a selection rather than cut it entirely, preserving the natural air surrounding the noise.
  • Process in the frequency domain: For clicks that occur on top of a vowel, spectral editing is far safer than broadband gain reduction.

Spectral editing is the ultimate precision tool, but it requires a trained ear and eye. Over‑use can create audible gaps or “sucking” effects. Use it as a surgical supplement to manual edits and plugins.

Advanced Workflow: Combining Techniques

In practice, most professional engineers combine several techniques:

  1. Prevent and record cleanly — start with a good take.
  2. First pass with a de‑click / de‑mouth plugin — set it aggressively but listen for artifacts. Apply only to problem sections, not the whole track.
  3. Second pass manual editing — zoom in to handle any remaining obvious clicks or smacks.
  4. Breath management using a de‑breath plugin or automation — reduce all breaths to a consistent, low level (e.g., −12 dB below the vocal average). For songs, you might leave some breaths for emotional impact.
  5. Final spectral check — use a spectrogram to spot any stray artifacts you might have missed audibly.
  6. Re‑integrate with context — listen to the vocal in the full mix. Sometimes a breath that sounds loud in solo is completely masked by instruments. Don’t over‑process.

This layered approach balances speed with quality. For a highly demanding project (e.g., a lead vocal in a pop ballad), you may need to devote an entire editing session to mouth and breath cleanup. For a podcast, a plugin‑only approach may suffice.

Best Practices and Common Mistakes

Even experienced engineers can over‑correct. Here are guidelines to maintain natural, professional results:

  • Never remove all breaths. A complete lack of breath sounds robotic and unnatural. Aim to reduce them to a consistent low level, not delete them entirely.
  • Work in a quiet monitoring environment. If your headphones or room have low‑end rumble, you may miss subtle clicks. Use closed‑back headphones with a flat response.
  • Check in mono and with low volume. Mouth noises can be masked by stereo width or by listening at high levels. A low‑volume mono check reveals them.
  • Save a backup of the raw vocal. Non‑destructive processing (e.g., using automation or clip gain) is always preferable to destructively altering the audio file.
  • Don’t rely solely on plugins. Every plugin introduces its own character. Over‑processing can make a vocal sound “processed” — hollow, without life. Always have manual editing as a fallback.

One common mistake is treating every mouth click as equal. Some are harmless and even natural. Develop your ear for which noises matter. A small click on an unstressed syllable might go completely unnoticed by listeners, while one on a held note can be jarring.

Conclusion

Removing mouth noises and breaths is both a technical skill and an artistic decision. The goal is not a sterile, lifeless vocal but one that sounds effortlessly pure — as if the singer were standing right next to you with perfect mic discipline. By understanding the origins of these noises, preventing them at the source, and applying a targeted combination of manual editing, intelligent plugins, and spectral tools, you can achieve clean, professional vocals that elevate any track. Experiment with each technique, trust your ears, and always preserve the emotional core of the performance.

For further reading on advanced vocal cleanup, consult Sound on Sound’s vocal editing guide or explore the iZotope learning resources on vocal repair. With practice, removing mouth noises and breaths becomes a fast, almost automatic step in your mixing workflow.