How to Remove Mouth Clicks from Vocals

Flat illustration of a studio microphone beside a voice waveform with small click spikes being picked out, showing how to remove mouth clicks from vocals

To remove mouth clicks from vocals, find each click by ear and on a zoomed waveform, then turn it down or cut it out by hand in an audio editor, or use a dedicated de-clicker for long takes. A Noise Reducer comes after that, not instead of it: a noise reducer is built for steady background hiss and room tone, and a click is a split-second tick inside the voice itself. The surest fix is the one before you record: water, a little lip balm, and a mic that isn't pressed against your mouth.

What mouth clicks are (and what they aren't)

Mouth clicks are the small wet ticks and smacks your mouth makes when you talk: saliva sticking and unsticking, lips parting, the tongue lifting off the roof of the mouth. You never notice them in conversation. A close microphone catches every one, and in a quiet podcast or an exposed vocal they make a voice sound "wet".

Two other vocal problems get mistaken for them, and each needs a different fix:

  • Plosives are low thumps from a burst of air on "p" and "b". Their energy sits mostly below about 150 Hz, so a low cut handles them. See How to Remove Plosives and Pops.
  • Sibilance is the harsh hiss on "s" and "sh", roughly 4–10 kHz depending on the voice. That's a de-esser's job; How to De-Ess Vocals With an Audio Equalizer covers it.
  • Mouth clicks are neither low nor confined to one band. A click lasts a few milliseconds and spreads its energy across a wide range of frequencies at once.

That's why an EQ cut doesn't work on clicks: there's no single band to pull down. Cut the highs in the Equalizer and the clicks get slightly duller while the whole voice loses its air.

How to spot mouth clicks

Start with headphones; laptop speakers hide clicks. They're most obvious in the gaps just before a sentence starts, when the mouth opens to speak.

Then look. On a waveform zoomed in to a second or two, a click is a thin, sharp spike that doesn't match the words around it. Some sit in pauses; others hide inside words.

The Spectrogram makes them easier to find. Speech shows up as stacked horizontal bands that bend and drift. A click is a thin vertical streak cutting up through many frequencies at one instant. Hiss is a flat haze; a plosive is a short blob along the bottom. How to Read a Spectrogram explains the axes if it's new to you.

What you'll need

  • The recording, in a format your browser or editor can open (WAV, MP3, M4A, FLAC).
  • Closed headphones.
  • For surgical removal: an audio editor or DAW with a zoomable waveform. Audacity is a free option.
  • Vocal Cut's tools for the edges and the cleanup: the Audio Cutter, Audio Joiner, Audio Fade and the Noise Reducer. Everything in Vocal Cut runs in your browser, and your audio never leaves your device.

Step 1: remove mouth clicks by hand

Manual editing is slow, but it's the most accurate method. In a DAW or editor:

  1. Zoom in on the click. Get close enough that the click is a visible spike, usually a few milliseconds wide.
  2. Select just the click. Keep it tight. Grabbing part of the next consonant is how edits become audible.
  3. Choose a fix.
    • Lower its gain. Turning the selection down by 10–20 dB, or using clip gain, often makes it disappear without leaving a hole.
    • Replace it with room tone. If the click sits in a pause, paste in a matching length of quiet room sound from the same recording.
    • Redraw it. Many editors have a pencil tool for redrawing a few samples by hand, which works on very short clicks inside a word.
  4. Cut at zero crossings and crossfade. An edit that slices the waveform away from zero creates a new click of its own. Snap cuts to zero crossings and add a crossfade of a few milliseconds at each edge. How to Cut Audio Without Losing Quality explains zero crossings.
  5. Listen back in context. A fix that looks clean can still sound wrong at full speed.

On a ten-minute voiceover that can mean dozens of edits. Fix the obvious ones and stop. Nobody hears the click hiding under a loud word.

Step 2: what you can do in Vocal Cut

Vocal Cut has no per-region editing inside a file and no de-click tool, so you can't select one click mid-sentence and turn it down. Here's what it can do:

  • Trim a click at the start or end. Clicks often sit right before the first word. In the Audio Cutter, move the Trim Start point and Trim End point sliders to keep only what you want, then click Cut & Download. The trim points move in 0.1-second steps, so this works when there's a short gap between the click and the first word. Export as WAV to keep the result lossless.
  • Cut around a click in a pause. Cut the audio before the click and after it as two files, then put them back together in the Audio Joiner. The cut points move in tenths of a second, so only do this in a pause long enough to lose a fraction of a second. The Joiner butts the two files straight together, so listen to the join for a new click.
  • Fade an edge. If a click lands at the very beginning or end of a clip, a short fade in or fade out in Audio Fade can soften it.

For anything inside the flow of speech, you need an editor from Step 1 or a de-clicker.

Step 3: remove background noise once the clicks are fixed

With the clicks dealt with, what's left in a home recording is usually a steady bed of hiss, fan or room tone. Now's the time to remove background noise from audio, which is what the Noise Reducer is built for.

  1. Open the Noise Reducer and drop in your edited file.
  2. Pick Voice for spoken word. It uses an AI speech model (GTCRN), downloaded once (about 0.5 MB) and cached, and outputs mono. Pick Music for sung vocals or anything you want kept in stereo. It uses spectral processing, not AI, with a maximum cut of about 16 dB.
  3. Tick "Also remove 50/60 Hz electrical hum (auto-detected)" only if you hear a hum.
  4. Click Reduce noise (Voice mode) or Reduce noise (Music mode). Long files take a while, and the tool shows progress as it runs.
  5. Play it back and adjust the Original ↔ Denoised slider (it starts at 75%) until the background drops without the voice going thin.
  6. Export as WAV, MP3, or Opus where your browser supports it.

Do the click fixes first. Noise reduction won't reliably remove clicks, because they aren't steady noise. For the general method, see How to Remove Background Noise from Audio.

Automated de-clickers: fast, with a catch

For a long take full of small clicks, a de-clicker saves hours. It scans for transients (short, sharp bursts of sound) that don't fit the surrounding audio and repairs them. Audacity has a Click Removal effect, though it was designed mainly for vinyl crackle and results on mouth noise vary. iZotope RX includes a dedicated Mouth De-click module. Many DAWs have their own.

The catch is that consonants are transients too. A "t", "k" or "ch" is a short, sharp burst, which is exactly what a de-clicker hunts for. Push the sensitivity too high and it softens those consonants, and the voice sounds lispy and blurred. Use the lightest setting that removes the clicks you can hear, check consonant-heavy lines, and fix leftovers by hand.

How to prevent mouth clicks

Every repair costs the voice something. Prevention doesn't:

  • Hydrate. A dry mouth clicks more. Sip room-temperature water before and during the session. A few slices of green apple before recording is a common voiceover booth trick for the same reason.
  • Skip dairy, coffee and sugary drinks beforehand. Milk can leave the mouth coated and sticky, and coffee dries it out. Both make clicks worse for many people.
  • Use lip balm. Dry lips crackle when they part. A thin layer helps.
  • Back off the mic and go slightly off-axis. Clicks are tiny sounds that shrink fast with distance. A few extra centimetres, with the mic angled from above or the side, picks up far less mouth noise. Mic Placement for Clean Voice Recordings shows the setups.
  • Breathe through your nose between lines and open your mouth a beat before you speak, so the "unsticking" happens before the word rather than inside it.
  • Test first. Speak a few lines into the Mic Test and listen back on headphones before a long take.

For a full session setup with the free Audio Recorder, see Record a Voiceover With an Online Voice Recorder.

Troubleshooting

The voice sounds dull or lispy after de-clicking. The de-clicker took consonants with the clicks. Lower its sensitivity, or run it only on the sections that need it.

My edit made a new click. The cut landed away from a zero crossing, or there's no crossfade. Move the edit point to a zero crossing and add a few milliseconds of crossfade.

There's a dead gap where the click was. The background dropped to digital silence. Fill it with room tone from elsewhere in the file.

The Silence Remover didn't catch the clicks in pauses. It isn't meant to. The Silence Remover removes gaps that fall below a level threshold. A click is louder than the silence around it, so it can survive while the quiet around it is removed.

The clicks are everywhere. That usually means a dry mouth and a mic that's too close. Re-recording with water, lip balm and more distance often beats an hour of editing.

Frequently asked questions

What causes mouth clicks in recordings?

Mouth clicks come from saliva, lips and tongue: moisture sticking and unsticking as your mouth opens and closes while you talk. A dry mouth makes them worse, and so do dairy, coffee and nerves. They exist in normal conversation too, but a close microphone in a quiet room picks up every one, which is why podcasts, voiceovers and exposed vocals reveal them.

Can a noise reducer remove mouth clicks?

Not reliably. A noise reducer is designed for steady background noise like hiss, fans and room tone, which stays the same from moment to moment. A mouth click is a split-second transient inside the voice itself, so noise reduction mostly leaves it alone. Fix clicks by hand or with a dedicated de-clicker first, then use a noise reducer to clean the background.

What's the difference between mouth clicks and plosives?

Plosives are low-frequency thumps caused by bursts of air on consonants like "p" and "b", and a low cut or high-pass filter removes most of them. Mouth clicks are very short ticks from saliva and lip movement that spread across a wide range of frequencies. An EQ cut won't fix clicks; they need manual editing or a de-clicker.

Does drinking water help with mouth clicks?

Yes. A dry mouth is one of the main causes of mouth clicks, so sipping room-temperature water before and during a session helps. Many voiceover artists also eat a few slices of green apple before recording and avoid dairy and coffee beforehand. Lip balm helps with lip smacks, and backing slightly away from the mic reduces how much mouth noise it picks up.

Do de-clickers damage the voice?

They can if pushed too hard. A de-clicker looks for short, sharp transients, and consonants like "t", "k" and "ch" are short, sharp transients too. At high sensitivity it can soften them, making speech sound blurred or lispy. Use the gentlest setting that removes the audible clicks, check consonant-heavy lines, and fix any stubborn clicks by hand.