How to Remove Breathing from Audio: 5 Methods That Actually Work

Distracting breath sounds are one of the most common problems in recorded audio, and one of the easiest to fix once you know the right approach. Whether you’re editing a podcast, voiceover, or audiobook, you have several routes available — from free manual techniques to professional AI-powered tools. This guide covers five proven methods and helps you choose the right one for your recording, skill level, and budget.

How to Remove Breathing from Audio: 5 Methods That Actually Work

Why Breathing Sounds Ruin Otherwise Clean Audio

Not every breath in a recording is a problem. The issue arises when audible inhales, heavy exhales, or chest and nose breath artifacts appear close to the microphone signal — loud enough to compete with the voice itself. This is especially common in close-mic setups used for voiceover work, podcasts, and audiobook production, where the microphone is positioned to capture maximum vocal detail.

Why Breathing Sounds Ruin Otherwise Clean Audio

Even a technically clean recording can feel amateur if a listener hears a sharp inhale before every sentence. In a short YouTube video, one or two noisy breaths between lines can distract from the message. Across a 40-minute podcast episode, dozens of breath artifacts accumulate into a listening experience that feels unpolished and tiring.

The good news is that removal is reliable and repeatable. The methods below range from zero-cost manual fixes to fully automated AI processing — and each serves a different editing scenario.

Method 1 — Manual Editing: Cut or Silence Breath Regions

Manual editing is the foundational method. It works in any DAW or audio editor — Audacity, Logic Pro, Reaper, Adobe Audition, GarageBand — and requires no plugins or paid software.

Steps:

1. Import your audio file and open the waveform view.

   

2. Play through the recording and scrub along the timeline to locate breath regions by ear.

3. Zoom in on the waveform to confirm the breath region’s exact start and end points.

4. Select the breath region using your DAW’s selection tool.

   

5. Either silence the selection (reduce gain to zero) or delete it and close the gap, depending on your timing needs.

   

6. Apply a short fade-in (10–30 milliseconds) at the beginning of the next audio segment to prevent an audible click at the edit point.

   

7. Play back the edit and adjust the selection boundaries if the transition sounds unnatural.

8. Repeat for each breath region in the session.

This method gives you complete control over every edit and produces clean, artifact-free results when done carefully. It is the right choice for short recordings with a handful of egregious breaths, or as a finishing pass after a faster automated method has handled the bulk of the work.

The honest limitation: for a one-hour podcast episode with hundreds of breath events, manual editing is not scalable. Use it selectively.

Tips for Identifying Breath Sounds in a Waveform

• Breaths appear as low-amplitude, irregular noise bursts — noticeably different from the flat silence of a quiet room and from the clean transients of consonants

• They typically occupy the full frequency spectrum at low levels, which distinguishes them from narrow-frequency noise like hiss or hum

• Zoom in to a 1:2 or 1:4 view ratio to see individual breath events clearly without confusing them with room noise

• Listen while watching the playhead to train your eye to recognize patterns specific to the speaker

Method 2 — Noise Gate: Automate Suppression Between Speech

A noise gate is a processing tool that automatically reduces the volume of any audio signal that falls below a set threshold level. When configured correctly, it silences breathing and room noise between words and phrases without requiring you to make individual selections.

How it works: You set a decibel threshold, and any signal below that threshold is attenuated or fully silenced. A noise gate is not intelligent — it does not distinguish between breath sounds and soft speech. Its effectiveness depends entirely on there being a meaningful amplitude difference between the speaker’s voice and their breath sounds.

Key parameters to configure:

Start with a threshold around -40 dB and move it upward in small increments until the breaths are suppressed. Then adjust attack and release times until transitions between gated and ungated audio sound natural.

Noise gates work best on recordings where breath sounds are relatively consistent in volume and the speaker’s voice is significantly louder. The key limitation is that aggressive threshold settings will clip soft word onsets — a word whispered or trailed off quietly will fall under the threshold and be cut. Always audition the full recording after applying a gate before committing the output.

Method 3 — iZotope RX: The Professional Standard for Breath Removal

iZotope RX is the industry reference tool for audio restoration, and its Breath Control module is purpose-built for this specific problem. Rather than using amplitude thresholds like a noise gate, Breath Control identifies breath events by their spectral signature — the characteristic frequency content of a breath sound — which makes it far more accurate and far less likely to cut soft speech.

Steps to use iZotope RX Breath Control:

1. Open your audio file in iZotope RX (standalone application or as a plugin within your DAW via RX Connect).

   

2. Navigate to the Breath Control module in the module browser.

3. Click Preview to let RX analyze the audio and detect breath events automatically. Review the result, then adjust the settings if needed before applying the effect. 

4. Adjust the Sensitivity slider to increase or decrease the number of detected events — increase it if breaths are being missed, decrease it if speech onsets are being flagged incorrectly.

5. Set the Target Level slider; a value between 60–80% typically produces natural-sounding suppression without creating dead-air artifacts between words.

6. Click Preview to audition the result on a selected region.

7. Click Render to apply the correction to the full file or selection.

   

The Learn function dramatically reduces setup time because it adapts detection sensitivity to the specific audio file rather than requiring you to find a threshold by trial and error.

Pricing note: Breath Control is included in iZotope RX Standard (approximately $99 on sale) and RX Advanced. It is not available in RX Elements, the entry-level tier. iZotope offers a free trial of RX Standard, which is worth running on a problem file before committing to a purchase.

For professional voiceover work, audiobook production, or any scenario where the breath removal needs to be both thorough and inaudible, iZotope RX Breath Control is the benchmark against which all other methods are measured.

Method 4 — Adobe Audition: Built-in Tools for Breath Cleanup

Adobe Audition users have two built-in paths for handling breath sounds, both available without additional plugins or subscriptions beyond Creative Cloud.

Path 1 — Noise Reduction Process (for background breath-level noise):

1. Find a region of audio that contains only breath sounds and no speech — a pause between sentences works well.

2. Select that region and go to Effects > Noise Reduction/Restoration > Capture Noise Print.

   

3. Select the full audio file or the section you want to process.

4. Go to Effects > Noise Reduction/Restoration > Noise Reduction (process).

   

5. Set the Noise Reduction percentage to 60–80%; preview the result and adjust before applying.

6. Click Apply.

   

This method works best when breath sounds have a consistent tonal quality that was captured in the noise print. It is most effective at reducing underlying breath noise across a recording rather than eliminating individual sharp inhale events.

Path 2 — Auto Heal for isolated breath removal:

1. Select a specific breath region in the waveform view.

   

2. Press Shift+Delete to trigger Auto Heal, which replaces the selected region with interpolated audio based on surrounding content.

Auto Heal is fast and works well for isolated, short breath events that sit in otherwise clean audio. It is less reliable on longer or more complex breath sounds.

Adobe Audition is a strong option for editors already in the Adobe Creative Cloud ecosystem who do not need the surgical precision of iZotope RX. For most podcast-level cleanup, combining the Noise Reduction process with selective Auto Heal passes handles the majority of breath artifacts effectively. For complex or dense breath events in professional productions, iZotope RX remains more precise.

Method 5 — AI and Online Tools for Fast, Automated Removal

For creators who want fast results without learning a DAW or purchasing dedicated audio software, AI-powered online tools offer a low-friction upload-process-download workflow. These tools process files in the cloud and return cleaned audio in minutes.

Cleanvoice AI is specifically designed for podcast post-production. Its detection engine targets breath sounds, mouth noise, and filler words as separate, individually controllable categories, making it one of the most targeted tools in this group for breath-specific removal.

Auphonic takes a broader approach, combining loudness normalization, noise reduction, and breath-noise processing in a single automated pass. It is well-suited for creators who want to handle multiple quality issues simultaneously rather than addressing breath sounds in isolation.

Descript Studio Sound is the relevant choice for creators already using Descript for transcription-based editing. It applies AI audio enhancement — including breath reduction — as a single toggle within the Descript editor, which keeps the workflow consolidated without switching between applications.

All three tools produce noticeably improved output on moderately noisy recordings. High-complexity breath events or recordings with audio quality issues beyond breath artifacts may still require a DAW-based pass after AI processing.

How to Choose the Right Method for Your Situation

The decision is not about which tool is best in the abstract — it’s about matching the method to your content volume, available tools, required output quality, and the amount of time you can invest. Consider four variables: whether you need a quick fix or a scalable workflow, whether you have DAW access and editing experience, what budget you’re working with, and how precise the final output needs to be.

If you are just starting out, the manual method in Audacity costs nothing and teaches you to hear and see breath artifacts accurately — skills that make every subsequent method easier to apply well. If you produce high volumes of content regularly, investing in iZotope RX or a dedicated AI tool will save more time than it costs within a few editing sessions.

Reduce Breathing Artifacts Before You Even Edit (Pro Tip)

The most efficient breath removal happens before the recording reaches your editor. A few simple capture-stage habits reduce the number of artifacts you have to handle in post.

Reduce Breathing Artifacts Before You Even Edit (Pro Tip)

• Position the microphone slightly off-axis from the direct path of the nose and mouth to reduce breath direction onto the capsule

• Practice controlled, low-volume breathing between sentences rather than releasing full exhales toward the mic

• Use a pop filter to add physical distance and reduce the impact of breath blasts on the capsule

• Choose hardware with built-in noise processing: the Hollyland LARK MAX 2, for example, includes AI Noise Cancellation that suppresses ambient and breath-adjacent noise at the capture stage, which reduces the volume of artifacts that reach your editing timeline in the first place

Upstream habits do not replace editing, but they reduce the scale of the problem before it starts.

FAQ

Q1: Will removing breathing affect how natural my voice sounds?

Well-tuned tools preserve natural speech when used conservatively. In iZotope RX, keeping Reduction Depth below 80% typically maintains natural speech onsets. The main risk of an unnatural result is over-processing: removing every trace of breath creates a sterile, airless sound. A light hand produces cleaner audio without making the recording feel artificially hollow.

Q2: What is the best free tool to remove breathing from audio?

Audacity is the most capable free option. Use the manual cut method for isolated breaths or apply the built-in Noise Gate effect (Effect > Noise Gate) for recurring breath events across a longer recording. Among AI online tools, both Auphonic (2 free hours per month) and Cleanvoice AI (limited free minutes) offer usable free tiers for smaller projects.

Q3: How do I remove breathing from audio in Audacity specifically?

For manual removal: select the breath region in the waveform view, then use Effect > Silence Audio to mute it. For automated suppression: go to Effect > Noise Gate, set the threshold to approximately -40 dB, adjust attack and release, and preview before applying. Adding short fades at edit boundaries prevents audible clicks between silence and speech.

Q4: Should I remove every breath or only the most intrusive ones?

Context determines priority. Voiceover work and audiobooks benefit from thorough breath removal because the format demands an uninterrupted listening experience. Podcasts can retain light, natural-sounding breaths — full removal across a conversational episode can feel clinical. Start with the loudest, most disruptive mid-sentence breaths, then decide whether quieter breaths warrant the additional editing time.

Q5: Can I batch-process a full episode or multiple files automatically?

Yes. iZotope RX supports scripted batch processing through its Actions feature, allowing you to apply Breath Control settings across multiple files without manual intervention. Auphonic accepts batch file uploads through its web interface. Cleanvoice AI processes files one at a time in its basic interface but supports RSS feed integration for automatic podcast episode processing. Manual editing and single-file noise gate setups do not scale to batch workflows.

Next Steps

For most recordings, the workflow is straightforward: use manual editing or a noise gate for quick fixes, step up to iZotope RX or Adobe Audition when precision matters, and reach for Cleanvoice AI or Auphonic when speed and simplicity are the priority. Start with the tool already in your workflow and apply the matching method from the table above. Once breath removal is handled, the same skills transfer directly to related editing tasks.

Related reading: - [How to reduce background noise in audio] - [Best podcast editing software] - [Voiceover recording tips for clean, studio-quality takes]