Voice Over Editing for Beginners: Removing Breaths, Fluffs, and Silences

Voice over editing is half the job. The recording is the easy part. The editing is what separates broadcast-ready audio from amateur. A 10-minute audiobook chapter might have 8 minutes of usable voice over in the raw recording. The other 2 minutes are breaths, mouth clicks, false starts, and silences that need to come out.

This guide covers the 5 things to remove from every voice over recording, the free and paid DAWs that work for editing, the manual workflow that pros use, and the AI-powered alternatives that turn 45-minute editing sessions into 90 seconds.

The 5 Things to Remove From Every Voice Over Recording

Every raw voice over recording contains the same five categories of imperfection. Removing them is what turns a recording into a deliverable.

1. Audible Breaths

Voice actors breathe between phrases. Some breaths are inaudible (a quiet inhale through the nose). Some are loud gasps that the mic captures clearly. The loud ones distract listeners and need to go.

Don't remove every breath. A natural-sounding read keeps small inhales between phrases. Remove only the gasps and the dramatically audible ones.

2. Fluffs, False Starts, and Restart Words

Stumbles, restarts, and 'sorry let me try that again' moments. The voice actor restarts the line. The good take and the bad take are both in the file. Edit the bad take out.

Look for: 'um,' 'uh,' stutters, restart words, and any take where the voice actor obviously broke character. Use the keeper take only.

3. Mouth Clicks and Lip Smacks

Saliva-driven clicks, tongue clicks, lip smacks. These show up as quick high-frequency spikes in the waveform. Most are at -50 to -55 dB. ACX and other professional spec wants them gone.

Manual removal is tedious. iZotope RX has a dedicated 'Mouth De-Click' module that catches most of them in one pass. Worth the subscription for serious audiobook narrators.

4. Long Silences and Excessive Pauses

When a voice actor pauses between sentences or paragraphs to think, those silences end up in the recording. A 4-second pause where the actor was reading ahead becomes 4 seconds of dead air in the final file.

Trim silences to natural lengths: 0.3 to 0.5 seconds between sentences, 0.7 to 1.0 seconds between paragraphs (in audiobook narration). Don't strip silences entirely. Trim them.

5. Environmental Noise and Mic Bumps

A chair creak. A page turn. A neighbor's car door. A mic bump from adjusting the boom. Anything that's not voice over and not intentional needs to come out.

Use the noise reduction or spectral repair tools in your DAW. Avoid over-processing, which introduces artifacts that are worse than the original noise.

DAWs to Use (Free and Paid)

Any DAW can edit voice over. For beginners, three options cover 95% of needs.

Audacity (Free)

Free, cross-platform, fine for basic VO editing. Lots of audiobook narrators use Audacity for their entire career. Interface is dated but functional.

Strengths: free, simple, gets the job done. Weaknesses: clunky workflow for complex edits, no native AI tools, slow on long files.

Reaper ($60-225)

The working narrator's secret weapon. Cheap (free trial that never expires, $60 personal license), fast, infinitely customizable. Most experienced VO editors recommend it.

Strengths: cheap, fast, deep customization, great for keyboard shortcuts. Weaknesses: steep learning curve, ugly default look (skinnable if you care).

Adobe Audition ($21/month)

Industry standard for many commercial voice actors. Tight integration with Adobe Creative Cloud. Excellent spectral editor for removing mouth clicks.

Strengths: pro features, polished interface, spectral repair. Weaknesses: subscription model, overkill for basic VO.

The Manual Voice Over Editing Workflow

If you're editing manually, here's the workflow most working VOs use. Total time: roughly 2-3 hours of editing per finished hour of audio.

1. Listen through once. Note where the obvious problems are. Take and finished version markers.
2. Remove fluffs and false starts. Cut the bad takes. Keep the keepers. This is the biggest time savings.
3. Tighten silences. Trim long pauses to natural lengths. Don't strip everything.
4. Remove audible breaths. Skip the natural-sounding ones. Cut the gasps and louder inhales.
5. De-click mouth noise. Use a dedicated tool or scan for spikes in silent areas.
6. Apply noise reduction sparingly. Only if the noise floor is too loud. Over-processing sounds worse than light noise.
7. Normalize to your loudness target. ACX, podcast, broadcast. See the loudness guide for spec details.
8. Final listen-through. One pass with fresh ears. Catch anything you missed.

AI-Assisted Editing (The 2026 Workflow)

The traditional manual workflow takes 2-3 hours per finished hour. AI tools introduced in the last two years can do the breath removal, silence trimming, and click cleanup in 90 seconds. Working voice actors now use AI for the heavy lifting and reserve manual editing for the fine details.

The standard 2026 VO editing workflow looks like this:

1. Run the raw recording through a best-take editor. Modern tools like Auto Clean Up keep the best version of every line and drop the fluffed retakes automatically. No manual selection needed.
2. Spot-check the output. Listen to a 30-second sample. Confirm the edit kept the right takes and didn't cut anything it shouldn't have.
3. Fine-tune in your DAW if needed. For anything requiring a judgment call the tool flagged as uncertain, do a quick manual pass.
4. Apply noise reduction and de-click as needed. Often the AI tool handles light de-clicking, but heavier processing still benefits from manual review.
5. Normalize to your spec target. Use a tool with built-in presets so you don't have to remember the numbers.

Total time with AI-assisted workflow: 15-25 minutes per finished hour. Compared to 2-3 hours manually, that's a 6-12x speedup.

Common Voice Over Editing Mistakes

What goes wrong in beginner voice over edits.

1. Removing every breath. Some breaths are part of natural speech. Strip them all and the read sounds robotic.
2. Over-normalizing. Pushing the audio too loud (above -18 dB RMS) introduces compression artifacts. Stay in spec.
3. Heavy noise reduction. Aggressive noise reduction adds artifacts that are worse than the original noise. If your noise floor is too loud, fix the room, not the file.
4. Tight crossfades on hard cuts. When you remove a section between phrases, use a 5-10 ms crossfade to avoid audible clicks. Hard cuts without crossfades often pop.
5. Editing while tired. Editing requires fresh ears. Don't edit a 90-minute audiobook chapter at the end of an 8-hour recording day. Sleep, then edit.
6. Not listening through the final version. Always do one full listen-through after editing. You'll catch cuts that sound wrong, silences that are too short, and clicks you missed.

How Editing Differs by Project Type

Different project types have different editing standards.

• Audiobooks. Strip aggressively. Listeners hear hours of your voice. Every distraction compounds. ACX has strict spec requirements.
• Commercials. Tight, punchy edits. Remove everything that doesn't sell. Loudness usually higher than audiobook spec.
• E-learning. Clarity-focused. Trim long pauses but keep natural ones. Sync to picture if there's video.
• Podcasts. Lighter touch. Listeners expect some natural breathing and pacing. Don't over-edit.
• Animation and character work. Don't remove breaths if they're character-related (a panting villain, a winded hero). Edit performance, not breath.

Editing Is Where Voice Actors Save Their Careers

A good performance with bad editing sounds amateur. A mediocre performance with great editing sounds professional. Most voice actors who don't book work have the performance but not the post-production. Tightening up your edits is often a faster path to better bookings than working on your read.

Once your editing workflow is dialed in, the next things to lock in are hitting your loudness spec and getting your home studio set up properly. Good editing on a good recording in a good room is the full stack of professional voice over.