To remove background noise from a video on a Mac, run noise reduction first, then loudness normalization to around -14 LUFS, then de-essing. Order matters: normalizing before you denoise just makes the noise louder too. Editors like Zella clean hum, hiss and room tone and balance the voice entirely on-device, so the footage never uploads.
Background noise - a fan, traffic, a buzzing fridge, room echo - makes a video feel amateur faster than any picture problem does. Viewers judge sound first.
The good news: steady background noise is exactly what automated reduction handles best, so one pass on a Mac usually fixes it without re-recording. The fastest route is an on-device AI cleanup that strips hum and hiss, evens out loudness, and never uploads your audio.
Below: the one-pass method, the iMovie and Audacity routes, and how to capture cleaner sound in the first place.
Remove background noise in one pass (the fast way)
If you want clean audio with the least fuss, run an AI cleanup pass in Zella:
- Open or record your clip in Zella.
- In AI Cleanup, apply noise reduction to remove steady hum and hiss.
- Run Polish Voice to normalize loudness to around −14 LUFS, de-ess harsh "s" sounds, and gently compress so quiet and loud moments sit evenly.
- Listen on headphones and adjust the amount - the goal is natural, not over-processed.
It all runs on-device, so even sensitive recordings never leave your device. See AI cleanup for the full breakdown. Everything is non-destructive, so you can preview and undo freely.
If the terms are new, Apple's audio fundamentals covers levels and headroom, and why hitting the right peak does not by itself make a mix sound right.
How to remove background noise in iMovie
iMovie ships free with every Mac and has a built-in noise reducer that's fine for light cleanup:
- Drop your clip into the timeline and select it.
- Click the Noise Reduction and Equalizer button above the viewer.
- Check Reduce background noise and drag the slider to set the amount.
- Optionally pick an Equalizer preset - Voice Enhance, Music Enhance, or Hum Reduction.
It's a single broadband slider, so it can muffle your voice if you push it. iMovie won't remove every sound, and it has no loudness normalization, so multi-clip projects still sound uneven.
How to remove background noise with Audacity (free)
Audacity is free and open-source, and its Noise Reduction effect is more controllable than iMovie's:
- Extract or import the audio into Audacity.
- Select a few seconds of pure room tone (no speech).
- Open Effect ▸ Noise Reduction and click Get Noise Profile.
- Select the whole track, reopen the effect, and apply.
It works well but means juggling separate audio and video files and re-syncing on export - more steps than an all-in-one editor.
Which method should you use?
| Method | Cost | Where it runs | Loudness normalize | Best for |
|---|---|---|---|---|
| Zella AI Cleanup | Free | On-device | Yes (−14 LUFS) | Clean voice in one pass, privacy |
| iMovie | Free | On-device | No | Quick light cleanup |
| Audacity | Free | On-device | Manual | Granular control, audio-only |
| Online AI tools | Free tier | Cloud upload | Varies | No-install browser quick fix |
The honest trade-off: online cloud tools are convenient but upload your audio to a server, and most cap free output or watermark it. A native on-device pass keeps private content private and skips the upload-wait-download loop entirely.
Types of noise and what reduction can (and can't) fix
Not all noise is equal, and knowing your type tells you whether to reach for the reduction slider or the cut tool.
| Noise type | Examples | Best fix |
|---|---|---|
| Steady broadband | Fan, AC hum, computer whir, tape hiss | Automated noise reduction |
| Room reverb | Echo in a hard, empty room | Reduction softens it; deader room is the real fix |
| One-off transients | Door slam, cough, notification ding | Cut it on the timeline |
| Wind / crowd chatter | Outdoor gusts, café murmur | AI reduction helps; close-mic capture helps more |
Steady, broadband noise is what reduction handles best, because the tool models a constant noise floor and subtracts it. One-off transients aren't really "noise" in that sense - the clean fix is to cut them on the timeline.
Prevent noise at the source (so cleanup does less)
The cheapest noise reduction happens before you hit record:
- Get the mic close to your mouth (a hand's width) and keep gain moderate to avoid clipping.
- Soften the room with rugs, curtains, or a closet of clothes to kill echo.
- Kill steady sources - fans, AC, a noisy laptop - for the take.
- Record 3-5 seconds of silence so you (and the tool) can hear the room's noise floor.
- Pick the quietest room and the quietest time of day.
The cleaner the input, the more natural the output - reduction is a fix, not a substitute for decent capture.
Match loudness across clips
Clips recorded at different times almost never sit at the same volume, and the jump between them is jarring. Polish Voice normalizes each one toward a consistent target of around −14 LUFS, so the whole video plays evenly and nobody reaches for the dial.
That matters most on compilations, multi-take tutorials, and anything with inserted B-roll or music. Noise reduction targets the noise floor; normalization brings your voice back to a clear, consistent level. Neither one replaces the other.
Pair tone fixes with pacing fixes
Noise reduction fixes how your audio sounds. It doesn't fix how it flows.
For a genuinely tight result, also remove silences and filler words in the same pass. Clean tone plus tight pacing is the whole difference between a produced video and a raw one, and both are one click. From there, add captions so the message lands on muted feeds too.
Common mistakes to avoid
- Over-reducing. Too much noise reduction makes voices sound underwater or robotic. Use the lightest setting that works.
- Fixing tone but ignoring loudness. Normalize so your video matches platform levels.
- Skipping headphones. Laptop speakers hide problems you'll only hear on earbuds.
- Uploading sensitive audio to a web cleaner. Keep it local if the content is private.
Free vs Pro in Zella
Noise reduction and Polish Voice are both in the free tier, so nothing in this guide is gated.
The free plan also covers unlimited recording, no watermark, 1080p export (up to 2 minutes free; Pro for longer), captions, and auto-zoom - all on-device, no account required. The optional one-time $89 Pro unlock adds unlimited export length, 4K export, and the pro-depth toolset: 4K/8K export, motion keyframes, Color Match and scopes, effect masks, speed ramps and per-drum Beat Sync.
Frequently asked questions
Is my audio uploaded for processing? No - noise reduction and Polish Voice run entirely on-device, with no connection needed, so confidential recordings never upload.
Will it make my voice sound robotic? Not if you keep it gentle - the defaults are conservative and you control the amount.
Can it remove echo or a one-off door slam? It reduces echo (a deader room is the real fix), but a single transient like a door slam is best cut on the timeline.
Does it work on a podcast, call, or screen-share too? Yes - any audio or video you import. It's tuned for voice, so apply it lightly on music-heavy clips.
The bottom line
To remove background noise from a video on a Mac, apply on-device AI noise reduction and Polish Voice to clean hum, hiss, and reverb and even out loudness - iMovie and Audacity work for lighter or audio-only jobs, but an all-in-one local pass keeps your audio private and tightens pacing in the same step. Capture clean audio where you can, and let reduction handle the rest.
Download Zella and clean up your sound.
Make your next video with Zella.
Record, edit and ship on Mac and iPhone - local, private, free to start.
RELATED