How to Remove Background Noise From a Video on Mac

The fastest way to remove background noise from a video on Mac is to select the clip in iMovie, click the Noise Reduction and Equalizer button, and raise the Reduce background noise slider while playing it back. Here's how far to take that, what the equalizer adds, and which noises won't respond at all.

Open the free Background Noise Remover →

Remove background noise from a video on Mac with iMovie

Select the clip in the timeline and click the Noise Reduction and Equalizer button to show the audio correction controls. The key one is Reduce background noise, and Apple describes what it's doing precisely: it reduces the noise as a percentage of the original clip's sound, where 0% is no reduction and 100% is the maximum, without pulling the overall volume down with it.

The workflow that matters is the one Apple recommends and most people skip: play the clip back, then refine the slider position. Noise reduction is not a value you set once. It's a value you audition, because the point where the noise stops being distracting and the point where the voice starts sounding processed are usually only a few percent apart, and where those points sit depends entirely on the recording.

There's a second control worth reaching for before you push the first one higher. The Equalizer pop-up holds presets for things like enhancing vocal quality or boosting bass and treble. Lifting the frequencies the voice lives in is frequently more effective than suppressing the ones the noise lives in, because it improves the ratio between them without removing anything.

Three noises, and only one of them answers to a slider

Background noise gets treated as one category and it isn't.

A continuous hiss , air conditioning, a computer fan, distant traffic , sits underneath the whole recording at a fairly constant level. This is what noise reduction is designed for and where the slider earns its keep.

A one-off event , a door, a horn, a chair scraping , is a brief spike. A tool tuned to remove a steady background has no particular reason to catch it, and the practical fix is usually editing around it rather than processing it.

Room reverberation is the hardest of the three, because it isn't layered on top of the voice at all. It's the sound bouncing off hard surfaces before it ever reaches the microphone, meaning it's already part of the recorded voice. No percentage slider separates them cleanly, and this is the case where re-recording in a softer room beats any amount of processing.

Naming which one you have takes ten seconds on headphones and saves an hour of dragging a slider that was never going to help.

A model that knows what a voice looks like

Montaj runs a neural denoising model on your device, tuned specifically for voice rather than for audio in general. That specialisation is the difference: a model that knows what speech looks like can hold onto it while removing what surrounds it, instead of treating everything below a threshold as noise.

The audio never leaves your machine to be cleaned, and you can compare the result against the original before committing. If you need to remove background noise from a video on Mac without handing the recording to a server, that's the shape of it.