# How to remove silences from a video on Mac

Canonical URL: https://getvidiyo.app/blog/remove-silences-from-video-mac
Last updated: 2026-10-08
Author: Ajay Pawriya (https://getvidiyo.app/about)
Published: 2026-10-08
Category: Talking head

> TL;DR: On a Mac, the fastest way is to let an editor detect quiet stretches in the audio and cut them with a little padding around speech. In Vidiyo, import the clip and type "cut the long pauses". By hand, use any editor's waveform to split and delete each pause, or find them first with ffmpeg's silencedetect filter.

Most raw talking-head recordings have more silence than you'd guess: the beat before you start, the pause while you think, the gap before a retake. Cutting it is the single biggest improvement you can make to pacing. On a Mac you can do it automatically in a minute, or by hand in any editor. Both are below.

## What's the fastest way to remove silences from a video on Mac?

The fastest way is automatic detection: an editor measures where the audio goes quiet, cuts those stretches and keeps a little padding around words. In Vidiyo that's a single request, "cut the long pauses", and the cuts appear on the timeline where you can check or adjust each one.

Doing it by hand works too, and it's free. You look at the audio waveform, split at the start and end of each flat stretch, and delete the gap. On a five-minute recording that's often dozens of cuts, which is why automatic removal exists. Long silent stretches are what editors call [dead air](https://getvidiyo.app/glossary/dead-air).

## How do I remove silences in Vidiyo?

Import the video, then ask the assistant to cut the pauses. Vidiyo measures quiet stretches in the audio on import, transcribes on your Mac for word timings, and removes only pauses where both agree. The whole request is one undo step, so you can compare before and after with ⌘Z.

1. Import your clip, or right-click its folder in Finder and choose **Edit in Vidiyo**.
2. Wait for transcription to finish. It runs locally; English is built in.
3. Type "cut the long pauses", or be specific: "tighten every pause longer than half a second".
4. Play it back. The reply says how much silence was removed and flags any spot it left for you to check.
5. If you want the jumps hidden, ask "punch in on alternate cuts" (see [jump cuts that don't feel jumpy](https://getvidiyo.app/blog/jump-cuts-that-dont-feel-jumpy)).

You don't need chat for this to work, but chat is what makes it one step. Without it, you can split and delete pauses by hand on the same timeline.

## How does Vidiyo decide which pauses to cut?

It cuts only where the audio and the transcript agree there's silence. A quiet stretch must measure below -35 dB for at least 0.35 seconds, and the transcript must show no word there. Gaps that look empty in the transcript but aren't quiet in the audio stay, because they might hold a breath, a laugh or a soft word.

Here are the numbers it uses, so you know what to expect:

| Setting | Value | Why |
|---|---|---|
| Silence threshold | -35 dB | Quiet enough for room noise, loud enough to skip soft speech |
| Shortest measured silence | 0.35 s | Ignores the natural gaps inside a sentence |
| Default minimum pause to remove | about 0.45 s | Keeps rhythm; ask for a different minimum if you like |
| Padding kept after a word | about 0.1 s | Word endings and breaths aren't clipped |
| Padding kept before the next word | about 0.13 s | The next word doesn't start abruptly |

If the transcript's timing for a word overlaps a measured silence, Vidiyo leaves that spot alone and tells the assistant, rather than cutting through a word it isn't sure about.

## How do I remove silences by hand on a Mac?

Use the waveform in any traditional editor: split where each flat stretch starts and ends, delete the gap, and close it up. It's slow but precise, and you decide every cut. Leave a few frames of air on each side so words don't clip, and listen back at normal speed.

1. Put the clip on the timeline and make the audio waveform visible.
2. Zoom in until you can see individual words as bumps.
3. Move to the end of a word before a long flat stretch, leave a couple of frames, and split.
4. Move to just before the next word, leave a few frames, and split again.
5. Delete the piece in between and close the gap.
6. Repeat, then play the whole thing back once without stopping.

## Can I find silences with a free tool before cutting?

Yes. ffmpeg, the free command-line media tool, has a `silencedetect` filter that prints the start and end time of every silence. Its defaults are a -60 dB threshold and a 2-second minimum, per the [ffmpeg docs](https://ffmpeg.org/ffmpeg-filters.html), which is far too strict for speech, so set your own.

A version tuned for a talking head:

```sh
ffmpeg -i talk.mov -af silencedetect=noise=-35dB:d=0.4 -f null - 2> silences.txt
```

That writes lines like `silence_start` and `silence_end` to `silences.txt`. You then cut at those times in your editor, or script the cuts. Remember to pull each cut point slightly inside the silence so you keep a little padding around words.

## How much silence should you leave in?

Leave short gaps between sentences and longer ones where a pause is doing a job. Cutting every gap to nothing makes speech sound breathless and machine-made. A good default is to remove pauses longer than about half a second and keep the short ones inside and between sentences, which is what Vidiyo does unless you ask otherwise.

Pauses worth keeping on purpose:

- Before a punchline or a surprising number, so it lands.
- After a question you want the viewer to think about.
- At a change of topic, where a beat signals "new section".
- At the very end, briefly, so the video doesn't stop mid-breath.

For fast short-form content you can ask for a tighter pass, such as "tighten every pause longer than a third of a second". For calm long-form content, a looser one reads better. You can always undo a pass and try again with a different minimum.

## Why does silence removal sometimes sound choppy?

Choppy results come from cutting too close to words or removing every pause, including the ones that give sentences room. Natural speech needs short gaps. Keep padding around words, set a minimum pause length, and leave pauses before a punchline or a big claim, because those are doing a job.

Background noise is the other cause. A fan, traffic or a busy room can keep the audio above the threshold, so real pauses aren't detected and only some get cut. Recording somewhere quieter helps more than any setting. After the silences are gone, the next step for most talking heads is [removing filler words](https://getvidiyo.app/blog/remove-filler-words-from-video).

## Honest limits

- In a noisy room, pauses may not measure as silent. Vidiyo leaves those for you to review rather than cutting through them.
- There's no speed control, so slow sections are tightened with cuts, not sped up.
- Timelines max out at 5 minutes.

## Frequently asked questions

### How do I automatically remove silence from a video on Mac?

Import the video into Vidiyo and type "cut the long pauses". It finds stretches that measure as silent in the audio, checks them against the transcript's word timings, and cuts them with padding kept around speech. One undo restores them all.

### What counts as a silence worth cutting?

For talking-head video, a pause of roughly half a second or more between words is usually worth cutting. Vidiyo's default removes confirmed pauses from about 0.45 seconds and you can ask for a different minimum.

### Will removing silences make my video sound choppy?

It can if cuts land right against words. Keep a little padding: Vidiyo leaves about 0.1 seconds after a word and about 0.13 seconds before the next one, so breaths and word endings survive.

### Can I remove silences for free on a Mac?

Yes. You can cut pauses by hand in any editor using the audio waveform, or list them with ffmpeg's free silencedetect filter and cut at those times. It takes longer than an automatic pass.

### Does silence removal upload my video?

Not in Vidiyo. Silence detection, transcription and export all run on your Mac. The assistant receives the transcript and measurements to decide what to cut.

## Related

- [Remove silences and filler words automatically](https://getvidiyo.app/features/silence-and-filler-word-removal)
- [Dead air](https://getvidiyo.app/glossary/dead-air)
- [Jump cut](https://getvidiyo.app/glossary/jump-cut)
- [How to remove filler words from a video](https://getvidiyo.app/blog/remove-filler-words-from-video)
- [How to add jump cuts that don't feel jumpy](https://getvidiyo.app/blog/jump-cuts-that-dont-feel-jumpy)

## Sources

- [FFmpeg filters documentation: silencedetect](https://ffmpeg.org/ffmpeg-filters.html) (accessed 2026-10-08)

Download Vidiyo for Mac (free during the beta): https://getvidiyo.app/download