How to remove silences from a video automatically
What "silence" actually means to an editor, how automatic silence removal decides where to cut, and how to do it in Vidova by asking for it in plain English.
Silence is the most common thing to cut from a video. It is also the most tedious to cut by hand. You find a gap, split the clip, delete the gap, and move the rest left. Then you repeat this for every pause in the recording.
Automatic silence removal does this for you. This article explains how it works, and how to run it in Vidova.
What "removing a silence" means
Some tools look at the volume of the audio track. They cut any part below a set threshold.
This method is not reliable. Room tone, mic hiss, and quiet consonants sit near the same volume as an actual pause. A tool that cuts by volume alone can cut in the middle of a word.
Vidova uses a different method. It reads the transcript instead of the audio.
A transcript already has a start time and an end time for every word. A silence is the gap between the end of one word and the start of the next word. Vidova finds these gaps directly from the transcript. It does not analyze the audio.
Not every gap is a silence to remove
Speech has natural pauses. A pause before a key point. A pause after a question. If a tool removes every gap, the video sounds cut up and unnatural.
Vidova uses two rules to avoid this:
How Vidova decides
- Minimum gap length: A gap must be at least 1.2 seconds long. Shorter gaps are treated as normal speech rhythm, not silence.
- Padding: Vidova keeps 150 milliseconds of audio on each side of a cut. This stops words from getting clipped.
- Vidova also removes silence before the first word and after the last word.
These settings are cautious by design. A leftover pause is a small problem. A clipped word is a bigger one.
Two ways to remove silences in Vidova
Method 1: Ask the agent in chat.
Type an instruction like "remove the silences." The agent reads the transcript, finds every gap over 1.2 seconds, and cuts them on the timeline.
This method needs two things: a transcript for the clip, and one recording clip on the timeline. If the agent cannot find a matching clip, it tells you. It does not guess.
Method 2: Use the Transcript Editor.
Open the transcript panel and select Silences. Vidova marks every gap of 0.5 seconds or more with a strikethrough. This threshold is lower than the chat method, so it also catches smaller pauses.
No cuts happen yet. You review each marked gap. You can unmark any gap you want to keep. Then you select Apply.
Two different workflows
The chat method is fast. The agent cuts the silences right away. You can undo the change if you want a pause back.
The Transcript Editor is a review step. It marks the cuts first and waits for you to approve them before anything changes.
When to use each method
- Talking-head videos, podcasts, and screen recordings: Use the chat method. The 1.2-second threshold is usually enough.
- Interviews with deliberate pauses: Use the Transcript Editor. You can keep the pauses that matter.
- Multi-take recordings: Remove silences first. A shorter timeline is easier to scan when you pick the best take.