How Long to Pause Between Sentences in a Video

Cut every pause out of a recording and it sounds like a machine. Leave them all in and it drags. The useful question is how long to pause between sentences in a video, and there is real research on what sounds natural. This post sets out what the studies found, which pauses to keep, and how to turn those numbers into silence-removal settings instead of guessing.

What the research says about pause length

Two studies are worth knowing, with their limits stated up front.

• 0.6 seconds sounded most natural. In a 2022 study in Frontiers in Psychology, Liu and colleagues played English speech with the pauses at commas and full stops set to different lengths, and asked native English, Chinese and Japanese listeners to rate how natural it sounded. Speech with 0.6-second pauses was rated the most natural.
• Full stops get about twice the pause of commas. The same paper summarises earlier measurements: pauses at commas of roughly 0.38 to 0.67 seconds, and at full stops of roughly 0.81 to 1.24 seconds, typically about 1:2. In commencement speeches, for example, that came to about 0.49 seconds at commas and 1.01 at full stops.
• Speed and sentence length matter. A 2020 study in Respectus Philologicus measured the pauses of eight readers. Between sentences, faster readers paused for 200 to 980 milliseconds on average and slower readers for 340 to 1,360, and longer sentences were followed by pauses up to 1.4 times longer.

The limits matter. Both studies used text read aloud, not off-the-cuff speech, and the second one measured eight speakers reading Lithuanian. Treat the numbers as a well-founded starting point, not a rule. Conversation in particular has its own rhythm.

What that means for an edit

Put together, the research sorts the pauses in a recording into three groups.

• Under about 0.7 seconds: the pauses inside a sentence, at commas and between phrases. These are part of how the sentence sounds. Keep them.
• About 0.8 to 1.2 seconds: the ordinary pause between sentences. Keep them in long-form video, a tutorial or a podcast. In a Short, where the pace is tighter, trimming them is a reasonable choice.
• Well beyond that: a speaker searching for the next thought, checking notes, or restarting. That is dead air, and it is what silence removal is for.

The one-to-two ratio is the part people break most often. Cutting every pause down to the same tiny length removes the difference between a comma and a full stop, and the speech loses its shape. If you tighten, tighten the long pauses more than the short ones.

Turning the numbers into settings

Silence removal in flshy has three decisions, and each one maps onto the research. The reference for every control is remove silence from a recording.

Minimum gap: which pauses count

The minimum gap is how long a silence has to last before it counts as a gap at all. Anything shorter is left exactly as it is. So it is the setting that decides which group of pauses survives untouched.

• To keep sentence breaks, set it above about 1.2 seconds. Only pauses longer than an ordinary full stop are touched, which suits long-form talk.
• To tighten sentence breaks but keep commas, set it a little above 0.7 seconds, if the control goes that low. Sentence-length pauses are shortened and pauses inside sentences are left alone, which suits short-form.

Padding: how much of a cut pause is left

Pad Before and Pad After keep a little audio on either side of every cut. Both default to 0.1 seconds and go up to 2. Because padding is kept on both sides, a pause that gets cut is left at roughly the two paddings added together, so the defaults leave about 0.2 seconds: tight, jump-cut pacing.

If you want cut pauses to land near the 0.6 seconds listeners found most natural, raise both to about 0.3. That is an approximation, so listen to the result rather than trusting the arithmetic.

Threshold: what counts as silence

The threshold decides what is quiet enough to be silence in the first place. It runs from -80 to -20 dB, and -40 is the default. A noisy room needs it higher, or the room noise keeps every pause above the line and nothing is found.

Doing it in flshy

This all runs free in flshy, in the browser, on your own machine.

1. Bring the recording in and drag it onto the timeline.
2. Open Remove silent gaps in the Audio tab of the inspector.
3. Set the three decisions from the section above, for the kind of video you are making.
4. Analyse, and read the gaps. The proposal shows how many gaps it found, the time they would save, the length the clip would end up, and the first gaps by timecode with how long each one is. If the list is full of short gaps you meant to keep, raise the minimum gap.
5. Cut, then listen. The cut is one step in the undo history, so it costs nothing to try a setting, listen at the joins, and go back.

Where this fits

For conversations, the beats between speakers matter more than any number, which is covered in removing long pauses from podcast video. For solo to-camera video cut tight, see jump cuts for talking head videos. Once the pauses are set, caption the clip at its final length, because caption cues are not moved by a silence cut: add captions to a video.