keptbits
← Blog

11 September 2026 · 6 min read

Clipping a conversation with two people in it

Interviews produce the best clips and are the hardest to cut. Where the boundaries go, what to do about interruptions, and why the question usually belongs in the clip.

A monologue is easy to clip: find the good bit, cut around it. A conversation is harder, because the good bit usually depends on something the other person said, and because two people talking means two of everything — two framings, two voices, two places a cut can go wrong.

Include the question

The most common mistake is clipping only the answer. An answer without its question opens on a viewer wondering what is being discussed, which is exactly the confusion that makes people scroll. The question is usually short, it provides the whole context, and it is a strong opening in its own right.

The exception is when the answer opens with something startling enough to stand alone. Then lead with that and let the context arrive a beat later.

Where a clip should end

End on the answer, not on the next question and not on the other person's reaction. "Right, interesting" is a terrible last line, and it is where a boundary placed on a timer will very often land.

Interruptions and overlap

Real conversations overlap. People finish each other's sentences, laugh over answers, start talking before the other has stopped. A cut placed inside an overlap catches half of two voices and sounds broken.

The workable rule is to move the boundary to the nearest point where only one person is talking, even if that costs a second either side. Clean silence is a better edge than an accurate one.

Framing follows the voice

In a vertical crop you cannot have both faces most of the time, so the frame has to be on whoever is talking. When the two of them go back and forth quickly, a frame that tries to follow every exchange becomes a mess — in that case, pick the person whose answer the clip is about and stay there. A listener's reaction shot is worth having only when the reaction is the point.

Two-speaker clips have a natural length

A question and its answer is usually somewhere between twenty and sixty seconds, which happens to be the sweet spot for short-form anyway. If a clip from an interview is running past ninety seconds, it probably contains two exchanges and would be better as two clips.

What makes interviews worth the trouble

Guests say things they did not plan to say. That unplanned quality is what makes an interview clip feel different from a scripted one, and it is why a two-hour conversation with someone interesting usually out-produces a carefully prepared talk of the same length.

Keptbits cuts on sentence boundaries.

Paste a link or upload an episode, say how many clips you want, and get them back captioned and vertical. Seven days free on any plan.

Start your free trial