Editing Talking-Head Video

A talking head is the cheapest format to record and the hardest to edit well, because there is nothing to hide behind. If the delivery drags, the cut has to fix it; if the framing is wrong, no overlay rescues it.

This lesson covers the three things that decide whether people watch: framing, delivery, and the cut.

Framing

Three decisions, and getting them right takes five minutes.

Camera height at eye level. A laptop camera below the chin is the single most recognisable amateur signal, because it produces a view up the nose and a double chin. Raise the camera to eye height; a stack of books works.

Eyes in the upper third, with a little headroom and no more. In vertical frames, headroom is expensive: frame tighter than feels natural, then check on a phone.

Distance of about an arm's length to a metre. Closer is more intimate and less forgiving; further is more professional and less engaging. For vertical, err closer.

Light from the side, slightly above. A window at 45 degrees is better than any ring light, which flattens the face and puts a reflection in the eyes.

Delivery

Most talking-head footage is unusable because of delivery, not technique.

• Speak in threes. Three points per clip, each one sentence.
• Start with the claim. The first sentence is the hook: how to create strong hooks.
• Stop when you have said it. The most common failure is continuing past the point, which forces the editor to cut the ending.
• Record the opening three times. The first take is a warm-up; the third is usually the one.

If you are scripting, write bullets, not sentences. A read script sounds read, and no edit fixes that.

The cut

Talking heads are edited almost entirely with jump cuts, and jump cuts have rules.

• Cut on the breath, never mid-word.
• Vary the framing rather than the position. A jump cut between two identical framings looks like an error; a cut from wide to close reads as a choice.
• Keep one gesture per idea. Hands that move constantly force more cuts.
• Hold one second after a strong line. The audience needs it.

A practical practice: cut a two-minute take down to forty seconds without changing the framing at all. It will be uncomfortable and it will teach you where your delivery is padded.

Audio

• Mic close, room soft. A cheap lavalier 20cm from the mouth beats an expensive mic across the room.
• Treat the room. A blanket behind the camera and one to the side removes most of the flutter echo.
• Level the voice, then add music under it if you want any: fixing bad audio in Shorts.

Text and captions

The talking head is where captions do the most work, because there is no other visual information.

• Position once, clear of the interface, and never move it.
• Emphasise the load-bearing word, once per sentence.
• Keep line length under about 38 characters: caption readability.

Practice

Record the same sixty-second answer three ways:

1. Seated, eye-level camera, bullets.
2. Standing, same framing, no notes.
3. Close framing, vertical, one claim only.

Cut all three to thirty seconds and watch them on a phone. The version that holds attention tells you what your setup should be, and it is usually not the one that felt most comfortable to record.

Next

For the b-roll and pacing that break up a long talking head, see editing for maximum watch time. For the vertical frame specifically, editing vertical video.