The project
Begin with a short spoken sequence and six or more supporting shots. Include a wide view, a medium action, a close detail, and a reaction or atmosphere shot. Your goal is to clarify what the audience hears while keeping the visual sequence coherent.
1. Describe the footage
Give sources useful names, notes, and keywords. Describe visible subjects and actions. This makes manual search more effective and gives the B-roll assistant meaningful information to match against transcript cues.
Review each source in motion. A beautiful still frame may occur inside a shot with an unusable camera move or abrupt ending.
2. Identify the words that need picture
Listen to the sequence and mark moments where an image can explain a place, action, object, or change. Avoid covering every sentence simply to keep the screen busy. Leave the speaker visible when expression or credibility matters.
The B-roll assistant can suggest matches from library names, notes, and keywords. Inspect the chosen source and range before accepting a placement.
3. Build visual continuity
Place B-roll over the ongoing spoken sequence. Use a wide shot to establish the scene, then a closer action or detail to develop it. Check screen direction and movement at each cut.
Trim the image to the useful action. Let it begin slightly before or linger after the named concept when that supports understanding, rather than mechanically changing picture on every noun.
4. Protect the underlying sound
Check which audio components entered with the B-roll. Keep, reduce, or remove location sound deliberately so it does not unexpectedly mask the narration. Maintain source synchronization when retaining linked sound.
Review your result
Watch once with sound and once muted. With sound, the images should clarify the story. Muted, the picture sequence should still have a comprehensible visual progression. Save a version before changing the main speech edit.