What is B-roll?

A-roll carries the words, B-roll carries the picture

Adobe's editing guides draw the line in plain terms. A-roll is footage that features key plot actions, talking characters or interview subjects. B-roll is everything that is not the main action: supplemental footage, cutaways, detail shots and the wide views that set a scene. In an interview the A-roll is the person answering, and it usually runs unbroken on the audio. The B-roll sits over it on the picture, so the viewer keeps hearing the answer while seeing the workshop, the finished job or the hands doing the task being described. Neither is better footage. They do different jobs, and a video that has only one of them tends to feel either static or aimless.

How editors use it

Four uses cover most of it. An establishing shot, which the Columbia film glossary describes as a long shot at the start of a scene that shows things from a distance, tells the viewer where they are before anyone speaks. A cutaway shows the thing being mentioned, so a line about a new kitchen is covered by a shot of the kitchen. Covering an edit is the practical one: when two parts of an answer are joined, the jump in the speaker's face is hidden under a few seconds of other footage, which is how interviews lose their pauses and false starts without looking chopped. And a run of short shots can compress time, a whole morning of work shown in a few seconds, which is the familiar montage.

When it matters most

B-roll earns its place whenever the speaker is describing something the viewer could be looking at. A physiotherapist explaining a stretch should be covered by the stretch. A founder describing a new warehouse should be covered by the warehouse. It matters just as much in advertising, where a feed viewer decides almost instantly whether to keep watching and a static frame gives them nothing to decide on. Adobe's guide makes the interview case directly: B-roll supplies visually interesting alternative footage so the result is not simply a person sitting still and talking into the camera. The plainer the speaker's surroundings, the more the video leans on it.

What good cover looks like

Good B-roll is specific, steady and varied in size. Specific means it shows this business and not any business: the actual counter, the actual van, the actual customer's finished bathroom. Steady means each shot is held long enough to trim, with the camera still or moving slowly and deliberately. Varied means a mix of wide, medium and close shots of the same subject, so a sequence can move from the room to the bench to the hands without two similar frames sitting side by side and looking like a mistake.

Common mistakes

The first is generic cover. A clip of people shaking hands over a line about customer service tells the viewer nothing, and they notice. The second is covering the moment that needed a face: an apology, a price, a promise, anything where the viewer is judging whether to believe the person. The third is cutting B-roll to its own rhythm rather than to the words, so the shot of the product arrives two sentences after the product was mentioned. The fourth is not shooting enough of it. Plenty of edits stall because there are only a few seconds of cover, and the useful two seconds of any shot are rarely the first two.

Where it shows up in the product

The video builder here does not film anything, so there is no camera and no footage bin in the usual sense. The closest equivalent is the uploaded clip scene: an MP4, WebM or MOV up to 200 MB, trimmed with a start and end point and fitted to its box with cover or contain, which is how your own footage of the shop or the job goes into a video. Its audio can be muted or kept, though a clip never pushes the background music down. Stills fill the same role in other scenes, from your uploads or from AI images generated to fit the slot. In a UGC video the talking scenes behave like A-roll and the short proof and credibility beats between them act as the cutaways, but uploaded clips are never used in that format.

Questions people ask