What is B-roll, and how do you shoot and use it?
The interview tells the story; B-roll shows it. Here is what B-roll is, what to shoot for different kinds of video, and how to cut it in so the film feels made, not recorded.
A-roll vs B-roll
A-roll is the footage that carries the story: the interview, the presenter talking to camera, the speech from the stage. If you turned the picture off, the A-roll's sound alone would still tell you what the film is about.
B-roll is everything you cut away to while the story keeps going: the workshop the founder is describing, hands on the product, the team at the whiteboard, the street outside the office. It usually plays over the A-roll's sound, so the viewer hears one thing and sees another that belongs to it.
The names come from film editing before video, when two rolls of film were run together to mix an interview with other shots. The terms stuck. Today “B-roll”, “cutaways” and “supporting footage” mean roughly the same thing.
Why B-roll matters
B-roll does three jobs in an edit, and most videos need all three.
- It covers cuts. When you remove a stumble or join two parts of an answer, the picture jumps. A shot laid over the join hides it, and the viewer hears one clean sentence.
- It shows what is said. “We rebuilt the whole line in a weekend” lands harder if the viewer sees the line. Words tell; pictures prove.
- It sets the pace. A talking head for two minutes feels long even when the words are good. Cutting away gives the eye something new and lets the film breathe between ideas.
Without B-roll you have two choices for every cut in an interview: a visible jump, or keeping the stumble. With it, you can shape the answer and still keep it honest.
A B-roll shot list by video type
Before the shoot, read the questions or the script and write down what the viewer should see while each answer plays. Then add the shots every film of that kind needs. Some starting points:
Interview or testimonial
- The person at work, doing the thing they will talk about, not posing.
- Their hands: typing, writing, holding a tool or the product.
- The place: the building from outside, the room they work in, a sign on the door.
- Objects from the story: the old machine, the first prototype, the photo on the wall.
- A wide of the interview set-up itself, useful as an opening or a reset.
Event or conference
- Arrivals, badges, the coffee queue.
- The full room at its fullest, from high up if you can.
- Faces listening, laughing, writing something down.
- Details: the stage name, the screen, the stand, the lanyards.
Product
- The product in use by a real person, in the place it is meant for.
- Close-ups of the parts that matter: the button, the screen, the texture, the packaging opening.
- Before and after, if the product changes something you can see.
- The product in the hand, to show its size.
Office or team
- People talking to each other, not to the camera: a meeting, a call, a pair at one screen.
- Walking shots through the space: corridors, stairs, the door opening.
- Workstations and the things on them that show what the team does.
- The building and the neighbourhood, at the time of day that suits the story.
How to shoot B-roll
Good B-roll is mostly discipline. The shot you will cut is a few seconds long, but you need a steady, clean stretch to cut it from.
- Hold every shot for ten seconds or more before you move or stop. Count it. Most unusable B-roll is a good shot that lasted three seconds.
- Keep it steady. Use a tripod, a wall, a table or your elbows against your body. Handheld is fine when it is a choice, not a wobble.
- Shoot movement in the frame, not just movement of the camera. A hand reaching, a door opening, a person crossing the room give the editor a natural place to cut.
- When you do move the camera, move slowly and in one direction, with a still moment at the start and the end. The editor can use the move or the still frames.
- Get each subject in three sizes: a wide to show where we are, a medium to show what is happening, a close-up to show the detail. Three sizes of one action cut together like a sequence.
- Record sound on every shot. Natural sound under B-roll, the clack of a keyboard or the hiss of a coffee machine, makes the film feel real, even when you mix it low.
- Shoot details on purpose. Hands, labels, screens, tools. Close-ups are the easiest shots to cut in anywhere and the ones people forget.
How much B-roll to shoot
More than you think. A finished interview film often spends much of its time cutting away from the speaker, and each cutaway is a few seconds. That adds up to a lot of different shots, not a lot of minutes of one shot.
- Aim for variety over length: twenty different ten-second shots are worth more than five two-minute ones.
- Shoot at least one shot for every concrete thing the speaker mentions: a place, an object, an action, a person.
- Shoot a few general shots that fit anywhere: the building, the team at work, details of the space. They save the edit when a line has nothing specific to show.
- If you can, shoot B-roll after the interview. Then you know what was said and can film exactly that.
Editing with B-roll
Cut the A-roll first. Lay the interview answers in order, trim them to clean starts and ends, and watch it with no B-roll at all. When the story works as sound, start covering it.
- Lay a shot over the words that name it. When the speaker says “the warehouse”, the warehouse appears on or just after the word. A shot that arrives a sentence late feels like decoration.
- Cover the joins. Every place you removed something from an answer is a candidate for a cutaway. Start the B-roll a moment before the join and end it a moment after.
- Use J-cuts and L-cuts. Let the speaker's voice start under the last shot before we see them (a J-cut), or let their line finish over the next shot (an L-cut). The sound carries the viewer across the cut.
- Come back to the face for the line that matters. The emotional moment, the admission, the laugh, the pause before an answer: let the viewer see it. Covering the key line with a nice shot of the office takes the feeling out of the film.
- Vary the length. A run of cutaways all the same length starts to feel mechanical. Short shots for a list of things, longer ones where the picture needs time.
- Let it breathe. After a strong line, a few seconds of B-roll with natural sound and no talking gives the viewer time to take it in.
For the whole route from raw interview to finished cut, see how to edit an interview video.
Stock, your own, or generated B-roll
Sometimes you don't have the shot. There are three places to get one, and each has a cost.
- Your own footage. Always the first choice. It shows the real place and the real people, and it matches the light and look of the interview. The cost is time on the shoot.
- Stock footage. Fast and often beautiful, but anyone can tell it is not your office, and the same clips appear in other companies' videos. Check the licence covers where you will publish. Use stock for general ideas, a city, a season, the sea, not for things that claim to be yours.
- Generated footage. AI video models can make a shot that matches a description. It is useful for an abstract idea, a visual metaphor, or a gap where nothing else fits. It is not footage of what happened, so don't use it to show a real event, a real product or a real person as if it were filmed. Results vary, and each attempt costs money on most services.
Common B-roll mistakes
- Shots too short to use. Hold for ten seconds, every time.
- Only wides. Without close-ups, cutaways have nowhere to go.
- Literal to the point of comedy. The speaker says “growth” and a plant appears. Show the real thing they mean, or show them.
- Wallpaper. B-roll that has nothing to do with the words, laid on to hide the face. The viewer stops listening.
- Covering the key line. The most important sentence usually needs the speaker's face.
- People who didn't agree to be filmed. Ask, especially in offices and at events, and keep screens with private information out of shot.
- Stock that pretends to be you. Viewers recognise it, and it undermines the real footage around it.
A B-roll checklist
- Shot list written from the questions or the script before the shoot.
- At least one shot for every concrete thing the speaker mentions.
- Wide, medium and close-up of each key action.
- Every shot held ten seconds or more, steady, with sound recorded.
- A handful of general shots that fit anywhere.
- Permission from the people in frame; nothing confidential readable.
- In the edit: A-roll first, B-roll over the words that name it and over every join.
- The key emotional line left on the speaker's face.
- Stock and generated shots only where they don't claim to be real.
Doing this in Endroll
Endroll is an AI video editor for Windows 10/11 and Macs with Apple silicon. Point it at your footage folder and it prepares everything on your computer: it listens to the speech, finds the faces and describes what is in each frame. Search everything then finds B-roll by what you remember seeing or by a spoken phrase, and returns shots, not files.
- In documentary drafts, a sit-down interview runs as voice-over under B-roll chosen by the words, and long interview runs get B-roll breathers.
- Single-camera interviews get B-roll cutaways over their jump cuts from the first draft, with the speaker's audio continuing underneath. Two shots of one camera setup are never cut together in drama or documentary.
- Sound across the cut in the Edit chapter: set a J-cut lead or an L-cut tail on any shot by hand.
- When you don't have a shot, Generate a picture or video makes one, with the price shown on the button, and Draft a prompt from the surrounding footage writes a prompt that matches the shots around the gap.
What the AI does not do: the draft doesn't invent footage or put words in anyone's mouth. Every shot in it is one you filmed unless you generated it yourself, and generated clips are marked in your library. Editing by hand never spends a credit. See how it works for interview documentaries and customer stories.