What Is B-Roll? A-Roll vs B-Roll

B-roll is the supplementary footage cut over your primary footage (a-roll) to illustrate points, cover edit cuts, and maintain visual pacing. Every edited video uses both layers: a-roll carries the story through dialogue or narration, while b-roll is everything placed around it to make that story land visually. Understanding how they interact is one of the most practical skills you can develop as a filmmaker or video creator.

What A-Roll Is

A-roll is the primary footage. It carries the core narrative: an interview subject speaking directly to camera, an actor delivering dialogue in a scene, a presenter walking through a tutorial. The audio on the a-roll is usually the main audio track — the words the audience needs to hear.

In a documentary, a-roll is the talking heads. In a drama, it's the main coverage of the scene. In a product explainer, it's the presenter. A-roll sets the structural spine of the edit.

What B-Roll Is

B-roll is supplementary footage cut away from the a-roll to illustrate, contextualize, or visually support what's being communicated. It's the city skyline shot while a narrator describes urban growth. It's hands typing while a voiceover explains a software feature. It's the establishing exterior of a building before we cut inside.

The term has a literal origin: in analog film editing, editors worked with two physical reels. The "A" reel held the primary footage; the "B" reel held the cutaways and supplementary material. The names stuck.

B-roll typically runs without primary audio. It might carry ambient sound — traffic, wind, room tone — but the main dialogue or narration continues underneath it from the a-roll track.

Why B-Roll Matters in the Edit

The most practical function of b-roll is hiding jump cuts. When you edit an interview and remove a section of dead air or a repeated phrase, the two remaining clips create a jarring visual jump on the same face. Lay a piece of b-roll over that cut and the seam disappears entirely.

Beyond covering edits, b-roll does several things at once:

Visual pacing. Staying on one talking head for three uninterrupted minutes is exhausting to watch. B-roll gives the viewer something to look at while the information continues.

Illustration. If someone says "the factory employs 400 workers," showing the factory floor makes the claim tangible. If a character references a location, cutting to that location anchors the story geographically.

Mood and atmosphere. Establishing shots — a slow pan across a foggy harbor, a time-lapse of a street at dusk — set tone before a scene begins. This is especially important in drama and micro-drama formats where the visual texture of the world needs to feel real.

Transition. B-roll bridges scenes. A shot of clouds passing or traffic moving can compress time between two story moments without a hard cut.

A-Roll vs B-Roll: A Practical Comparison

| | A-Roll | B-Roll | |---|---|---| | Content | Interviews, main dialogue, on-camera presenter | Cutaways, establishing shots, illustrative footage | | Audio | Carries primary audio | Usually muted or ambient only | | Function | Delivers the narrative | Supports and illustrates the narrative | | Edit role | Structural spine | Covers cuts, adds visual variety |

Neither category is more important than the other. A talking-head documentary with no b-roll is technically complete but visually monotonous. B-roll with no a-roll is just footage — there's no story holding it together.

How AI Generates B-Roll in 2026

Sourcing b-roll used to mean either paying a stock footage subscription or sending a second camera crew. Both options added cost and time. AI video generation has changed that calculus significantly.

Text-to-video models can now produce custom b-roll clips from a written description. On Leyline's platform, models including Veo 3 (Google), Seedance (ByteDance), and Kling (Kuaishou) generate motion from a keyframe image and a prompt. Each produces clips with distinct characteristics: Veo 3 handles motion coherence and native audio well; Kling is cost-effective for high volume; Seedance produces strong results for stylized or dramatic sequences. See how to generate b-roll with AI for a step-by-step breakdown of the full generation process.

The basic workflow for AI b-roll:

1. Identify what the a-roll is talking about at each point in the script
2. Write specific prompts — shot type, subject, lighting, mood
3. Generate several variants per scene and select the strongest
4. Color-match to your primary footage in post

Prompt specificity matters. "A wide establishing shot of a rain-soaked Tokyo street at night, neon reflections on wet pavement, no faces, slow dolly" will produce usable footage. "Tokyo street" will produce something generic. One action per shot, concrete visual details, positive framing — the same principles that apply to keyframe prompts in scripted production apply here.

Stock B-Roll Sources Worth Knowing

For footage you don't generate from scratch, several libraries provide free commercial-use clips:

• Pexels and Pixabay: both CC0, no attribution required, broad subject variety, strong first stop for most needs
• Mixkit: curated lifestyle and tech clips, free commercial license, no attribution needed
• Coverr: smaller library, higher average quality, CC0
• Dareful: 4K and 8K nature and urban footage, CC0, useful when aerial or landscape shots are needed

For higher volume or cinematic quality, Storyblocks (subscription) and Artgrid (subscription) cover professional agency workflows. AI-generated b-roll is increasingly the preferred route for branded or stylized productions where stock footage looks too generic.

Frequently Asked Questions

What is the difference between a roll and b roll?

A roll and b roll refer to the two layers of footage in an edit. The a-roll is the primary footage — interviews, dialogue, on-camera narration — that carries the main audio and story. The b-roll is supplementary footage cut over the a-roll to illustrate points, cover edit cuts, and add visual variety. Both are usually present in any edited video.

Does b-roll always have to be silent?

No. B-roll often carries ambient or natural sound — the hum of a crowd, wind, traffic. What it typically doesn't carry is the primary dialogue or narration, which continues from the a-roll underneath it. In some cases, a b-roll clip will include its own audio that becomes the focus momentarily before cutting back to the main track.

Can AI completely replace stock b-roll?

For custom or branded productions, AI generation is now a practical alternative to stock footage. Models like Veo 3, Seedance, and Kling can produce on-brief clips with specific visual treatments that stock libraries can't match. For generic needs — a sunset, a crowd, a city street — free stock libraries like Pexels and Mixkit remain faster and cheaper. Most productions use both.

How much of a finished video is typically b-roll?

It varies widely by format. Documentaries and corporate interviews often rely heavily on b-roll to cover extended interview sections. A scripted drama scene might use far less, relying instead on coverage from multiple angles of the same action. Social short-form content often mixes a-roll and b-roll in roughly equal proportion to maintain pacing for short attention spans.

Keep learning

If you're building a scripted production on Leyline and want to understand how b-roll fits into a structured AI workflow, AI filmmaking workflow walks through the full pipeline from script to export. For the specifics of choosing the right AI video model for your b-roll clips, which AI video model to use covers each model's strengths, cost profile, and what to expect from a generation run. If you're still deciding between AI generation and stock libraries, where to get b-roll footage compares the main options side by side.