Basics

What Is B-Roll? A-Roll vs B-Roll, Explained With Examples

B-roll is any footage that cuts away from your main shot. If A-roll is you talking to camera, b-roll is everything else the viewer sees while you keep talking — graphics, animated text, charts, cutaways, product shots. It illustrates the point, hides your edits, and keeps the frame moving so attention doesn't drift.

The definition, precisely

B-roll is supplementary footage intercut with the main footage. The test is simple: if you closed your eyes and the video still made sense, what you were hearing is the A-roll. Everything you were seeing that wasn't the speaker is b-roll.

Note that "b-roll" describes a role in the edit, not a type of file. The same clip can be A-roll in one video and b-roll in another. A screen recording is A-roll in a software tutorial and b-roll in a video about productivity.

A-roll vs B-roll

A-rollB-roll
CarriesThe messageThe illustration
TypicallyTalking head, interview, primary screen captureGraphics, cutaways, charts, product shots, stock
AudioUsually the source of the audioUsually silent, plays under the narration
If removedVideo is incomprehensibleVideo is comprehensible but flat, and every cut is visible

Why it's called "b-roll"

It's a survival from physical film. Editors kept the primary take on one reel — the A reel — and duplicates or supplementary shots on a second reel, the B reel, so they could cut between them without cutting into the original negative. The reels disappeared decades ago. The word didn't.

The three jobs b-roll actually does

  1. It shows the thing. You say "revenue tripled in two quarters." A chart appearing at that moment does more than the sentence can on its own.
  2. It hides your cuts. Removing an "um" from a talking-head shot leaves a visible jump. Cut away for those two seconds and the jump disappears. This is why b-roll is often called the editor's safety net.
  3. It resets attention. A static frame decays. Changing what's on screen buys back another stretch of attention — which is why retention graphs tend to dip during long uninterrupted talking-head sections.
"It shows the thing" in practice: generated b-roll for the line "say your revenue tripled in two quarters" — the chart lands on screen as the line is spoken.

Kinds of b-roll (footage is only one)

For knowledge content the ranking is usually: motion graphics > screen recording > self-shot > stock > AI video. Most creators reach for stock first because it's the most obvious, and it's the one that matches their words least.

B-roll shot types

"Kinds" is about where the footage comes from. Shot types are about what the camera is doing — the vocabulary editors actually use when they mark up a script. Each one answers a different question the viewer is silently asking.

For screen-based and explainer content, most of these have a motion-graphics equivalent that works better than filming anything: the establishing shot becomes a title card, the insert becomes a callout on the number you just said, the detail shot becomes an animated diagram. The grammar is the same; only the medium changes.

When to use b-roll, by format

The rule of thumb — cut away whenever the frame has stopped earning attention — plays out differently depending on what you're making.

If you're making the third or fourth kind, the practical question isn't "where do I find footage" — it's "how do I get a visual for every claim I made, at the second I made it." That's a different problem, and it's the one generating b-roll from your subtitles solves.

How much do you need?

Ignore ratios like "30% b-roll." Two practical rules do more work:

Why the 10–20 second rule exists: attention decays on a static frame, and every new shot buys some of it back. (Generated b-roll for the line "viewers decide in the first three seconds whether they stay or scroll away.")

Where to get it

The traditional route is a stock library: search, license, download, drag onto the timeline, then trim and slide each clip until it lines up with the right sentence. It works. It's also hours per video, and the result sits near your point rather than on it.

There's a faster route for narrated content specifically. Once you have subtitles for your narration, the .srt file already contains both what you said and exactly when you said it — which is the entire cut list. Ahacut reads that file and generates a full-length motion-graphics b-roll track cut to your timings, which you drop onto your existing timeline in CapCut, Premiere or Final Cut. Because the visuals are built from your timeline rather than fitted to it afterwards, there's nothing left to nudge.

The generated route, as generated b-roll: the narration says "transcribe, generate, drop it on your timeline" and the diagram animates step by step in sync with it.

See it on your own script

Upload subtitles, get a synced b-roll track back. Free credits to start, no card.

Start free

Frequently asked questions

What is b-roll?
B-roll is any footage that cuts away from your main shot. If A-roll is you talking to camera or your screen recording, b-roll is everything else the viewer sees: graphics, animated text, charts, product shots, cutaways. It exists to illustrate what's being said and to hide edits in the main footage.
What is the difference between A-roll and B-roll?
A-roll carries the message — usually the person speaking, or the primary screen capture. B-roll supports it. A-roll is what you'd still understand with your eyes closed; b-roll is what makes it clearer and keeps the frame moving.
Why is it called b-roll?
It's a leftover from physical film. Editors kept the primary footage on the A reel and duplicate or supplementary footage on a second reel — the B reel — so they could cut between them without damaging the original. The reels are gone; the vocabulary stayed.
Does b-roll have to be video footage?
No. For knowledge and explainer content, the most useful b-roll is usually motion graphics — the number you just quoted, the term you just defined, the three steps you just listed — animated on screen at the moment you say them. Stock footage is only one kind of b-roll, and often the weakest kind for this content.
How much b-roll does a video need?
There's no ratio worth memorizing. The practical rule is that a talking-head shot can hold attention for roughly 10–20 seconds before the frame needs to change, and every concrete noun, number or list you say is an opportunity for a cutaway.
Where do you get b-roll footage?
Four routes: shoot it yourself, license it from a stock library, generate it with AI video, or generate motion graphics from your own script. For explainer content the last one matches your words most closely, because it's authored from what you actually said rather than searched for afterwards.
What are the types of b-roll shots?
The common ones are the establishing shot (wide, sets the location), the cutaway (anything that leaves the main subject), the insert (a tight shot of one specific detail), over-the-shoulder and POV shots, detail or macro shots, reaction shots, and transition shots such as a whip pan. For screen-based content each has a motion-graphics equivalent: a title card, a callout on a number, an animated diagram.
When should you use b-roll?
It depends on the format. In documentary, b-roll carries the story between interview segments. In interviews and news, it mainly hides edits. In online explainers and talking-head video, its job is comprehension and retention — cut away whenever the frame stops earning attention, and put a visual on every number or term you say.

Related: How to add b-roll to videos automatically · How to write a video script · How to write a transcript of a video