1. What B-roll is for

A-roll is your main story: the person talking, the screen recording, the thing the narration is about. B-roll is supporting footage laid on top of it to show what the A-roll says. Good B-roll adds clarity. Bad B-roll is decoration that pulls the eye away from the message.

In this lesson you add B-roll and overlays to your retention cut from lesson 1.2 so it becomes the finished module deliverable.

2. Primary story track vs support tracks

In CapCut desktop the bottom video track is the main track. Anything you drag above it becomes an overlay. Keep this vertical order for every project:

• A-roll base - carries the narration. Its audio keeps playing under B-roll (that is an L-cut in practice).
• B-roll support - mute its audio in most cases so it does not fight the voice.
• Text accents - one short idea at a time.
• Graphic markers - arrows, circles, highlights that point at something.

3. B-roll timing and relevance rules

1. Relevance: the B-roll shows exactly the noun or action being said. If you cannot say why a clip is there, it goes.
2. Enter slightly early: start the B-roll on or just before the keyword, not after.
3. Hold long enough to parse: roughly 1.5-4 s per shot. Less and viewers cannot read it; more and it becomes wallpaper.
4. Exit on a beat: return to the speaker for emotion, emphasis or a new idea.
5. Cover weak moments: B-roll is ideal for hiding jump cuts, stumbles and long static stretches.

Map the script to B-roll before you search for footage:

SCRIPT-TO-B-ROLL MAP - retention_cut
LINE (A-roll audio)                          B-ROLL                  IN        OUT
"Most people lose viewers in 3 seconds"      analytics graph drop    00:00.4   00:02.6
"Open the project settings..."               screen rec: settings    00:05.0   00:08.2
"...and set 30 fps"                          zoom on fps field       00:08.2   00:09.8
"Here is the result"                         before/after split      00:26.0   00:30.5
(no B-roll - emotional line, stay on face)   -                       00:30.5   00:35.0

4. Shot selection matrix

5. Picture-in-picture basics

Picture-in-picture (PiP) is simply an overlay clip scaled down so the clip below stays visible. In CapCut desktop: drag the clip onto a track above the main track, select it, and in the Video - Basic panel change Scale and Position (or drag the corners in the player).

• Typical tutorial layout: screen recording full frame, speaker in a corner at 25-35% scale.
• Keep PiP inside the safe area - platform UI (captions, buttons, usernames) covers the edges, especially the bottom and right side on vertical video.
• Add a subtle border or rounded mask (Mask tab) so the frame reads as a deliberate window.
• Use keyframes on scale/position if the PiP must move away from something important.

6. Opacity and blending for clean composition

7. Composition checklist

• The subject stays visible - no overlay covers a face or the thing being explained.
• Text is readable - enough contrast, on screen long enough to read twice.
• Every overlay lasts long enough to parse (about 1.5 s minimum).
• B-roll audio is muted or ducked under the voice.
• Nothing important sits in the platform UI zones.

8. Exercise

1. Open your retention_cut project from lesson 1.2.
2. Write a script-to-B-roll map and a shot selection matrix for it.
3. Replace three weak talking-head moments with contextual B-roll without changing the total duration or pacing.
4. Add one picture-in-picture segment and one overlay using a blend mode.
5. Run the composition checklist, then export the rough cut and the retention-optimized cut side by side - this is your module deliverable.

Further reading

• CapCut Help Center
• CapCut resources and tutorials
• Wikipedia - B-roll
• PremiumBeat blog - editing and B-roll articles