What Is A-Roll? A-Roll vs B-Roll for YouTube Videos

A-roll (also written A roll) is the main footage of a video: the shot that carries the story, usually you talking to camera or an interview subject answering questions. B-roll is everything you cut over it: the screen recording, the product close-up, the clip of the thing you are describing. Cut the A-roll first, then lay B-roll where the viewer needs to see what you mean.

That is the whole idea, and it has not changed since editors were splicing 16mm film. What has changed is how long the A-roll takes to cut. A 20 minute talking-head recording is usually eight minutes of usable footage buried in silences, false starts and repeated takes, and digging those out by hand is most of an editing day. AutoCut does that pass for you, and it is near the end of this post. First, the definitions, because the top results for this query stop there.

What is A-roll in video?

A-roll is the primary video and audio of a project: the footage where the subject speaks, acts or does the thing the video is about. On YouTube it is you, on camera, delivering the script. It sets the pace and carries the sound. Remove it and there is no video.

It decides how long the video is, where the cuts land and what the viewer hears. When an editor says "cut the A-roll first", they mean build the spine before you decorate it.

Examples of A-roll on YouTube:

• A talking-head explainer shot on one camera at a desk
• Both hosts of a podcast, on camera, mid-conversation
• An interview subject answering questions
• A tutorial presenter introducing the topic before the screen recording starts
• A vlog piece to camera, phone held at arm's length

The name is older than YouTube. In film-era editing the cut negative was split across two reels, an A roll and a B roll, with black leader between the shots, so the printer could alternate between them and hide the splices (Wikipedia's history of the term). Television news kept the words: the reporter's interview was the A-roll, the cover shots were the B-roll. Dictionary.com still defines A-roll as the primary footage shot by the principal crew, picture and sound together.

What is B-roll footage?

B-roll is the supplementary footage you cut over the A-roll: close-ups, screen recordings, product shots, location shots, stock clips, archive, anything that shows what the speaker is describing. It usually carries no sync audio. B-roll covers cuts, illustrates a point and gives the viewer's eyes something new to look at.

B-roll has two jobs, and they are different jobs:

1. Illustrate. You say "the new dashboard", the viewer sees the new dashboard. The picture does the work the sentence cannot.
2. Cover. You cut a sentence out of the A-roll and the frame jumps. Lay three seconds of B-roll over the join and nobody sees it. This is the classic fix for a jump cut you do not want.

Examples of B-roll footage on YouTube:

• Screen recording of the app you are reviewing
• Hands unboxing the product, shot on a phone
• Drone or walking shots of the place you are vlogging
• Stills, charts and archive clips in a video essay
• Stock footage of a city, a server room, a stock chart

A-roll vs B-roll: the differences

If you only remember one row of this table, remember the audio one: A-roll is the footage you can hear, B-roll is the footage you can only see.

One clarification: a second angle of the same speaker is not B-roll. If you run two cameras on a podcast, the wide and the singles are all A-roll, because they all carry the sync audio and the story. B-roll is footage of something other than the speaker speaking.

A-roll and B-roll examples by YouTube format

The split looks different depending on what you make. The A-roll is always the spine, but how much of the screen time it gets varies a lot.

Notice the tutorial row. The screen recording can be most of the runtime and it is still B-roll, because the voice driving it came from the A-roll. Runtime share does not decide which is which. Audio does.

How much B-roll should a video have?

Rule of thumb for a talking-head YouTube video: B-roll covers 20 to 40 percent of the runtime, in inserts of 3 to 8 seconds, and the presenter's face is never gone for more than 15 seconds at a stretch. That is a rule of thumb, not a law: a sane default when you have nothing else to go on.

The reasoning behind the numbers:

• Below 20 percent the video reads as static unless jump cuts, punch-ins and on-screen text do the visual work instead. Plenty of commentary channels run at zero B-roll and hold fine. They earn it with pace.
• Above 40 percent the viewer loses the person they clicked on. A face builds trust; B-roll proves a point. Too much proof, not enough person.
• The 15 second cap exists because long stretches of B-roll with nobody talking to camera are where retention dips tend to show up. YouTube's own audience retention report marks dips as moments viewers skipped or left, and it singles out the first 30 seconds as its own metric. Do not open on a B-roll montage.

Turned into minutes:

Tutorials, cinematic vlogs and video essays break the rule on purpose. For everything else, when to cut away is simpler than how much:

• Cut to B-roll when you name a thing the viewer has not seen yet
• Cut to B-roll to cover an edit you do not want noticed
• Cut back to the face for the punchline, the promise and the ask
• Never cut away in the middle of eye contact that matters

How do you shoot A-roll so it cuts well?

Shoot A-roll for the edit, not for the take. Lock the camera, lock exposure and white balance, leave the mic where it is, and record one continuous file. When you fluff a line, pause, then restart from the top of the sentence. Every take then matches every other take, and any two of them can be cut together without a visible seam.

Here is the shoot-day checklist, in order:

1. Lock the frame. Tripod, one framing for the whole session. If you deliver in 1080p, shoot in 4K: a crop gives you a clean punch-in as a second "angle" without a second camera.
2. Go manual. Exposure, white balance and focus all off auto. Auto modes re-evaluate every time you stop and start, and the cut between two takes shows up as a brightness or colour shift.
3. Match frame rate to delivery. YouTube's recommended upload settings list 24, 25, 30, 48, 50 and 60 frames per second and say to upload at the frame rate you recorded. Pick one, keep it for the whole shoot, and set shutter to double the frame rate.
4. Fix the mic and record room tone. Ten seconds of silence at the start of the file. Do not move the mic between takes; a mic that moved is the mismatch you cannot fix later.
5. Keep rolling. One file per session, not one file per take. When you stumble, pause for two seconds, restart the sentence (not the paragraph), and pause two seconds after. Silence is what you, and any editing software, cut on.
6. Return to a home position. Between sentences, come back to the same posture and eye line. Two takes that begin in the same place cut invisibly.
7. Leave handles. Hold your look at the lens for a beat after the last word of every section. That beat is where the cut lands.
8. Say the take. "Take three" out loud, or a hand clap. Every retake then lines up in the edit.

The most common failure is number 2. Stop the camera between takes with auto white balance on, and every cut in the finished video has a tiny colour pop, even when the framing is identical.

If your whole channel is this format, the talking head video guide goes deeper on lighting, lens choice and delivery.

How do you edit A-roll fast?

Cutting A-roll by hand is a watching job before it is an editing job. You import the file, sync the audio, then play the whole recording while cutting silences and choosing the best take of each line. Only then do you lay B-roll, grade the picture, clean the audio and export.

The manual pass, step by step:

1. Import the footage and sync any external audio to the camera track.
2. Play the whole recording. Cut every silence longer than a beat.
3. Where you repeated a line, compare the takes and keep one.
4. Tighten: remove the ums, the breaths, the sentence that went nowhere.
5. Now, and only now, place B-roll, grade, fix the audio and export. This is the rough cut to final cut sequence, and steps 2 to 4 are the rough cut.

Steps 2 to 4 take longer than the recording itself, because you have to watch all of it, and the repeated takes twice. That is the boring part, and it is the part that stops most people posting weekly.

AutoCut does steps 2 to 4 for you. Drop in the raw talking-head footage, one file or several video and audio files, MP4 or MOV, up to 30 minutes. No prompts, no setup. It removes the silences, the rambles, the dead air and the repeated takes and returns a clean A-roll. Where you recorded a line more than once, Auto Retakes lines every version up side by side so you pick the one you like, which is step 3 exactly. Audio Enhance removes background noise and improves the voice, including phone and cheap-mic recordings. A clean A-roll comes back in minutes.

It works standalone or as the first step before Premiere Pro or DaVinci Resolve: take the clean A-roll into your editor and spend the saved time on the B-roll. It is not a clip maker. It produces the long video, not vertical highlights.

Your first AutoCut is free, on up to 30 minutes of raw footage.

A-roll first, B-roll second: the order that saves the most time

B-roll placed before the A-roll is locked has to be moved every time you tighten a sentence, and every move risks the insert landing on the wrong word. So the working order is:

• Lock the A-roll. Silences out, takes chosen, rambles gone. Nothing else touches the timeline until this is done.
• Lay B-roll to the words. Start each insert on the noun it illustrates, not before it. Cut back to the face before the next idea starts.
• Grade and audio last. Both are global passes; doing them before the cut is locked means redoing them.

If you are new to all of this, start with the video editing for beginners guide, then the full how to edit YouTube videos walkthrough. Both assume you know what A-roll and B-roll are, which you now do.

Frequently asked questions

What is an A-roll and B-roll?

A-roll is the main footage of a video, the shot that carries the story and the sync audio, usually a person talking to camera or being interviewed. B-roll is the supplementary footage cut over it: close-ups, screen recordings, location shots and stock clips that illustrate what the speaker is saying or cover an edit. Cut the A-roll first, then lay B-roll on top.

What is an example of A-roll?

A YouTuber sitting at a desk explaining a topic to camera is A-roll. So is a podcast host mid-conversation, an interview subject answering a question, or a vlogger talking to a phone at arm's length. The test is simple: if the footage carries the voice and the story, and deleting it would leave no video, it is A-roll.

What is the difference between A-roll, B-roll and C-roll?

A-roll is the main story footage, B-roll is the supporting footage cut over it. C-roll is an informal label some editors use for a third layer: reaction shots, second-camera cutaways of the speaker, or behind-the-scenes clips. It is not a standard term and most editors never use it. Two layers cover almost every YouTube video.

What is B-roll on Instagram?

The same thing at a smaller size. In an Instagram Reel, B-roll is the footage laid over a voiceover or a talking clip: the product, the place, the process. Creators say "post B-roll" when they mean a clip with no one talking to camera, often with text or a voiceover carrying the story instead of a face.

What does "a roll" mean outside video?

Three things, depending on who is talking. "On a roll" means a run of success, as in winning several games in a row. In food, a roll is a small round bread or a sushi roll. In video production, A-roll is the primary footage, which is what this article is about. Two of the three are not about cameras.