Most of these words describe a shape in time or a position in frame. That's why they never quite land as text definitions. Each one below is shown, then named.
A name/title graphic placed in the bottom third of the frame — low enough to stay clear of the face, inset from the edge so it isn't clipped on any screen.
Why it's called that: split the frame into three horizontal bands — the graphic lives in the lowest one. That's the whole etymology.
The inset margin (~5%) you keep text inside, so nothing gets cropped by an overscanning TV, a rounded phone corner, or a platform's UI overlay.
Where it bites: captions that look perfect in your editor and get eaten by the Like button on a phone.
Put the subject on a third-line rather than dead centre; leave a little space above the head (headroom) and more space in the direction they face or move (lead room).
A-roll is your main thread — the person talking, the thing actually driving the story. B-roll is supporting footage you cut over the top: hands on a keyboard, the screen, the room. The A-roll audio usually keeps running underneath.
Practical value: B-roll is what lets you delete someone's "umm, so, like" without a visible jump. Shoot far more of it than feels necessary.
Three specific jobs B-roll does. A cutaway looks away from the action (the listener nodding). An insert pushes in on a detail within it (the finger pressing the key). An establishing shot opens a scene by showing where we are (the house from outside).
These are the ones that genuinely need a picture. In every diagram below, the top lane is picture and the bottom lane is sound. The orange line marks the moment of the cut.
Picture and sound change at the same instant. Clean, and slightly abrupt — everything below is a way of softening it.
You hear the next scene before you see it. The audio of shot B starts early, underneath the tail of shot A.
What it feels like: anticipation. You hear the door open a beat before you see who walked in, so the cut feels motivated rather than arbitrary. Extremely common in dialogue — the reply starts while you're still on the listener's face.
The mirror image. The picture moves on, but the audio of the previous shot keeps running underneath the new one.
What it feels like: continuity. We cut to the listener's reaction while the speaker is still talking. J and L are the same tool, just which side overhangs.
Two shots of the same subject from nearly the same angle, spliced together — the subject appears to teleport slightly. A mistake in drama; the native dialect of YouTube.
Cut between two shots that share a shape, motion, or composition, so the eye glides across the join. A round window becomes a full moon; a thrown object becomes a passing car.
Why bother: it makes an edit feel like an idea rather than a splice. Use sparingly — it draws attention to itself.
A cross dissolve fades one shot into the next; conventionally it signals passing time or a soft change of place. A montage is a run of short shots compressing a long process into a few seconds, usually carried by music rather than dialogue.
Room tone is 30 seconds of "silence" recorded in the actual room — silence is never truly silent, and you need this to patch gaps invisibly. A bed is the music or ambience running under everything. VO is voiceover, recorded separately from picture. Sync means picture and its own sound are lined up correctly.
The one that catches everyone: forgetting room tone. Cut a word out and the background hiss vanishes with it — the gap is audible as a tiny hole.