Clip modes

What kind of video it is decides how clips are chosen and framed. Pick the one that matches what you uploaded.

General

The default. Made for vlogs, day-in-the-life videos, talking to camera and ordinary long-form YouTube.

• Moments come from the sound. What you say, your reactions, laughter, music and movement on screen all count towards picking a clip.
• Framed on the person in shot. Clips crop around the face on screen, and show the whole frame when nobody is in view or several people are spread out.
• Nothing to hear and nobody talking? Clips are taken at evenly spaced points through the video instead.

Gaming

For gameplay footage. Listens to your gameplay rather than reading a transcript to find moments.

• Talking over it? Your commentary is the strongest signal for a clip.
• No commentary? It goes by the game itself - gunfire, explosions, big sounds and fast action.
• Layout. Choose how your gameplay is framed for a vertical screen: the whole frame (safest for any game), zoomed in, centre-cropped for shooters, or your gameplay stacked over your own webcam if your recording has one.

Tutorial

For screen recordings and demonstrations: software and coding walkthroughs, cooking, DIY, whiteboards, presentations - anything that shows a process rather than a face.

• Moments come from what you say. Explaining a key point, a tip, a warning or a reaction is the strongest signal for a clip, with activity on screen counting too.
• Shows your whole screen by default. Nothing is cut off. The same layout choices as Gaming are available if you'd rather zoom in or keep a webcam in view.
• A recording that barely changes while something important is said is not skipped.

Podcast

For conversations: interviews, panels, two or more people talking.

• Framed on whoever is speaking. It works out who is talking at each moment and crops to them, following the conversation as it moves between people.
• Moments come from the transcript. Clips are chosen for complete exchanges - a question and its answer, a whole story - rather than for loud or busy moments.
• Just one person talking to camera? General mode frames a single speaker the same way, for less processing - Podcast's speaker-tracking only earns its keep with more than one person on screen.

Which one should I pick?

• Vlog, talking-head video, or you're not sure - General.
• Recorded gameplay - Gaming.
• Screen recording, walkthrough or demo - Tutorial.
• Two or more people in conversation - Podcast.