Audio and captions

Choosing the audio

Audio is generated alongside the picture, and one control sets its character. Your modes are a voiceover that carries the clip with ambience underneath, a voiceover over a music bed, ambient only with no spoken words, music only, and fully silent. The default is a voiceover-led track. A music bed is opt-in, so choose one of the scored modes when you actually want music.

Here is a detail worth knowing. Left to its own devices, the scene planner tends to write spoken dialogue to fill the duration, the way a radio host fills dead air. That is exactly why ambient only and silent are real, deliberate choices rather than just the absence of a voiceover. If you want a visual-only clip, say so with the audio mode instead of hoping it stays quiet.

Captions

Captions are on by default and burned into the picture after the render finishes. They are timed from a transcript of the finished clip, so they line up with what is actually said, and they are biased toward your brand terms so your product and company names come out spelled right. Because they are burned from the rendered audio, you still get captions even when you left the spoken-lines field blank and let the planner write the talk track.

The captions control is where you turn them off or set their basic look before you generate. Deeper caption styling belongs to the editor once the clip exists, not to this form. Decide on or off and the rough look here, then fine-tune later.