Scenes & images
Narration and ambient sound
Two separate audio tracks per scene, and when each is worth adding.
A scene can carry two independent audio tracks.
Narration
A spoken introduction to the scene. Record it in the browser or upload a file. It plays once when the learner arrives and can be replayed from the scene's controls.
Narration is the single most effective addition to a scene that is otherwise a wall of pins: thirty seconds of "here is what you are looking at, and what I want you to notice" orients someone far better than a paragraph they have to find first.
Ambient sound
A looping background bed — kitchen noise, a busy lobby, birdsong. It sets the scene's atmosphere and is deliberately quieter than narration.
Use it sparingly. Ambient audio that plays under an assessment is a distraction, and learners on shared devices may not be able to turn it down.
Accessibility
Anything said only in audio is unavailable to some of your learners. Write the narration a transcript: a learner can then open it from the scene and turn the sound off, which is what makes it a real alternative rather than a note to yourself. See Recording narration, and writing its transcript.
If the narration carries instructions, it is still worth repeating them in a pin, where they can be found without playing anything.