H3 Max can generate audiovisual output from one creative request. Instead of treating sound as a generic add-on, describe the dialogue, ambience, foley, and music cues that belong to visible moments in the shot.
| Layer | What to describe | Example |
|---|---|---|
| Dialogue | Exact short line and speaker | She says, “We leave before sunrise.” |
| Ambience | Continuous environment | Quiet station room tone and distant rain |
| Foley | Sound tied to visible actions | Coat rustle, cup placed on wood, bicycle chain |
| Music cue | Mood and timing, only when useful | One restrained bass pulse at the reveal |
Weak:
Cinematic music and sound effects.Stronger:
As the glass bottle touches the volcanic stone, play one clean glass chime and a short low bass pulse. Keep the rest of the shot to quiet studio room tone; no continuous music.The stronger version explains the source, timing, and relative importance of the sound.
[Speaker and framing] says, “[short exact line].” Keep the camera movement restrained and the face visible through the line. Audio: natural voice close to camera, [room ambience], and [one relevant background sound]; no music.Short dialogue is easier to fit into a short clip. Leave time for the speaker to begin, deliver the line, and finish the visible reaction. Synchronized generation can improve coherence, but it does not guarantee perfect speech, mouth motion, accent, or pronunciation on every run.
no music when natural ambience is the goal.