Music, sound effects & dialogue

Build a soundtrack on Sunra beyond voiceover: generate music, design sound effects, re-voice clips with the voice changer, script multi-voice dialogue, and mix the layers.

Voiceover and dubbing are the speech side of the audio studio. This page covers the rest of the soundtrack — music, sound effects, the voice changer, and multi-voice dialogue. For the full studio overview, start with Voiceover and audio basics.

The audio studio's composer bar — the left dropdown switches mode (music, sound effects, voice changer, dialogue); the others pick provider and voice.
The audio studio's composer bar — the left dropdown switches mode (music, sound effects, voice changer, dialogue); the others pick provider and voice.

Music

Generate a backing track in the music mode (also a standalone AI music generator). Prompt a style and energy, not a specific song:

Warm lo-fi, relaxed, mellow keys — for a calm product montage.
Driving electronic, upbeat, punchy bass — for a sneaker ad.
  • Toggle instrumental when you don't want vocals competing with a voiceover.
  • Set the length to match your edit.
Tip
Describe the *feeling and the use* ("driving electronic, upbeat, for a sneaker ad") rather than naming an artist. Naming a specific artist, song, or copyrighted lyrics doesn't just work less well — the music model rejects it outright. Mood-and-purpose prompts land reliably and never trip the filter.

Sound effects

Describe a sound in the sound effects mode (or the AI sound effect generator) — "heavy rain on a tin roof," "sci-fi door whoosh," "distant city traffic." You can set the clip length and how strictly it follows the prompt.

Layer effects under a clip to make a silent render feel alive — footsteps, ambience, a single accent hit on a cut.

Voice changer

Already have a recording? The voice changer re-voices it in a different voice while keeping your timing and delivery — useful for anonymizing or restyling narration. There's a denoise option to clean up a rough source first.

Dialogue

The dialogue mode generates a multi-voice conversation: write the script line by line and assign a different voice to each speaker (there are 37 voices to choose from). Good for skits, explainer back-and-forths, and character scenes.

Mixing the layers

A finished mix usually stacks three things, loudest to quietest:

  1. Voice — a voiceover or dialogue track, the loudest element and the one viewers follow.
  2. Music — a bed underneath, instrumental and noticeably quieter so it never fights the voice.
  3. Effects — accents and ambience for texture, lowest of all.

Generate each, then assemble and balance them against your visuals in Flow or Studio.

Note
For simple clips you can skip the separate audio pass entirely: generate with an audio-native video model like Veo 3.1 or Kling 3.0, which produce picture and sound together — see the AI video with audio use case.

Related articles