# Seedance audio reference prompting

Canonical: https://ai-video.holycrab.ai/seedance/capabilities/audio-reference/
Markdown: https://ai-video.holycrab.ai/seedance/capabilities/audio-reference.md
Last updated: 2026-06-19
Source kind: official_manual
Rights status: embed_only
Curation: published
HolyCrab conversion: https://holycrab.ai/?model=seedance&utm_source=ai-video-trends&utm_medium=seedance_seo&utm_campaign=seedance_prompt_playbook&utm_content=seedance-capabilities-audio-reference&source_kind=official_manual&page_kind=capability&prompt=Seedance+2.0+handoff+for%3A+Seedance+audio+reference+prompting.+Goal%3A+Audio+references+can+guide+mood%2C+pacing%2C+and+audio-video+alignment+when+the+prompt+explains+how+the+sound+should+influence+the+shot..+Prompt+structure%3A+subject%2C+motion%2C+camera+path%2C+reference+role%2C+lighting%2C+duration%2C+and+final+payoff.

Citation summary: Use audio references for mood-matched trailers, music-driven pacing, sound-synced product reveals, and character moments with clear beats.

## Direct Answer

Audio references can guide mood, pacing, and audio-video alignment when the prompt explains how the sound should influence the shot.

## Source Attribution

- ByteDance Seed official Seedance 2.0 model page: https://seed.bytedance.com/en/seedance2_0
- Seedance 2.0: Advancing Video Generation for World Complexity: https://arxiv.org/abs/2604.14148

## Publication And Release Gate

- Curation gate: this page can stay public only while curationStatus is published.
- Rights gate: source rights must be known, non-blocked, and attribution must remain visible.
- Discovery gate: public HTML, Markdown, llms.txt, sitemap, and search-index entries are generated only from published records.
- Official-source policy: paraphrase BytePlus or paper material and cite the original source.

## Prompt Reuse Checklist

- Name the exact input type: text, image, audio, video, first frame, or final frame.
- State what each reference controls so the model does not blend competing instructions.
- Use one camera move, one main action, and one final-frame payoff.
- Test the smallest version in HolyCrab before expanding the shot.

## Prompt Utility

- Evidence status: official-source paraphrase
- Confidence: official
- Reusable pattern: Audio references can guide mood, pacing, and audio-video alignment when the prompt explains how the sound should influence the shot.
- Rewrite rule: Keep the capability structure, then replace subject, reference role, camera, action, and final-frame payoff.
- Source evidence: Use audio references for mood-matched trailers, music-driven pacing, sound-synced product reveals, and character moments with clear beats.
- HolyCrab next step: Test the smallest official-style Seedance prompt in HolyCrab before adding more reference inputs.

## Prompt pattern

Explain the emotional tone, beat timing, and visual moments that should react to the audio.

## Example prompt

Create a 12 second neon street dance clip. Match the camera cuts to the snare hits, keep the dancer centered, and end on a wide pose as the final chord lands.

## Common failure modes

Audio is referenced but visual timing is not specified; Too many scene cuts; Dialogue or sound intent is ambiguous

## HolyCrab Handoff

1. Start with the source-attributed Seedance capability or example brief.
2. Name each input type and what it should control.
3. Write one camera path and one final-frame payoff.
4. Generate a short HolyCrab test, then revise only the weakest prompt variable.

## HolyCrab Handoff Prompt

```text
Seedance 2.0 handoff for: Seedance audio reference prompting. Goal: Audio references can guide mood, pacing, and audio-video alignment when the prompt explains how the sound should influence the shot.. Prompt structure: subject, motion, camera path, reference role, lighting, duration, and final payoff.
```

## Related Seedance Pages

- Audio-synced Seedance trailer prompt: https://ai-video.holycrab.ai/seedance/examples/audio-synced-trailer/ | A short trailer prompt that uses an audio reference to shape pacing and visual beats.