The street-performance recipe
Plan the performer, speaker, crowd and camera together. The street preset keeps the lead in a waist-up frame with a reacting audience behind. Photo remix emphasizes a readable face; crowd hype gives the audience more reaction; football crew uses invented jersey-wearing extras.
Reference guidance and variation
A reference clip can guide the motion and camera when the service is configured for that workflow. A photo-only task relies on the preset description. Both are generative: face shape, gestures and background details can drift. Review your own result rather than expecting the gallery clip with only one face changed.
Choose a preset by its actual shot plan
All four presets request one continuous 15-second shot with restrained movement. Start with the scene whose framing matches your portrait, then keep the same photo when comparing another scene. Changing photos, framing and models together makes it difficult to tell which choice changed the result. These descriptions summarize our configured scripts, not measured quality scores.
| Preset | Configured action | What to review |
|---|---|---|
| Street freestyle | Waist-up, steps and open palms | Hands stay clear of the face |
| Photo remix | One clap, two gestures, relaxed smile | Face remains recognizable |
| Crowd hype | Medium-wide frame and small cheers | Crowd stays behind the lead |
| Football crew | Generic jerseys and a celebratory pose | Photo clothes and face are retained |
Sources and editorial notes
- EvoLink: Seedance 2.0 Mini reference-to-video
- Bulin47-ai.com: third-party examples
- Bulin47ai.org: freestyle reference
Instructions are based on the controls and configuration of this studio. Reference media comes from the credited third-party pages; it is not our own generation test. Planning prompts do not change the built-in scripts. External documentation is linked where used. Read our editorial approach