Bernini R2V - Character to 2x15s Video

Generates 2x 15-second videos of the same character in the same

setting, using ByteDance Bernini (R2V) with a single character

sheet image as reference.

Input:

- Character sheet image (1 image, referenced as image0)

- Scene-specific prompts (per-cut structure, 2x 15-second scripts)

Process:

- Bernini R2V generates new scenes while preserving character

identity from the reference image

- Two-stage HIGH/LOW noise sampling + LightX2V LoRA for fast

generation (8 steps)

Output:

- 2x 15-second videos (832x480, 16fps)

- Combine both clips in an editor to complete a 30-second concept film

Use case:

- Romance / lifestyle concept film production

- Multi-scene storytelling with a consistent character (dates,

travel, daily life, etc.)