3. Image-to-Video (I2V) Consistency

Anchoring the Generation

Text-to-Video (T2V) often hallucinates. Image-to-Video (I2V) uses a seed image (e.g., from Midjourney) to anchor the first frame, guaranteeing character consistency.

Workflow: Midjourney to Runway

1. Generate character sheet in Midjourney V6.
2. Upscale the chosen frame using Magnific AI.
3. Import to Gen-3 Alpha using I2V, keeping the motion brush subtle.

Common Mistakes

Pushing the motion slider too high in I2V destroys the spatial structure of the anchor image. Keep motion below 40% for realistic character movements.