Diverse Face Motion Video Dataset

I’m building a fresh dataset of face-motion videos to feed directly into a machine-learning training pipeline. I need original, consented recordings of real people looking toward the camera while moving naturally—turns, nods, blinks, smiles, frowns, and other everyday expressions.

To ensure the model sees a broad range of human variety, every batch you deliver must cover:

- Different age groups
- Different ethnicities
- Various facial expressions

All footage must be supplied in MOV format so it can drop straight into my macOS preprocessing scripts. Raw or lightly compressed files are ideal; audio is optional. Shoot against a clean background with even lighting, keep each take between 5-15 seconds, and avoid stock content or AI-generated faces.

Acceptance criteria
• Each clip satisfies the three-point diversity checklist above
• Video plays flawlessly in QuickTime (MOV)
• A signed appearance release accompanies every participant

Let me know how many unique faces you can source, the approximate turnaround, and any questions about file naming or consent.