Jul 2026
Moving Alphabet: A Controlled Study of Training Data for Text-to-Video Generation
This work introduces Moving Alphabet, a procedural testbed that renders letters with varying fonts, colors, sizes, and positions, moving in different directions and speeds against a black background, to study how data distribution and caption quality impact text-to-video models.
Amber Yijia Zheng, Lu Liu, Raymond A. Yeh et al.
· arXiv.org · 1 citation