MV-Forcing: Long Multi-View Video Generation via
4D-Grounded Spatio-Temporal Self-Forcing

Supplementary Material

Please click on the sections below to view our interactive results, comparisons, and ablations:

• Our Results
• Real-World Results
• Short Sequence Comparisons
• Long Sequence Comparisons
• Component Ablations
• View Scaling Ablations
• Temporal Scaling Ablations

Our Results

We show multi-view video generation results from MV-Forcing across diverse scenes and camera trajectories. Each scene shows multiple synchronized views generated autoregressively from a single text prompt. Hover over any video to play.

"A person wearing a white shirt and black pants, performing stretching exercises in a dimly lit room with wooden floors."

"A woman in black performing yoga poses in front of a modern building with large windows and trees casting shadows."

"A man in a t-shirt is dancing in a room with large windows and filled with large plants."

"A man in a black t-shirt, blue shorts and wearing sneakers is moving his arms and legs in various directions in front of a modern building with a tree casting shadows on the ground."

"A man in a black t-shirt with a logo on the front and blue jeans is performing a series of dance moves near a large brick wall, a metal staircase, and stacked wooden pallets."

"A woman performing dancing in an urban setting with industrial structures in the background."

"A man in a black sweatshirt dancing in a dimly lit storage room with a hanging light bulb and metal shelves."

"A man in a black shirt stretching his arms and legs in a modern room with large windows."

"A man in a gray t-shirt and black shorts stretching in a clothing store."

"An older man in a white t-shirt and dark pants dancing inside a vintage diner with a counter, stools, and jukebox."