The year/Independent research

Paper 2512.03041

MultiShotMaster: A Controllable Multi-Shot Video Generation Framework

Published
Dec 2025
Research lab
Independent
Citations
28
GitHub
174 stars

01 In brief

Summary

MultiShotMaster is a framework for controllable multi-shot video generation, extending a pretrained single-shot text-to-video model.

It introduces two RoPE variants: Multi-Shot Narrative RoPE, which applies phase shifts at shot boundaries for flexible shot arrangement while preserving narrative order, and Spatiotemporal Position-Aware RoPE, which enables grounded reference injection by applying region-specific positional encodings to reference tokens.

A Multi-Shot & Multi-Reference Attention Mask manages information flow, and an automated data curation pipeline extracts multi-shot videos, hierarchical captions, grounding signals, and reference images.

The framework supports text-driven inter-shot consistency, customized subjects with motion control, and background-driven scene customization, with configurable shot counts and durations.

Experiments show superior performance over baselines like CineTrans, EchoShot, VACE, and Phantom in text alignment, inter-shot consistency, transition deviation, narrative coherence, and reference consistency.

Limitations include reliance on a ~1B parameter model at 384x672 resolution and coupling between subject motion and camera control.

02 From the paper

Abstract

Current video generation techniques excel at single-shot clips but struggle to produce narrative multi-shot videos, which require flexible shot arrangement, coherent narrative, and controllability beyond text prompts. To tackle these challenges, we propose MultiShotMaster, a framework for highly controllable multi-shot video generation. We extend a pretrained single-shot model by integrating two novel variants of RoPE. First, we introduce Multi-Shot Narrative RoPE, which applies explicit phase shift at shot transitions, enabling flexible shot arrangement while preserving the temporal narrative order. Second, we design Spatiotemporal Position-Aware RoPE to incorporate reference tokens and grounding signals, enabling spatiotemporal-grounded reference injection. In addition, to overcome data scarcity, we establish an automated data annotation pipeline to extract multi-shot videos, captions, cross-shot grounding signals and reference images. Our framework leverages the intrinsic architectural properties to support multi-shot video generation, featuring text-driven inter-shot consistency, customized subject with motion control, and background-driven customized scene. Both shot count and duration are flexibly configurable. Extensive experiments demonstrate the superior performance and outstanding controllability of our framework.