Research map / Video / Video generation
ShotStream: Streaming Multi-Shot Video Generation for Interactive Storytelling
Yawen Luo, Xiaoyu Shi, Jun-hao Zhuang, Yutian Chen, Quande Liu, Xintao Wang, Pengfei Wan, Tianfan Xue
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add dynamic context-frame sampling vs first-only frames (ablation) | bidirectional teacher · Multi-shot eval, teacher ablation | inter-shot semantic consistency: 0.671 -> 0.762 | Improved |
| Inject per-shot captions for condition frames (ablation) | bidirectional teacher · Multi-shot eval, teacher ablation | aesthetic quality: 0.422 -> 0.571 | Improved |
| Frame concatenation instead of channel concatenation (ablation) | bidirectional teacher · Multi-shot eval, teacher ablation | transition control: 0.912 -> 0.978 | Improved |
| Fine-tune only 3D attention vs full fine-tuning (ablation) | bidirectional teacher · Multi-shot eval, teacher ablation | dynamic degrees: 60.85 -> 63.06 | Improved |
| Add RoPE discontinuity indicator between global and local caches (ablation) | causal student · Multi-shot eval, causal student ablation | inter-shot subject consistency: 0.507 -> 0.654 | Improved |
| RoPE offset vs learnable embedding for cache distinction (ablation) | causal student · Multi-shot eval, causal student ablation | inter-shot subject consistency: 0.518 -> 0.654 | Improved |
| Add inter-shot self-forcing stage after intra-shot stage (ablation) | causal student · Multi-shot eval, causal student ablation | inter-shot subject consistency: 0.604 -> 0.654 | Improved |
| Add intra-shot stage before inter-shot stage (ablation) | causal student · Multi-shot eval, causal student ablation | dynamic degrees: 52.86 -> 63.06 | Improved |
| Causal ShotStream vs bidirectional multi-shot models | ShotStream · Multi-shot eval (100 Gemini prompts) | FPS: 0.643 -> 15.95 | Improved |
| ShotStream vs best baseline on transition control | ShotStream · Multi-shot eval (100 Gemini prompts) | transition control: 0.715 -> 0.978 | Improved |
| ShotStream vs EchoShot on dynamic degrees | ShotStream · Multi-shot eval (100 Gemini prompts) | dynamic degrees: 65.92 -> 63.56 | Hurt |
Later papers that compared with this method
1 later accepted paper reports ShotStream as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- CineOrchestra: Unified Entity-Centric Conditioning for Cinematic Video Generation (NeurIPS 2026 · better on 3 of 3 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most shared results.
- LongLive — LongLive: Real-time Interactive Long Video Generation (ICLR 2026) · better on 2 of 3 shared results
- Mask2DiT — Mask^2DiT: Dual Mask-based Diffusion Transformer for Multi-Scene Long Video Generation (CVPR 2025) · better on 3 of 3 shared results
- Rolling Forcing — Rolling Forcing: Autoregressive Long Video Diffusion in Real Time (ICLR 2026) · better on 2 of 2 shared results
- Infinity-RoPE — Infinity-RoPE: Action-Controllable Infinite Video Generation Emerges From Autoregressive Self-Rollout (CVPR 2026) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- MovieDreamer: Hierarchical Generation for Coherent Long Visual Sequences (ICLR 2025 · similarity 0.87)
- ShotAdapter: Text-to-Multi-Shot Video Generation with Diffusion Models (CVPR 2025 · similarity 0.87)
- StreamingT2V: Consistent, Dynamic, and Extendable Long Video Generation from Text (CVPR 2025 · similarity 0.87)
- EntityBench: Towards Entity-Consistent Long-Range Multi-Shot Video Generation (NeurIPS 2026 · similarity 0.86)
- Advancing Narrative Long Video Generation via Training-Free Identity-Aware Memory (NeurIPS 2026 · similarity 0.86)
- Storybooth: Training-Free Multi-Subject Consistency for Improved Visual Storytelling (ICLR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
