arXiv:2609.00646 · 评测基准
DramaChain Bench: An End-to-End Benchmark for Short-Drama Generation
DramaChain Bench：面向短剧生成的端到端基准
DramaChain Bench: An End-to-End Benchmark for Short-Drama Generation
- 类型：arxiv
- 标识：2609.00646
- 链接：https://arxiv.org/abs/2609.00646
- 主分类：evaluation
- 形态：benchmark
- TLDR：Commercial short-drama production follows a multi-stage chain: script, storyboard, keyframe imagery, shot-level video, and the finished short drama. Most existing benchmarks evaluate solely the video-generation stage using pre-authored inputs instead of real upstream pipeline outputs. This leaves two critical questions unanswerable: whether each stage adheres to the original script intent (rather than only its immediate input prompt), and whether disparate shots remain coherent after assembly into multi-episode releases. We present DramaChain Bench, the first short-drama benchmark that evaluat
- 副分类：multimodal
- 待LLM分类：否
- 标题中文：DramaChain Bench：面向短剧生成的端到端基准
- TLDR中文：商业短剧制作遵循多阶段链路：剧本、分镜、关键帧图像、镜头级视频，再到成片。现有大多数基准仅在预制输入上评估视频生成阶段，而非真实的上游流水线输出。这留下两个关键问题无法回答：每个阶段是否忠实于原始剧本意图（而非仅其直接输入的提示），以及不同镜头在拼装为多集剧集后是否仍保持连贯。本文提出 DramaChain Bench，这是首个评估
- 来源文件：
- /inbox/tom/_candidates/2026-09-03-agent-rag-longcontext-candidates.json