TalkCuts: A Large-Scale Dataset for Multi-Shot Human Speech Video Generation

Jiaben Chen (University of Massachusetts at Amherst) · Zixin Wang (University of Massachusetts at Amherst) · AILING ZENG (International Digital Economy Academy (IDEA)) · Yang Fu (Fudan University) · Xueyang Yu (University of Massachusetts at Amherst) · Siyuan Cen (University of Massachusetts at Amherst) · Julian Tanke (University of Bonn) · Yihang Chen (University of California, San Diego) · Koichi Saito (Sony AI) · Yuki Mitsufuji (Sony Group Corporation) · Chuang Gan (IBM)

2d keypoints3d smpl-x annotationsaudio-driven settingscamera transitionscinematographic coherencecontrollable generationhuman speech synthesisllm-guided frameworklong-form video synthesismulti-shot video generationmultimodal learningpose-guided settingsspeaker gesticulationstalkcutsvisual appealvocal modulation

In this work, we present TalkCuts, a large-scale dataset designed to facilitate the study of multi-shot human speech video generation. Unlike existing datasets that focus on single-shot, static viewpoints, TalkCuts offers 164k clips totaling over 500 hours of high-quality 1080P human speech videos with diverse camera shots, including close-up, half-body, and full-body views. The dataset includes detailed textual descriptions, 2D keypoints and 3D SMPL-X motion annotations, covering over 10k identities, enabling multimodal learning and evaluation. As a first attempt to showcase the value of the dataset, we present Orator, an LLM-guided multi-modal generation framework as a simple baseline, where the language model functions as a multi-faceted director, orchestrating detailed specifications for camera transitions, speaker gesticulations, and vocal modulation. This architecture enables the synthesis of coherent long-form videos through our integrated multi-modal video generation module. Extensive experiments in both pose-guided and audio-driven settings show that training on TalkCuts significantly enhances the cinematographic coherence and visual appeal of generated multi-shot speech videos. We believe TalkCuts provides a strong foundation for future work in controllable, multi-shot speech video generation and broader multimodal learning.