metadata
license: cc-by-nc-4.0
extra_gated_prompt: >-
SingMoSub is available for non-commercial research purposes only. You agree
not to reproduce, modified, duplicate, copy, sell, trade, resell or exploit
any portion of the dataset and any portion of the derived data for commercial
purposes. You agree not to further copy, publish or distribute any portion of
the SingMoSub dataset to any third party for any purpose. Except, for internal
use at a single site within the same organization it is allowed to make copies
of the dataset.
extra_gated_fields:
First Name: text
Last Name: text
Country: country
Affiliation: text
Email(Institutional Email Only): text
I want to use this dataset for:
type: select
options:
- Research
- Education
I agree to use this dataset for non-commercial use ONLY: checkbox
SingMoSub
Overview
SingMoSub is the first dataset featuring motion subtitles specifically designed for singing-driven 3D head motion. It provides temporally aligned, region-level motion annotations along with acoustic descriptions, enabling detailed modeling of expressive head and facial dynamics in singing scenarios.
The dataset contains over 37 hours of synchronized multimodal data, including:
- Singing audio corresponding to each motion sequence;
- 3D facial motion captured as FLAME parameters (56 dimensions: 50 expression, 3 neck pose, 3 jaw pose);
- Multimodal subtitles for 4 regions motion subtitles (eyebrows, eyes, mouth, neck) and acoustic description subtitles in SRT format.
Data Structure
SingMoSub
├── split.json # Dataset split information
├── audio.zip # Audio files
│ ├── clip_name.wav
│ └── ...
├── flame_seqs # FLAME motion parameter sequences
│ └── motion.pt
├── sub.zip # Multimodal SRT subtitles
│ ├── clip_name
│ │ ├── brow.srt # Brow motion subtitles
│ │ ├── eye.srt # Eye motion subtitles
│ │ ├── mouth.srt # Mouth motion subtitles
│ │ ├── neck.srt # Neck motion subtitles
│ │ └── acoustic_description.srt # Acoustic description subtitles
│ └── ...
Citation
If you use this dataset, please consider citing:
@article{huang2025think2sing,
title={Think2sing: Orchestrating structured motion subtitles for singing-driven 3d head animation},
author={Huang, Zikai and Zhou, Yihan and Xu, Xuemiao and Xu, Cheng and Xing, Xiaofen and Qin, Jing and He, Shengfeng},
journal={arXiv preprint arXiv:2509.02278},
year={2025}
}