SingMoSub: 37 Hours of Singing-Driven 3D Head Motion with Subtitles

Over 37 hours of synchronized multimodal data for singing-driven 3D head motion, featuring motion subtitles and acoustic descriptions. The dataset, named SingMoSub, was created by ZikaiHuang and was last updated on March 1, 2026. It provides temporally aligned, region-level motion annotations for modeling expressive head and facial dynamics.

Use Cases

• Train models for singing-driven 3D head motion generation based on synchronized audio and motion sequences.
• Develop region-level facial animation controls based on the provided motion subtitle annotations.
• Research multimodal alignment between audio features and expressive facial dynamics in singing scenarios.
• Create systems for generating acoustic descriptions of motion based on the provided annotations.

Strengths

• Over 37 hours of synchronized multimodal data, providing substantial material for training.
• First dataset featuring motion subtitles specifically for singing-driven 3D head motion.
• Provides temporally aligned, region-level motion annotations for detailed modeling.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for specific model training needs.

Provenance

Source

ZikaiHuang on Hugging Face

Collection Method

Likely involves motion capture and audio recording of singing performances, with subsequent annotation.

Time Range

null

Freshness

Last updated 2026-03-01 12:07:12; freshness should be verified.

Geography

null