VibraVerse is a large-scale multimodal dataset designed to bridge 3D geometry, material physics, and acoustics. It explicitly encodes the causal chain from geometry to acoustic signals, unlike unconstrained audiovisual recordings. The dataset was created by technetium66 and was last updated on 2026-01-03.
Use Cases
- Train multimodal models for geometry-to-acoustic synthesis based on the described causal chain.
- Develop physically-informed neural networks for material property prediction based on 3D geometry and acoustic signals.
- Benchmark multimodal alignment algorithms using the dataset's geometry-acoustics pairs.
- Simulate acoustic environments for virtual or augmented reality applications based on 3D object data.
Strengths
- Dataset is described as large-scale, suggesting a substantial number of samples.
- It is designed with a physically-consistent causal chain from geometry to acoustics.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
- Source
- huggingface user technetium66
- Freshness
- Last updated 2026-01-03 19:44:36; freshness should be verified.