AiEson2/SonicGauss

---
license: cc-by-4.0
language:
- en
- zh
tags:
- Audio
- ImpactSound
size_categories:
- 100M<n<1B
---

# SonicGauss Dataset

This dataset contains processed **ObjectFolder 2.0** and **ObjectFolder Real** data for training the [SonicGauss](https://chunshi.wang/SonicGauss/) model - a position-aware physical sound synthesis framework for 3D Gaussian representations.

## Dataset Description

The dataset includes:
- **3D Gaussian Splatting (3DGS) PLY files** extracted from object scans
- **Impact sound recordings** at various positions
- **Rendered images** from multiple viewpoints
- **Training/validation split files** in JSON format

### Dataset Statistics

- **ObjectFolder 2.0**: 1,000 objects with synthetic impact sounds
- **ObjectFolder Real**: Real-world recordings with diverse materials
- **Total Size**: ~24.7 GB (split archive)

## Dataset Structure

```
datas/
├── objectfolder_2.0_train.json
├── objectfolder_2.0_val.json
├── objectfolder_real_train.json
├── objectfolder_real_val.json
├── OF_Real/
│ └── ObjectFolderResults/
└── OF_2.0/
├── audio_results/ # Impact sound recordings (.wav)
└── vision_results/ # 3DGS PLY files (.ply)
```

## Usage

```bash
# Download using huggingface-cli
pip install huggingface-hub
huggingface-cli download AiEson2/SonicGauss --repo-type dataset --local-dir ./datas/
```

## Citation

If you use this dataset, please cite:

```bibtex
@inproceedings{wang2025sonicgauss,
title={SonicGauss: Position-Aware Physical Sound Synthesis for 3D Gaussian Representations},
author={Wang, Chunshi and Li, Hongxing and Luo, Yawei},
booktitle={Proceedings of the 33rd ACM International Conference on Multimedia},
pages={10886--10895},
year={2025}
}

@inproceedings{gao2023objectfolder,
title={The objectfolder benchmark: Multisensory learning with neural and real objects},
author={Gao, Ruohan and Dou, Yiming and Li, Hao and Agarwal, Tanmay and Bohg, Jeannette and Li, Yunzhu and Fei-Fei, Li and Wu, Jiajun},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
pages={17276--17286},
year={2023}
}

@inproceedings{gao2022objectfolder,
title={Objectfolder 2.0: A multisensory object dataset for sim2real transfer},
author={Gao, Ruohan and Si, Zilin and Chang, Yen-Yu and Clarke, Samuel and Bohg, Jeannette and Fei-Fei, Li and Yuan, Wenzhen and Wu, Jiajun},
booktitle={Proceedings of the IEEE/CVF conference on computer vision and pattern recognition},
pages={10598--10608},
year={2022}
}
```

## License

This dataset is derived from ObjectFolder and is distributed under the same terms. Please refer to the original [ObjectFolder](https://objectfolder.stanford.edu/) project for licensing details.