Haoyu123123/loom-of-time-data

Loom of Time — Training Data

Spectrogram dataset used to train the ResNet18 regional classifier in Loom of Time (岁月织机), presented at ACM Multimedia 2026 (Interactive Art Track).

GitHub repository: HAoyu-baller/Loom-of-Time-Music-Viz
Model weights: Haoyu123123/loom-of-time-models

---

Contents

spectrograms/
  jiangnan/   # 江南 — Jiangnan (Yangtze River Delta)
  shanxi/     # 陕西 — Shaanxi (Northwest)
  yungui/     # 云贵 — Yungui (Yunnan-Guizhou Plateau)
  huanan/     # 华南 — Huanan (South China)
  dongbei/    # 东北 — Dongbei (Northeast)

Each file is a 224×224 PNG mel-spectrogram image (magma colormap, 128 mel bins, hop length 512, FFT size 2048) generated from a 6-second audio clip at 22,050 Hz.

---

Dataset Statistics

Region	Clips
Jiangnan	varies
Shaanxi	varies
Yungui	varies
Huanan	varies
Dongbei	varies

Audio sources: regional Chinese folk music recordings. See the GitHub README for the full data pipeline documentation.

---

License

CC BY-NC 4.0 — non-commercial research use only.

---

Citation

@inproceedings{ruan2026loom,
  title     = {Loom of Time: Real-Time Chinese Folk Music Decomposition
               and Generative Weaving Visualization},
  author    = {Ruan, Haoyu},
  booktitle = {Proceedings of the 34th ACM International Conference on
               Multimedia (Interactive Art Track)},
  year      = {2026},
  publisher = {ACM}
}