README.md

---
license: cc-by-nc-4.0
task_categories:
- audio-to-audio
language:
- en
pretty_name: Diff-SSL-G-Comp
size_categories:
- 1K<n<10K
extra_gated_prompt: >-
  Terms of Access: The researcher has requested permission to use the Diff-SSL-G-Comp
  dataset. In exchange for such permission, the researcher hereby agrees to the following
  terms and conditions:
  
  1. The researcher shall use the dataset under the CC-BY-NC license. The
     researcher shall also conform to the usage constraints and limitations 
     listed in the Cambridge Multitrack Library。
  2. The authors make no representations or warranties regarding the datasets,
     including but not limited to warranties of non-infringement or fitness for
     a particular purpose.
  3. The researcher accepts full responsibility for their use of the dataset and
     shall defend and indemnify the authors of Diff-SSL-G-Comp, including their 
     employees, trustees, officers, and agents, against any and all claims arising 
     from the researcher's use of the datasets, including but not limited to the 
     researcher's use of any copies of copyrighted content that they may create 
     from the datasets.
  4. The researcher may provide research associates and colleagues with access
     to the datasets, provided that they first agree to be bound by these terms
     and conditions.
  5. The authors reserve the right to terminate the researcher's access to the
     datasets at any time.
  6. If the researcher is employed by a for-profit, commercial entity, the
     researcher's employer shall also be bound by these terms and conditions,
     and the researcher hereby represents that they are fully authorized to enter
     into this agreement on behalf of such employer.
extra_gated_fields:
  Name: text
  Email: text
  Affiliation: text
  Position: text
  Your Supervisor/manager/director: text
  I agree to the Terms of Access: checkbox
---

# Diff-SSL-G-Comp: Towards a Large-Scale and Diverse Dataset for Virtual Analog Modeling

This is the official repository 👑 for the **Diff-SSL-G-Comp** dataset. 

## Diff-SSL-G-Comp Overview ⭐️

The **Diff-SSL-G-Comp** dataset is an extensive and diverse dataset with the following features:
- we use *real-world* unmastered songs as input instead of noises and test signals;
- *175* songs are selected with diverse genres, instruments, tempos, and keys;
- comprehensive *220* parameter combinations are recorded, resulting in *2528* hours of data.

## Diff-SSL-G-Comp Dataset Structure ⛪️

Diff-SSL-G-Comp is formatted using the following structure:

```
|-- cambridge_unmastered_songs
|   |-- 54_UnmasteredWAV.wav
|   |-- ...
|-- processed_normalized
|   |-- 54_UnmasteredWAV.wav
|   |-- ...
|-- ...
|-- processed_ground_truth
|   |-- threshold_-12_attack_10_release_0.1_ratio_2
|   |   |-- 54-exported.wav
|   |   |-- ...
|   |-- ...
|-- ...

```

The raw audios are in the folder "cambridge_unmastered_songs." The normalized input audios are in the folder 
"processed_normalized." The ground truth hardware output audios are in the folder "processed_ground_truth."

Note that we used annotated thresholds illustrated on the analog device instead of the digital workstation.

## Reference 📖
If you use the Diff-SSL-G-Comp dataset, please cite the following papers:

```bibtex
@inproceedings{diffsslgcomp,
    author={Yicheng Gu and Runsong Zhang and Lauri Juvela and Zhizheng Wu},
    title={Diff-SSL-G-Comp: Towards a Large-Scale and Diverse Dataset for Virtual Analog Modeling},
    booktitle={arXiv:2504.04589},
    year={2025}
}
```