mshoxxDB - a Versioned Dataset for Electronic Music

Description: Version 1 consists of 18 full-length pieces of music in the genre of Electronic Music, totaling 61 minutes of audio. The dataset spans a number of sub-genres of Electronic Music: Video Game, 8-Bit (Chiptune), EDM, Pop, House, Chillout/Dreamy. The dataset is suited for a variety of tasks in the field of Music Information Retrieval (MIR), such as Source Separation, Multi-Pitch Estimation, Beat Detection, Tempo Estimation. It is particularly interesting for instrument-agnostic methods and evaluations of model generalization due to the wide variety of synthetic and traditional timbres.

Contents:- mixtures and multi-tracks in FLAC format (44.1 kHz, 16-Bit, Mono, compression level 6)- track-wise MIDI files- CSV metadata with genre, tempo, time signature, and artist information

Technical Properties: Not all mixtures are exact sums of their respective multi-tracks. The mixtures may contain additional processing in the form of limiters and compression (on the full mix or side-chain compression from one track to another). No harmonic effects were added onto the mixtures (= effects such as reverbs, echos, and delays that add more harmonic information and would result in mismatches between MIDI and audio).

License: All contents distributed under Creative Commons BY-NC-SA 4.0.

Demo Page: For a few listening examples, please visit this dataset's github page at https://mic-tae.github.io/mshoxxdb/.

Repo: The mshoxxDB repo is located at https://github.com/mic-tae/mshoxxdb, and may contain more up-to-date information (README.md).

Citation: Should you use this dataset in your work, please cite it the following way (bibtex):

@misc {taenzer:mshoxxDB:2024,  author = {Taenzer, Michael},  title = {{mshoxxDB - a Versioned Dataset for Electronic Music}},  booktitle = {{Late-Breaking and Demo Session of the 25th International Conference on Music Information Retrieval (ISMIR)}},  address = {{San Francisco, CA, USA}},  year = {2024},}

描述：版本1包含18首完整时长的电子音乐（Electronic Music）作品，总音频时长共计61分钟。本数据集涵盖多个电子音乐子流派：电子游戏音乐、8位（Chiptune，芯片音乐）、电子舞曲（EDM）、流行、浩室、弛放/梦幻风格。其适用于音乐信息检索（Music Information Retrieval, MIR）领域的多项任务，例如源分离、多音高估计、节拍检测、速度估计。由于涵盖了丰富的合成音色与传统乐器音色，本数据集尤其适合与乐器无关的模型方法研发，以及模型泛化能力的评估。

数据集内容：
- FLAC格式的混音轨与分轨音频（采样率44.1kHz，位深16比特，单声道，压缩等级6）
- 按曲目划分的MIDI文件
- 包含曲风、速度、拍号与艺术家信息的CSV元数据文件

技术特性：并非所有混音轨都严格等于对应分轨的直接叠加。混音轨可能经过额外处理，包括限幅器处理与压缩（如对整体混音的压缩，或是轨道间的侧链压缩）。未向混音轨添加谐波类效果（即诸如混响、回声与延迟这类会引入额外谐波信息，导致MIDI与音频不匹配的效果）。

授权协议：所有数据集内容均采用知识共享署名-非商业性使用-相同方式共享4.0（Creative Commons BY-NC-SA 4.0）协议分发。

演示页面：如需试听部分样例，请访问本数据集的GitHub页面：https://mic-tae.github.io/mshoxxdb/。

代码仓库：mshoxxDB 仓库地址为 https://github.com/mic-tae/mshoxxdb，其中可能包含更新的信息（如README.md文件）。

引用方式：若您在研究工作中使用本数据集，请按如下BibTeX格式进行引用：
@misc {taenzer:mshoxxDB:2024, author = {Taenzer, Michael}, title = {{mshoxxDB - a Versioned Dataset for Electronic Music}}, booktitle = {{Late-Breaking and Demo Session of the 25th International Conference on Music Information Retrieval (ISMIR)}}, address = {{San Francisco, CA, USA}}, year = {2024},}