MOSA
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/yufenhuang/MOSA-Music-mOtion-and-Semantic-Annotation-dataset
下载链接
链接失效反馈官方服务：
资源简介：
该数据集名为MOSA，包含了高质量的三维动作捕捉数据、同步的音频录音以及针对742场专业音乐表演的逐音符语义标注。这些表演由23位专业音乐家完成，数据总量超过30小时，包含57万音符，是迄今为止最大的具有音符级别标注的跨模态音乐数据集。该数据集适用于进行跨模态音乐信息检索（Mir）以及音乐内容生成任务。
This dataset, named MOSA, comprises high-quality 3D motion capture data, synchronized audio recordings, and note-level semantic annotations for 742 professional music performances. Performed by 23 professional musicians, the dataset has a total duration of over 30 hours and contains 570,000 musical notes, making it the largest cross-modal music dataset with note-level annotations to date. This dataset is applicable to cross-modal music information retrieval (MIR) and music content generation tasks.
