Vocadito

Vocadito数据集由声学研究所创建，包含40段单声道歌唱短片段，涵盖7种不同语言，由不同训练水平的歌手演唱，并使用多种设备录制。数据集提供了包括f0、歌词和两种不同音符注释在内的多种注释类型。创建过程中，所有注释均由音乐家完成，旨在解决歌唱声音的音符估计问题，特别是在音符和f0注释之间的交互研究。

The Vocadito dataset was created by the Acoustic Research Institute. It comprises 40 monophonic short singing segments covering seven distinct languages, performed by singers with varying training levels and recorded using multiple devices. The dataset provides a range of annotation types, including f0, lyrics, and two distinct note annotations. All annotations were completed by musicians during the dataset's development, with the objective of addressing the problem of singing voice note estimation, particularly for research on the interaction between note and f0 annotations.