jaCappella Corpus
收藏arXiv2023-02-24 更新2024-06-21 收录
下载链接：
https://tomohikonakamura.github.io/jaCappella_corpus/
下载链接
链接失效反馈官方服务：
资源简介：
jaCappella Corpus是一个专门为日本无伴奏合唱团分离和合成而构建的数据集。该数据集包含35首版权清晰的合唱团歌曲及其个人声部的音频记录，这些歌曲改编自无版权的日本儿童歌曲，包含六个声部。数据集分为七个子集，每个子集代表一种典型的音乐风格，如爵士乐和演歌。该数据集反映了社交媒体服务中广泛传播的合唱团风格，适用于合唱团分离的研究。
The jaCappella Corpus is a dataset specifically constructed for the separation and synthesis of Japanese a cappella choirs. This dataset contains 35 copyright-cleared choral songs along with audio recordings of their individual vocal parts. These songs, which consist of six vocal parts, are adapted from copyright-free Japanese children's songs. The dataset is divided into seven subsets, each representing a typical musical genre such as jazz and enka. This dataset reflects the choral styles widely spread on social media services, and is suitable for research on choral separation.
提供机构：
东京大学信息科学与技术研究生院
创建时间：
2022-11-29
