ChoralSynth

ChoralSynth数据集由庞培法布拉大学音乐技术组创建，旨在解决合唱音乐信息检索研究中缺乏全面数据集的问题。该数据集包含20个合唱作品，每个作品包括完整的音乐XML乐谱、MIDI文件和多轨录音，以及详细的元数据，如声部类型、语言和作曲家信息。数据集的创建过程涉及与多元化的团队合作，包括音乐家、软件工程师和研究人员。应用领域包括音乐信息检索任务，如源分离、节拍检测、旋律分析和和弦分析，旨在推动合唱音乐研究的进步。

The ChoralSynth dataset was developed by the Music Technology Group at Pompeu Fabra University, aiming to address the lack of comprehensive datasets in choral music information retrieval research. This dataset contains 20 choral compositions, each paired with complete MusicXML scores, MIDI files, multi-track recordings, and detailed metadata including voice part types, language, and composer information. The development of the dataset involved collaboration with a diverse team consisting of musicians, software engineers and researchers. Its application scenarios cover music information retrieval tasks such as source separation, beat detection, melody analysis and chord analysis, with the goal of advancing choral music research.

搜集汇总

背景与挑战

背景概述

ChoralSynth是一个合成数据集，包含20首多轨合唱歌曲，通过仔细分析和筛选合成合唱歌曲生成。该数据集可用于音乐信息检索研究，如源分离、旋律分析和和弦分析，遵循非商业CC BY-NC-SA 4.0许可，并可通过Zenodo下载。