Data of research project: About the impact of audio quality and codecs on genre classification and BPM recognition in Essentia

A dataset of songs used in the Research Project of Q2 2021-2022 at the TU Delft. Research paper was about the impact of audio quality and codecs on genre classification and BPM recognition in Essentia. This dataset has been compiled for data collection from Muziekweb. Muziekweb could provide our research group with music in high quality FLAC format. This was essential to this research, since we were researching the impact of audio quality. FLAC is one of the highest quality (lossless) audio codecs.

本数据集为代尔夫特理工大学（TU Delft）2021-2022年第二季度研究项目所使用的歌曲数据集。其配套研究论文聚焦于音频质量与编解码器对Essentia框架内音乐流派分类及每分钟节拍数（BPM，Beats Per Minute）识别的影响。本数据集从Muziekweb平台采集汇编所得，Muziekweb可为本研究团队提供高质量的FLAC（无损音频编解码器）格式音频文件。鉴于本次研究的核心主题为音频质量的影响探究，高保真音频文件是本研究不可或缺的基础条件，而FLAC作为当前顶级的无损音频编解码器之一，恰好契合本研究的需求。