Lyra

Lyra数据集是由国立雅典理工大学电气与计算机工程学院等机构创建，包含1570首希腊传统和民间音乐作品，总计约80小时的数据。数据集内容来源于一部在线希腊纪录片系列，涵盖丰富的元数据信息，如乐器、地理和流派等。创建过程中，通过机器自动化和人工评估/标注相结合的方式，确保数据集的一致性和信息丰富性。Lyra数据集主要用于音乐信息检索（MIR）领域的研究，旨在通过计算分析揭示希腊传统和民间音乐的音乐学特征和现象。

The Lyra Dataset was developed by institutions including the School of Electrical and Computer Engineering of the National Technical University of Athens and other relevant organizations. It contains 1570 works of Greek traditional and folk music, with a total audio duration of approximately 80 hours. The content of the dataset is sourced from an online Greek documentary series, and it includes rich metadata such as musical instruments, geographic origins, musical genres, and other related information. During its development, a combination of automated machine processing and manual evaluation/annotation was employed to ensure the consistency and informational richness of the dataset. The Lyra Dataset is primarily utilized for research in the field of Music Information Retrieval (MIR), with the goal of uncovering musicological characteristics and phenomena of Greek traditional and folk music through computational analysis.