EMIR

EMIR是由西北大学信息科学与技术学院创建的第一个针对埃塞俄比亚音乐的信息检索数据集，包含600个样本录音，涵盖正统特瓦希多吟唱、传统阿兹马里歌曲和现代埃塞俄比亚世俗音乐。数据集由五位专家评委对每个样本进行分类，分为四个著名的埃塞俄比亚Kiñits：Tizita、Bati、Ambassel和Anchihoye。每个Kiñit使用其独特的五声音阶和风格特征，使得Kiñit分类需要结合音阶识别与流派识别。数据集旨在解决埃塞俄比亚音乐分类的挑战，特别是缺乏训练数据的问题，并推动埃塞俄比亚音乐的研究和模型实验。

EMIR is the first information retrieval dataset focused on Ethiopian music, developed by the School of Information Science and Technology of Northwest University. It consists of 600 audio recordings spanning three categories: Orthodox Tewahedo chanting, traditional Azmari songs, and modern Ethiopian secular music. Each sample was categorized by five expert judges into four prominent Ethiopian Kiñits, namely Tizita, Bati, Ambassel, and Anchihoye. Each Kiñit possesses its unique pentatonic scale and stylistic characteristics, requiring Kiñit classification to combine scale recognition and genre identification. This dataset aims to address the challenges in Ethiopian music classification, particularly the lack of training data, and promote research on Ethiopian music and model experimentation.