SGER: A Music Search Engine Based on Aesthetic Similarity

基于审美相似性的音乐搜索引擎这个SGER项目旨在开发一个基于识别审美相似性的音乐搜索引擎原型。该引擎将利用幂定律度量来提取音乐理论和音乐作品的其他属性(例如，音高、持续时间、音调距离、持续时间距离、旋律间隔、和声间隔、旋律二元语法等)的统计比例。引擎使用均方误差(MSE)方法搜索在美学上与输入作品相似的片段。使用古典音乐档案库(14,695首MIDI曲目)，结合500多首其他风格的MIDI曲目(如爵士、摇滚、乡村等)进行了初步测试。类似的衡量标准已经在文本材料的美学属性上得到了验证。文本结果(作者归因、风格识别和愉悦预测)表明准确率很高。将进行评估和验证实验，以与表明检索到的作品的美学相似性的计算结果进行比较。这些实验将由Dwight Krehbiel教授(伯特利学院分奖)进行，他是一位认知神经科学和音乐心理学专家，在测量对音乐的情感和生理反应方面拥有丰富的经验。

AbstractIIS - 0736480Manaris, Bill Z.College of CharlestonA Music Search Engine based on Aesthetic SimilarityThis SGER project aims to develop a prototype music search engine based on identifying aesthetic similarities. This engine will utilize power-law metrics to extract statistical proportions of music-theoretic and other attributes of music pieces (e.g., Pitch, Duration, Pitch Distance, Duration Distance, Melodic Intervals, Harmonic Intervals, Melodic Bigrams, etc.).The engine searches for pieces that are aesthetically similar to the input piece using a mean squared error (MSE) approach. Preliminary testing has been done using the Classical Music Archives corpus (14,695 MIDI pieces), combined with 500+ MIDI pieces from other styles (e.g. Jazz, Rock, Country, etc.). Similar metrics have already been validated on aesthetic attributes of textual materials. Text results (author attribution, style identification, and pleasantness prediction) indicated an high level of accuracy. Assessment and validation experiments will be conducted to compare to computational findings indicating aesthetic similarity of retrieved pieces. These experiments will be conducted by Prof. Dwight Krehbiel (subaward, Bethel College), a specialist in cognitive neuroscience and psychology of music, who has extensive experience in measuring emotional and physiological responses to music.