Clef: An Extensible, Experimental Framework for Music Information Retrieval

music;information retrieval;symbolic data;search

DeCurtins, Max

University:Havard University

Department:Software Engineering

关键词: music;    information retrieval;    symbolic data;    search;   

Others  :  https://dash.harvard.edu/bitstream/handle/1/37364558/DECURTINS-DOCUMENT-2018.pdf?sequence=1&isAllowed=y

美国|英语

来源: Digital Access to Scholarship at Harvard

【 摘 要 】

Text-based searching of textual data has long been a mainstay of computing, andas search technology has evolved so has interest in searching non-textual data. In recentyears efforts to use image files as queries for other image files (or for information aboutwhat is in the query image file) have profited from advances in machine learning, as haveother alternative search domains.While searching for music using musical data has met with considerable successin audio sampling software such as Shazam, searching machine-readable, music notation-based data—also known as symbolic music data—using queries written in musicnotation has lagged behind, with most development in this area geared toward academicmusic researchers or existing in ad hoc implementations only. Music information retrieval—the field concerned with developing search techniques for music—requires a frameworkthat can move beyond predetermined combinations of algorithms and datasets.The Clef system demonstrates that this is possible using container-based servicesthat communicate with each other over HTTP. Clef offers an extensible approach tobuilding a musical search engine that allows new algorithms and datasets to be accessedthrough a consistent, music notation-based user interface for query input. Extending thesystem with a new container for running a music information retrieval algorithm requiressignificant development, but once operational, new algorithm containers integrate seamlesslyinto the user interface.