作者: Thomas Etienne Aussenac , Jordi Janer , Perfecto Herrera , Jordi Bonada , Emilia Gomez
DOI:
关键词:
摘要: Systems and methods for determining similarity between two or more audio pieces are disclosed. An illustrative method musical similarities includes extracting one descriptors from each piece, generating a vector of the pieces, features calculating values feature, distance containing normalized vectors outputting response to user process indicating pieces. The can be used in performing content-based classification music. that extracted piece include tonal descriptors, dissonance rhythm spatial descriptors.