Towards Describing Perceived Complexity of Songs: Computational Methods and Implementation

Providing valuable semantic descriptors of multimedia content is a topic of high interest for music content processing. Such descriptors should merge the two predicates of (1) being useful for different operations such as retrieval, visual representation of collections, classification, etc., and (2) being automatically extractable from the source. In this paper the semantic descriptor concept music complexity is introduced, and the advantages of their usage for music retrieval and for automated music recommendation are addressed. The authors provide a critical review of existing related proposals and also prospect new methods for automated music complexity estimation.