Traditional music indexing treats music recordings mainly in the same way as text documents: it tries to express content in subject headings. The Internet made it possible to search for music fragments by entering the music itself as a ‘question’. In order to retrieve the music fragments that match this ‘question’ from a database certain mathematical techniques must be applied, which are already implemented in some (experimental) music retrieval sites.
automatic music indexing
query by humming
query by representation
Traditionally, songs and other musical works are indexed by textual elements: the title, the name of the composer, the name of the orchestra or the singer(s), terms ...