Abstrakt
Interactive music retrieval systems using microphone input have become popular, with applications ranging from whistle queries to robust audio search engines capable of retrieving music from a short sample recorded in noisy environment. The availability for mobile devices brought them to millions of users.
Underlying methods have promising results in the case that user provides a short recorded sample and seeks additional information about the piece. Now, the focus needs to be switched to areas where we are still unable to satisfy the user needs.
Such a scenario can be the choice of a favorite music performance from the set of covers, or recordings of the same musical piece, e.g. in classical music. Various algorithms have been proposed for both basic retrieval and more advanced use cases.
In this paper we provide a survey of the state-of-the-art methods for interactive music retrieval systems, from the perspective of specific user requirements.