Abstract
An important source of intelligence for music emotion recognition today comes from user-provided community tags about songs or artists. Recent crowdsourcing approaches such as harvesting social tags, design of collaborative games and web services or the use of Mechanical Turk, are becoming popular in the literature. They provide a cheap, quick and efficient method, contrary to professional labeling of songs which is expensive and does not scale for creating large datasets. In this paper we discuss the viability of various crowdsourcing instruments providing examples from research works. We also share our own experience, illustrating the steps we followed using tags collected from Last.fm for the creation of two music mood datasets which are rendered public. While processing affect tags of Last.fm, we observed that they tend to be biased towards positive emotions; the resulting dataset thus contain more positive songs than negative ones.
Highlights
Music Information Retrieval (MIR) and Music Emotion Recognition (MER) are two important research directions that are changing the way people find and listen to music
Noisy and imperfect as they might be, social tags are a source of human-generated contextual information that is become an essential part of the solution to many MIR and MER problems
In this paper we presented the various crowdsourcing approaches that are being experimented for collecting subjective human judgment about emotionality of musical pieces
Summary
Music Information Retrieval (MIR) and Music Emotion Recognition (MER) are two important research directions that are changing the way people find and listen to music. A social tag is just a free text label that a user applies in music-related objects like songs, artists or albums. These tags capture contextual and descriptive information about the resource they are associated to. Noisy and imperfect as they might be, social tags are a source of human-generated contextual information that is become an essential part of the solution to many MIR and MER problems. For this reason, many researchers and developers are experimenting with different ways of obtaining tags.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have