Design and Evaluation of a Probabilistic Music Projection Interface

We describe the design and evaluation of a probabilistic interface for music exploration and casual playlist generation. Predicted subjective features, such as mood and genre, inferred from low-level audio features create a 34dimensional feature space. We use a nonlinear dimensionality reduction algorithm to create 2D music maps of tracks, and augment these with visualisations of probabilistic mappings of selected features and their uncertainty. We evaluated the system in a longitudinal trial in users’ homes over several weeks. Users said they had fun with the interface and liked the casual nature of the playlist generation. Users preferred to generate playlists from a local neighbourhood of the map, rather than from a trajectory, using neighbourhood selection more than three times more often than path selection. Probabilistic highlighting of subjective features led to more focused exploration in mouse activity logs, and 6 of 8 users said they preferred the probabilistic highlighting mode.