Ruffle: A User-Controllable Music Shuffling Algorithm

Music shuffling is a common feature, available in most audio players and music streaming platforms. The goal of this function is to let songs be played in random, or constrained random, order. The results obtained by in-use shuffling algorithms can be unsatisfactory due to several factors including: the variability of user expectations to what constitutes a “successful” playlist, the common bias of being unable to recognize true randomness, and the tendency of humans to find nonexistent patterns in random structures. In this paper, a new shuffling algorithm called Ruffle is presented. Ruffle lets the user decide which aspects of the music library have to be actually shuffled, and which features should remain unchanged between consecutive extractions. First, an online survey was conducted to collect users’ feedback about the characteristics used for shuffling. It is worth noting that, in general, the algorithm could address any metadata and/or audio extracted feature. Then, in order to test the algorithm on personal playlists, a Web version based on Spotify API has been released. For this reason, a second survey is marking an ongoing effort placed on validating the effectiveness of the algorithm by collecting users’ feedback, and measuring the level of user satisfaction.