Right now, customers can take heed to music and uncover new artists, songs or albums on quite a lot of music streaming platforms, together with Spotify, Apple Music, Amazon Music Limitless and extra. Many builders have been attempting to create instruments that would enhance these providers, equivalent to music advice methods that recommend new songs or playlists to customers based mostly on their preferences and on music they listened to prior to now.
Researchers at Seoul Nationwide College lately created an interactive knowledge visualization software that would improve each current and rising music streaming providers. This software, referred to as Music Circles, can symbolize songs as distinctive vectors after which calculate similarities between totally different vectors to group comparable songs into clusters.
“As music lovers with totally different tastes, we got here collectively for a challenge that might discover novel methods of visually representing and grouping summary music knowledge,” Seokgi Kim, Jihye Park, Kihong Seong, Namwoo Cho, Junho Min and Hwajung Hong, the researchers who carried out the examine, advised TechXplore through e mail. “We needed to diverge from conventional methods of discovering comparable music by genres, artists, and many others. The core concept was to symbolize songs with numbers by assigning embeddings based mostly on numerical audio function values, equivalent to acousticness and danceability.”
The first goal of the examine carried out by Kim and his colleagues was to assist customers to seek for music they may like and discover music streaming catalogs in methods which are extra intuitive and fascinating. Music Circles, the system they created, calculates the similarity between totally different songs by representing them as vectors, to make searching for out customized music extra entertaining.
“The sequence of interactions and visualizations in our challenge makes knowledge exploration more practical and environment friendly,” the researchers defined. “Our visualizations, which resemble circles (therefore the identify), present attention-grabbing data (e.g., developments in music) based mostly on relationships between audio options of songs.”
Basically, Music Circles arranges songs as totally different cluster visualizations that match the music style of particular person customers. To entry tune clusters aligned with their musical preferences, customers merely have to take a survey about their tune preferences. Music Circles makes use of the information collected by this survey to generate visualizations of tune clusters aligned with a person’s preferences.
“We stray away from the normal music advice outlook (album covers + listing of songs) and supply visualizations of traits of sure clusters,” the researchers stated.
“With applicable annotations and thoroughly chosen designs, we really feel that the challenge is each gratifying and informative. Whereas visualization in music advice is scarce on the whole, our challenge highlights the truth that knowledge visualization could make looking out/shopping for music extra gratifying and efficient.”
In distinction with different music advice methods developed prior to now, Music Circles locations versatile artists who produce a variety of various songs into multiple cluster. For example, if Ed Sheeran’s songs have been to be really useful to customers solely based mostly on what artists they listened to prior to now, his songs would solely be really useful to a restricted viewers. Music Circles, however, locations totally different songs by Ed Sheeran in several clusters, based mostly on their distinctive attributes and traits, thus recommending them to a wider vary of customers.
Sooner or later, the system may very well be used to enhance music streaming providers; for example, permitting customers to achieve a greater understanding of audio options, uncover new songs they may like, view present music developments and uncover what music cluster they belong to. The Music Circles framework is now out there on-line and may be accessed at: https://musiccircles.netlify.app/ .
“As knowledge scientists, we wish to make the most of underrated attributes of songs equivalent to producers, lyricists (individuals who may be extra associated to the music than the precise artist) and supply distinctive music suggestions that differs from streaming powerhouses like Spotify and Apple Music,” the researchers stated. “We additionally wish to make the challenge scalable to massive knowledge. We want to show a bigger set of music in a extra environment friendly method to supply our challenge to extra music lovers.”
Apple Music’s new Replay function will present your most-played music of 2019
Music-Circles: can music be represented with numbers? arXiv: 2102.13350 [cs.HC]. arxiv.org/abs/2102.13350
© 2021 Science X Community
Music Circles: An interactive knowledge visualization software that helps customers uncover new music (2021, March 16)
retrieved 17 March 2021
This doc is topic to copyright. Aside from any truthful dealing for the aim of personal examine or analysis, no
half could also be reproduced with out the written permission. The content material is offered for data functions solely.