K-Means Outperforms Complex Audio Models When Guided by Segmented Musical Features
An empirical evaluation of 25,000 music tracks demonstrates that straightforward K-Means clustering applied to structurally segmented audio features outperforms more ...






