Cover Song Identification with Timbral Shape Sequences
This addresses the problem of identifying cover songs for music information retrieval, offering an alternative to existing note-based methods.
The paper tackles cover song identification by introducing a novel low-level feature based on timbral shape sequences, achieving correct identification of 42 out of 80 songs in the Covers 80 dataset.
We introduce a novel low level feature for identifying cover songs which quantifies the relative changes in the smoothed frequency spectrum of a song. Our key insight is that a sliding window representation of a chunk of audio can be viewed as a time-ordered point cloud in high dimensions. For corresponding chunks of audio between different versions of the same song, these point clouds are approximately rotated, translated, and scaled copies of each other. If we treat MFCC embeddings as point clouds and cast the problem as a relative shape sequence, we are able to correctly identify 42/80 cover songs in the "Covers 80" dataset. By contrast, all other work to date on cover songs exclusively relies on matching note sequences from Chroma derived features.