Data-Driven Chord-Sequence Representations of Songs and Applications

Chih-Li Wang, Qian Zhong, Szu-Ying Wang, and Vwani Roychowdhury


Hidden Markov Model, Smith-Waterman Algorithm, Cover song, Music Summarization


Content-based music analysis has drawn much attention due to the rapidly growing digital music market. We introduce a method for determining approximate chord-sequence representations of songs, purely from the chromagram of songs and without assuming any prior music knowledge. Each song is represented by a sequence of states, which represent the different chords, of an underlying Hidden Markov Model (HMM). The method is then applied to the problems of cover song identification and music summarization by robust and local sequence alignment algorithms. The proposed method has a number of advantages, including elimination of the unreliable beat estimation step and the ability to match parts of songs. The invariance of key transpositions among cover songs is achieved by cyclically rotating the chromatic domain of a chromagram. Our data driven method is shown to be robust to reordering, inserting, deleting parts of songs, and provide superior performance compared with other known methods for cover song identification.

Important Links:

Go Back