Model based multiple musical sequence alignment using tempo extraction as preprocessing

摘要

In this work, we propose tempo extraction as preprocessing to a model based method for aligning multiple unsynchronized musical audio sequences. Each music excerpt has a certain tempo information and it can be assumed that two sequence with different tempo are not overlapping on a generic timeline. The simulation results on real data sets suggest that the performance of the model based approach increases and computational time decreases with grouping the sequences according to their tempo information.