Tempo Induction from Music Recordings Using Ensemble Empirical Mode Decomposition Analysis

摘要

In this article, a system for tempo estimation from audio signals, based on EEMD analysis, was presented. The novel aspect of the proposed approach is the generation of a temporal signature of a music excerpt in different time scales using the EEMD analysis. The signal is decomposed into a series of IMFs that represent rhythmic information in different time scales. Comparing the periodicities in different time scales, the mode representing the time scale closer to tempo is identified and used for tempo extraction. The performance of the system was evaluated by direct comparison against various musical genres.