New Journal of Physics, vol. 11 (2009)
There is growing evidence that nonlinear time series analysis techniques can be used to successfully\r characterize, classify, or process signals derived from real-world dynamics even though these are\r not necessarily deterministic and stationary. In the present study, we proceed in this direction by\r addressing an important problem our modern society is facing, the automatic classification of\r digital information. In particular, we address the automatic identification of cover songs, i.e.\r alternative renditions of a previously recorded musical piece. For this purpose, we here propose a\r recurrence quantification analysis measure that allows the tracking of potentially curved and\r disrupted traces in cross recurrence plots (CRPs). We apply this measure to CRPs constructed from\r the state space representation of musical descriptor time series extracted from the raw audio\r signal. We show that our method identifies cover songs with a higher accuracy as compared to\r previously published techniques. Beyond the particular application proposed here, we discuss how our\r approach can be useful for the characterization of a variety of signals from different scientific\r disciplines. We study coupled Rössler dynamics with stochastically modulated mean frequencies as one\r concrete example to illustrate this point.
Mendeley saves you time finding and organizing research
Choose a citation style from the tabs below