MIREX 2010 SUBMISSION: BEAT TRACKING WITH BIDIRECTIONAL LONG SHORT-TERM MEMORY NEURAL NETWORKS

We present a new universal beat tracker. It is based on auditory spectral features and relative spectral differences processed by a bidirectional Long Short-Term Memory recurrent neural network, which acts as reduction function for the tracking of beats. The network is trained with a large database of beat data covering various genres and music types.

[1]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[2]  George Tzanetakis,et al.  An experimental comparison of audio tempo induction algorithms , 2006, IEEE Transactions on Audio, Speech, and Language Processing.

[3]  Alex Graves,et al.  Supervised Sequence Labelling with Recurrent Neural Networks , 2012, Studies in Computational Intelligence.

[4]  Mark B. Sandler,et al.  A tutorial on onset detection in music signals , 2005, IEEE Transactions on Speech and Audio Processing.