ResearchBib Share Your Research, Maximize Your Social Impacts
Sign for Notice Everyday Sign up >> Login

Musical Audio Beat Tracking using Hidden Markov Model

Journal: International Journal of Science and Research (IJSR) (Vol.3, No. 5)

Publication Date:

Authors : ; ;

Page : 1054-1058

Keywords : Beat-tracking; onset detection; k-nearest neighbor (k-NN) regression; Hidden Markov Model; Musical Audio Processing;

Source : Downloadexternal Find it from : Google Scholarexternal


The proposed method describes the beat tracking system of musical audio signals. Here, the beat tracking deals with the explicit modeling of non-beat states and estimates the time between consecutive beat events. The expected accuracy of the estimated beats is also provided additionally. To predict the accuracy of the beat estimate a k-nearest neighbor regression algorithm is used. A database of 222 musical signals of various genres is used to statistically evaluate the performance for beat tracking system. A new perspective for beat detection is presented which can be used to enhance the performance of the proposed algorithm and relates it with the human tapping.

Last modified: 2014-07-02 19:42:59