Downloads: 115
M.Tech M.E PhD Thesis | Electronics Communication Engineering | Volume 3 Issue 5, May 2014 | Pages: 1054 - 1058 | India
Musical Audio Beat Tracking using Hidden Markov Model
Abstract: 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.
Keywords: Beat-tracking, onset detection, k-nearest neighbor k-NN regression, Hidden Markov Model, Musical Audio Processing
How to Cite?: Lata Dhulekar, Dr. S. K. Shah, "Musical Audio Beat Tracking using Hidden Markov Model", Volume 3 Issue 5, May 2014, International Journal of Science and Research (IJSR), Pages: 1054-1058, https://www.ijsr.net/getabstract.php?paperid=20131975, DOI: https://dx.doi.org/10.21275/20131975