Downloads: 111 | Views: 234
M.Tech / M.E / PhD Thesis | Electronics & Communication Engineering | India | Volume 3 Issue 5, May 2014 | Rating: 6.6 / 10
Musical Audio Beat Tracking using Hidden Markov Model
Lata Dhulekar | Dr. S. K. Shah
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
Edition: Volume 3 Issue 5, May 2014,
Pages: 1054 - 1058