The Community for Technology Leaders
2015 Eighth International Conference on Contemporary Computing (IC3) (2015)
Noida, India
Aug. 20, 2015 to Aug. 22, 2015
ISBN: 978-1-4673-7947-2
pp: 106-110
Harsh M. Vyas , Department of Computer Science and Engineering, National Institute of Technology Karnataka, Surathkal, India
Suma S. M. , Department of Computer Science and Engineering, National Institute of Technology Karnataka, Surathkal, India
Shashidhar G. Koolagudi , Department of Computer Science and Engineering, National Institute of Technology Karnataka, Surathkal, India
Guruprasad K. R. , Department Mechanical Engineering, National Institute of Technology Karnataka, Surathkal, 575025, India
ABSTRACT
In this work, an effort has been made to identify the gamakas present in a given piece of Carnatic music clip. Gamakas are the beautification elements used to improve the melody. The identification of gamaka is very important stage in note transcription. In the proposed method, features that correspond to melodic variations such as pitch and energy are used for characterizing the gamakas. The input pitch contour is modelled using Hidden Markov Model with 3 states, namely Attack, Sustain and Decay. These states correspond to ups and downs in the melody of the music. The system is validated using a comprehensive data set consisting 160 songs from 8 different ragas. The average accuracy of 75.86% is achieved using this method.
INDEX TERMS
Hidden Markov models, Feature extraction, Multiple signal classification, Instruments, Databases, Testing, Rendering (computer graphics)
CITATION

H. M. Vyas, Suma S. M., S. G. Koolagudi and Guruprasad K. R., "Identifying gamakas in Carnatic music," 2015 Eighth International Conference on Contemporary Computing (IC3), Noida, India, 2015, pp. 106-110.
doi:10.1109/IC3.2015.7346662
87 ms
(Ver 3.3 (11022016))