We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
SPECTROGRAM BASED MUSICAL INSTRUMENT IDENTIFICATION USING HIDDEN MARKOV MODEL (HMM) FOR MONOPHONIC AND POLYPHONIC MUSIC SIGNALS.
- Authors
Bhalke, D. G.; Rama Rao, C. B.; Bormane, D. S.; Vibhute, M.
- Abstract
Spectrogram is generated for musical notes, which is used to calculate the spectral, temporal and modulation features. To detect the musical instruments from polyphonic and monophonic musical notes , 23 features are analyzed . Out of 23 features 12 specific features are used to generate feature vector . Hidden Markov model (HMM) is used to calculate the conditional instrument existence probability. In this work, ten musical instruments from wind and string categories are used for identification. The musical instruments are recognized using different HMM algorithms: Forward, Backward, Posterior decoding and Viterbi algorithm and their results are compared . Recognition accuracy achieved for monophonic musical notes are 91% and 87% for polyphonic musical notes.
- Subjects
MUSICAL instruments; ALGORITHMS; SPECTROGRAMS; MONOPHONIC chansons; POLYPHONIC chansons; HIDDEN Markov models
- Publication
Acta Technica Napocensis. Electronica-Telecomunicatii, 2011, Vol 52, Issue 2, p1
- ISSN
1221-6542
- Publication type
Article