BINUS UNIVERSITY

Analysis and Identification Gamelan Bonang Sound Spectrum

CIMSiM 2010 : Second International Conference on Computational Intelligence, Modelling and Simulation

Bali, Sept 28 – 30, 2010

Ford Lumban Gaol

Abstract

In this research will show a method for sound recognition with artificial neural network back propagation concept. The artificial neural network use sigmoid activation function to all layer. Steps to the extraction, first, devide into ten component, second do the Fast Fourier Transform, third continue with Power Spectral Density, fourth count the average. The end result show that the pattern will recognize by the Artificial Neural Network. Extraction performed, as well as Fourier transform, and also calculate the power density, and then process them with artificial neural networks, allows the system to do the identification of voice data Gamelan Bonang using In this research will show a method for sound-recognition with artificial neural network backpropagation concept. The artificial neural network use sigmoid activation function to all layer. Steps to the extraction, first, devide into ten component, second do the Fast Fourier Transform, third continue with Power Spectral Density, fourth count the average. The end result show that the pattern will recognize by the Artificial Neural Network. Extraction performed, as well as Fourier transform, and also calculate the power density, and then process them with artificial neural networks, allows the system to do the identification of voice data Gamelan Bonang.

Keywords

Identification, Component, Fourier Transform (FT), Power Spectral Density (PSD), Backpropagation