JIKSI (Jurnal Ilmu Komputer dan Sistem Informasi)
Vol 3, No 1 (2015): Jurnal Ilmu Komputer dan Sistem Informasi

PENGENALAN SUARA UNTUK MENCARI JUDUL LAGU DENGAN BACKPROPAGATION NEURAL NETWORK

Ilham Samuel (Unknown)
Chairisni Lubis (Unknown)



Article Info

Publish Date
31 Jan 2015

Abstract

Voice recognition is a method which can be used to identify the title of a song. Song title recognizition is a difficult process because not every song our brain can remember and sometime only a fragment of song's lyrics or just a part of tone that can remember. This software is made as a solution to voice recognition to find the title of a song with backpropagation neural network method. The first method of this application is to find resemble extraction for every fragment of a song to get important resemble and then BPNN are going to look the similarity resemblance from every song's fragment. Recognition is executed by looking for similiraty of song's fragment, the most similar fragment will decided as the title of a song. The result of voice that had been tested are the voice which already trained the degree of recognition up to 100% an the voice which has not been trained have degree of recognition up to 42.1% Kata Kunci:Pengenalan Suara, Mel Frequency Cepstrum Coefficeints, Backpropagation Neural Network

Copyrights © 2015






Journal Info

Abbrev

jiksi

Publisher

Subject

Computer Science & IT Mathematics Other

Description

Jurnal Ilmu Komputer dan Sistem Informasi (JIKSI) diterbitkan oleh Fakultas Teknologi Informasi Universitas Tarumanagara (FTI Untar) Jakarta sebagai media publikasi karya ilmiah mahasiswa program studi Teknik Informatika dan Sistem Informasi FTI Untar. Karya-karya ilmiah yang dihasilkan berupa hasil ...