PENGENALAN SUARA MENGGUNAKAN MEL FREQUENCY CEPSTRAL COEFFICIENTS DAN SELF ORGANIZING MAPS

Meutia Puspitasari, Julian Supardi, Yoppy Sazaki

Abstract


The interaction between human and computer can not be enjoyed by everyone, especially for people with disabilities, they will be difficult to interact using a computer. Speech to text recognition is one way to make it easy for anyone to operate a computer. Therefore, developed speech to text recognition research using Mel Frequency Cepstral Coefficients (MFCC) method to extract features from the voice data and Self Organizing Maps (SOM) to classify voice. This study uses primary data of 200 voice data as input for training and testing steps. The accuracy of the results achieved in this study is, testing for the same data with training data is 100 %, while for the data that is different from training data is 82 %. The results is obtained by comparing the amount of voice data that recognized successfully by the entire amount of voice data that tested.

Full Text:

A54-A61

Refbacks

  • There are currently no refbacks.



Konferensi Nasional Teknologi Informasi & Aplikasinya
Jurusan Sistem Informasi | Fakultas Ilmu Komputer
Universitas Sriwijaya