• Mohammad Taufan Asri Zaen
  • Ahmad Susan Pardiyansyah
  • Budi Darmawan
Keywords: Speaker verification, HMM, MFCC.


In this research, we design and build a speaking verification system that use MFCC as voice extraction and Hidden Markov Models for pattern recognition. Data was collected from 10 speakers. We choose one of these speakers and from this one we build HMM model, and the other speakers used as testing data. The speaker which we chose the voice as the HMM model says “hadir” as many as 30 times as samples. Then we take 10 of these samples to be used as training process and the rest of the samples used as testing. So, we have 20 data samples for testing purpose which we choose to build the HMM model from, and we have 180 more voice samples from the other speakers. The result shows that the testing with smaller threshold raise the accuracy of recognized speaker. But, the testing result with voice samples from other then the voice of the speakers shows that, the bigger the threshold value we use, the better accuracy we get. So, from the whole chart image we can see that the more states we use effecting and makes the eshold range we can use to get accuracy result more then 90% for both testing process is narrow.


[1] B. Darmawan, R. Hidayat, A. Susanto, “Sistem Pengenalan Pembicara Menggunakan Hidden Markov Models dan MFCC,” in Proc. CITEE 2011, pp201-205, July 28, 2011.
[2] O. Toledo-Ronen, H. Aronowitz, R. Hoory, J. Pelecanos, and D.Nahamoo, “Towards Goat Detection in Text-Dependent Speaker Verification,” in International Conference on Speech Communication and Technology, 2011.
[3] M. F. BenZeghiba and H. Bourlard, “User-customized password speaker verification using multiple reference and background models,” Speech Communication, vol. 48, no. 9, pp. 1200–1213,2006.
[4] A. Charisma, “sistem verifikasi penutur menggunakan metoda mel frequency cepstral coefficients-vector quantisation (mfcc-vq) serta sum square error (sse) dan pengenalan kata menggunakan metoda logika fuzzy” Jurnal Teknik Eletro ITP, Volume 2 No. 2; Juli 2013
[5] I.W.A. Resmawan “Verifikasi Suara Menggunakan Metode MFCC Dan DTW” Skripsi Jurusan Teknik Elektro, Fakultas Teknik Universitas Udayana, 2010
[6] A. Buono, B. Kusumoputro, “Pengembangan Model HMM Berbasis Maksomum Lokal Menggunakan Jarak Euclid Untuk Sistem Identifikasi Pembicara,” Proceedings of National Conference On Computer Science&Information Technology. 2007, pp.49-54.
[7] L.R. Rabiner (1989), “A Tutorial on Hidden Markov Models and Selected applications in Speech Recognition,” Proc. IEEE. vol.77, 1989, pp.257-286.