Research of Speech Signal Acoustic Models for Speaker Recognition
Authors: Gauti Oberoi
DOI: https://doi.org/10.17605/OSF.IO/36UMW
Short DOI: https://doi.org/ggkv84
Country: India
Full-text Research PDF File: View | Download
Abstract:
The research of speech signal acoustic models for speaker recognition been described in this paper. The aim of this research – to investigate acoustic speech signal models suitable for speaker recognition. In the analytical practical part, voice records were investigated, MFCC features were extracted, acoustic speech signal models were trained and tested.
Furthermore, investigation results have shown that components in records distributed differently. The six most common acoustic models components were chosen. The most common voice and background components are different. Statistical analysis has shown that log-likelihoods are not statistically significant different for different languages records when same type and same languages acoustic models were applied. Moreover, log-likelihoods are not statistically significant different for different languages records when English acoustic models were used. Finally, log-likelihoods differ mostly in Spanish and English language records. Increasing the number of English and Spanish records log-likelihoods are statistically significant different when English acoustic models are used.
Keywords: MFCC features, GMM, speaker recognition, acoustic model.
Paper Id: 477
Published On: 2018-03-10
Published In: Volume 6, Issue 2, March-April 2018
Cite This: Research of Speech Signal Acoustic Models for Speaker Recognition - Gauti Oberoi - IJIRMPS Volume 6, Issue 2, March-April 2018. DOI 10.17605/OSF.IO/36UMW