Robustness of speaker recognition from noisy speech samples and mismatched languages

Al-Noori, AHY; Li, FF; Duncan, PJ

Abstract

Speaker recognition systems can typically attain high performance in ideal conditions. However, significant degradations in accuracy are found in channel-mismatched scenarios. Non-stationary environmental noises and their variations are listed at the top of speaker recognition challenges. Gammtone frequency cepstral coefficient method (GFCC) has been developed to improve the robustness of speaker recognition. This paper presents systematic comparisons between performance of GFCC and conventional MFCC based speaker verification systems with a purposely collected noisy speech data set. Furthermore, the current work extends the experiments to include investigations into language independency features in recognition phases. The results show that GFCC has better verification performance in noisy environments than MFCC. However, GFCC show more sensitivity to language mismatch between enrolment and recognition phase.

Citation

Al-Noori, A., Li, F., & Duncan, P. (2016, June). Robustness of speaker recognition from noisy speech samples and mismatched languages. Presented at 140th Convention-AES, Paris

• Presentation Conference Type: Other
• Conference Name: 140th Convention-AES
• Conference Location: Paris
• Start Date: Jun 4, 2016
• End Date: Jun 7, 2016
• Acceptance Date: Jun 3, 2016
• Online Publication Date: May 26, 2016
• Publication Date: May 26, 2016
• Deposit Date: Apr 11, 2016
• Publisher URL: http://www.aes.org/e-lib/browse.cfm?elib=18275
• Related Public URLs: http://www.aes.org/events/140/
• Additional Information: Event Type : Conference; Funders : The Ministry of Higher education and Scientific research - Iraq