| Paper: | SP-P6.4 | ||
| Session: | Feature Analysis for ASR, TTS, and Verification | ||
| Time: | Wednesday, May 19, 09:30 - 11:30 | ||
| Presentation: | Poster | ||
| Topic: | Speech Processing: Feature Extraction | ||
| Title: | APPLICATION OF THE MODIFIED GROUP DELAY FUNCTION TO SPEAKER IDENTIFICATION AND DISCRIMINATION | ||
| Authors: | Rajesh Hegde; Indian Institute of Technology | ||
| Hema Murthy; Indian Institute of Technology | |||
| V. Ramana Rao Gadde; Star Laboratory, SRI International | |||
| Abstract: | In this paper we explore new methods by which speakers can be identified and discriminated by using features derived from the Fourier transform phase. The Modified Group Delay Feature (MODGDF) which is a parameterized form of the modified group delay function is used as a front end feature in this study. A Gaussian mixture model (GMM) based speaker identification system is built with the MODGDF as the front end and is tested on bothclean(TIMIT) and noisy telephone(NTIMIT) speech.The results obtained arecompared with traditional Mel frequency cepstral coefficients (MFCC)which is derived from the fourier transform magnitude. When both MFCC and MODGDF were combined, the performance improved by about4% indicating that both phase and magnitude contain complementary information.It has been shown earlier that the MODGDF does posess phoneme specific characteristics. In this paper we show that the MODGDF has speaker specificproperties. We also make an attempt to understand speaker discriminating characteristics of the MODGDF through the non linear mapping techniquebased on sammon mapping and find that the MODGDF empirically demonstrates a certain level of linear separability among speakers in the lower dimensional speaker space. | ||
| Back | |||
Home -||-
Organizing Committee -||-
Technical Committee -||-
Technical Program -||-
Plenaries
Paper Submission -||-
Special Sessions -||-
ITT -||-
Paper Review -||-
Exhibits -||-
Tutorials
Information -||-
Registration -||-
Travel Insurance -||-
Housing -||-
Workshops