Short-time phase spectrum in speech processing: A review and some experimental results
Digital Signal Processing
Compensating Function of Formant Instantaneous Characteristics in Speaker Identification
IAS '09 Proceedings of the 2009 Fifth International Conference on Information Assurance and Security - Volume 01
Speaker Identification Using Instantaneous Frequencies
IEEE Transactions on Audio, Speech, and Language Processing
Hi-index | 0.00 |
The paper proposes the sub-band main peak frequencies( SMPF) for speaker identification (SI). The SMPF could be derived from the sub-band first formant frequencies by all-pole model of speech signal. Compared with MFCC features for SI based on a Gaussian mixture model (GMM), only SMPF features for SI is better than only the MFCC, with one of improved relative rate up to 15%. Experimental utterances are Chinese mandarin under clean background recording circumstances.