Effect of MFCC normalization on vector quantization based speaker identification

• Published in: The 10th IEEE International Symposium on Signal Processing and Information Technology pp. 250 - 253
• Main Authors: Shirali-Shahreza, M H, Shirali-Shahreza, Sajad
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.12.2010
• Subjects: Accuracy; Feature extraction; Mel frequency cepstral coefficient; Mel Frequency Cepstral Coefficients (MFCC); Normalization; Speaker Recognition; Speech; Speech recognition; Vector quantization; Vector Quantization (VQ)
• ISBN: 9781424499922; 1424499925
• ISSN: 2162-7843
• DOI: 10.1109/ISSPIT.2010.5711789

Mel Frequency Cepstral Coefficients (MFCC) are widely used in speech recognition and speaker identification. MFCC features are usually pre-processed before being used for recognition. One of these pre-processing is creating delta and delta-delta coefficients and append them to MFCC to create feature vector. Another pre-processing is coefficients mean normalization. In this paper, the effect of these two processes on the accuracy of a Vector Quantization (VQ) speaker identification system is compared. Additionally, it is shown that coefficient variance normalization, which is less common, can improve the accuracy.