On the Importance of Components of the MFCC in speech and speaker recognition

摘要

In this paper, we analyzed the relative importance of components of MFCC for both speech recognition and speaker recognition using DTW recognizer in various noise environments. For English digit and under the Euclidean distance definition, the experiment results show cepstral components from C_2 to C_16 contain the most useful speaker information, while C_0 and C_1 are usually harm to speaker recognition. Cepstral terms from C_1 to C_12 are found to contain the most useful speech information. In both tasks, the additive noise decreases the relative importance of low MFCC terms faster than that of the middle and high MFCC terms, and the decrement depends on the speech SNR. The channel distortion will deteriorate low terms more than the middle and high MFCC terms in both tasks, also.