Whether Mfcc or Gfcc Is Better for Recognizing Emotion from Speech? a Study

A major challenge for automatic speech recognition (ASR) relates to significant performance reduction in noisy environments. Recently, the study of the emotional content of speech signals got more importance and hence, many systems have been proposed to identify the emotional content of a spoken utterance. The important aspects of the design of a speech emotion recognition system are pre-processing, feature extraction, training and classification, recognition. Typically, extracted speaker features are short-time cepstral coefficients such as Melfrequency cepstral coefficients (MFCCs) and perceptual linear predictive (PLP) coefficients, or long-term features such as prosody. Such systems usually do not perform well under noisy conditions because extracted features are distorted by noise, causing mismatched likelihood calculation. By introducing a novel speaker feature, gammatone frequency cepstral coefficient (GFCC), based on an auditory periphery model, and show that this feature captures speaker characteristics and performs substantially better than conventional speaker features under noisy conditions. An important finding in the study is that GFCC features outperform conventional MFCC features under noisy conditions..