Linear channel compensation in speech recognition typically involves estimating an additive shift in the cepstral domain. This paper explores both Bayesian and maximum likelihood techniques to transform either the features or the model parameters. Experiments on the Macrophone corpus show error rate reductions over cepstral mean subtraction for short utterances.

Cepstral normalization has been popularly used as a powerful approach to produce robust features for speech recognition. Good examples of approaches include the well known Cepstral Mean Subtraction (CMS) and Cepstral Mean and Variance Normalization (CMVN), in which either the first or both the first and the second moments of the Mel-frequency Cepstral Coefficients (MFCCs) are normalized [1, 2]....

In this paper our main aim to provide the difference between cepstral and non-cepstral feature extraction techniques. Here we try to cover-up most of the comparative features of Mel Frequency Cepstral Coefficient and prosodic features. In speaker recognition, there are two type of techniques are available for feature extraction: Short-term features i.e. Mel Frequency Cepstral Coefficient (MFCC)...

Evolutionary algorithms provide flexibility and robustness required to find satisfactory solutions in complex search spaces. This is why they are successfully applied for solving real engineering problems. In this work we propose an algorithm to evolve a robust speech representation, using a dynamic data selection method for reducing the computational cost of the fitness computation while impro...

The noise robustness property for an automatic speech recognition system is one of the most important factors to determine its recognition accuracy under a noise-corrupted environment. Among the various approaches, normalizing the statistical quantities of speech features is a very promising direction to create more noise-robust features. The related feature normalization approaches include cep...

In prior work we have demonstrated the noise robustness of a novel microphone solution, the PARAT earplug communication terminal. Here we extend that work with results for the ETSI Advanced Front-End and segmental cepstral mean and variance normalization (CMVN). We also propose a method for doing CMVN in the model domain. This removes the need to train models on normalized features, which may s...

We provide a supervised speech-independent voice recognition technique in this paper. In the feature extraction stage we propose a mel-cepstral based approach. Our feature vector classification method uses a special nonlinear metric, derived from the Hausdorff distance for sets, and a minimum mean distance classifier. Keywords—Text-independent speaker recognition, mel cepstral analysis, speech ...

The mel-cepstral coefficients are often calculated from the linear prediction coefficients by using recursion formulas. However, the obtained mel-cepstral coefficients have errors caused by truncation in the quefrency domain. The purpose of this report is to point out that the melcepstral coefficients can be calculated from the LP (Linear Prediction) coefficients using the recursion formulas wi...