1Capturing Local Variability for Speaker Normalization in Speech Recognition

This paper presents a new modeling framework which naturally extends the Hidden Markov Model (HMM) approach for capturing speaker variability. The new model reduces the impact of local spectral and temporal variability by means of a mismatch reduction, which is based on a finite set of spectral and temporal warping factors applied at the frame level. Optimum warping factors are obtained while decoding in a locally constrained search. The model involves augmenting the states of a standard HMM, providing a new degree of freedom. It is argued in the paper that this represents an efficient and effective method for compensating local variability in speech which may have potential application to a broader array of speech transformations. The technique is presented in the context of existing methods for frequency warping based speaker normalization for ASR. The novel model is evaluated in clean and noisy task domains using subsets of the Aurora 2, the Spanish Speech-Dat-Car and the TIDIGITS corpora. In addition, some experiments are performed on a Spanish language corpus of speakers with pathological and non-pathological speech. It has been found that, under clean or not severely degraded conditions, the new model provides improvements over the standard HMM baseline. It is argued that the framework of local warping is an effective general approach to providing more flexible models of speaker variability