An investigation of modelling aspects for ratedependent speech recognition

For the modelling of speech rate variation in speech recognition many approaches have been suggested. However, the training of speech-rate dependent models has by far received most of the attention. In order to investigate problematic a spects related with the classification of the speech data whic h represents one of the major problems of these approaches extensive experiments were carried out on a German corpus of read speech. The results indicate that while the kind of the modeldriven speech-rate measure is only of minor importance a datadriven classification of the speech data significantly impro ves the performance of rate-dependent models. Further results suggest a detailed modelling of speech rate based on more general models. This means that it might be possible to model speech rate adaptation by means of a transformation based on a continuous measure.