A generalised model for utilising prosodic information in continuous speech recognition | IEEE Conference Publication | IEEE Xplore