Low Cost Duration Modelling for . . .

State transition matrices as used in standard HMM decoders have two widely perceived limitations. One is that the implicit Geometric state duration distributions which they model do not accurately reflect true duration distributions. The other is that they impose no hard limit on maximum duration with the result that state transition probabilities often have little influence when combined with acoustic probabilities, which are of a different order of magnitude. Explicit duration models were developed in the past to address the first problem. These were not widely taken up because their performance advantage in clean speech recognition was often not sufficiently great to offset the extra complexity which they introduced. However, duration models have much greater potential when applied to noisy speech recognition. In this paper we present a simple and generic form of explicit duration model and show that this leads to strong performance improvements when applied to connected digit recognition in noise. Keywords: duration models, noise robust ASR, HMMs Acknowledgements: This work was carried out within the EC/OFES RESPITE (REcognition of Speech by Partial Information TEchniques) project. Thanks to Jon Barker at Sheffield University for discussions on decoder optimality. Contents 1