As mentioned earlier, ASR problem can be attacked from two sides; namely
The Hidden Markov Model(HMM) is a result of the attempt to model the
speech generation statistically, and thus belongs to the first
category above. During the past several years it has become the most
successful speech model used in ASR. The main reason for this success
is it's wonderful ability to characterize the speech signal in a
mathematically tractable way.
In a typical HMM based ASR system,the HMM stage is proceeded by the preprocessing (parameter extraction) stages. Thus the input to the HMM is a discrete time sequence of parameter vectors, such as those described in the previous chapter. The parameter vectors can be supplied to the HMM, either in vector quantized form or in raw continuous form. It can be designed HMMs to handle any of the cases, but important point is how the HMM deals with the stochastic nature of the amplitudes of the feature vectors which is superimposed on the time stochasticity.