In speech recognition for closed captioning, the process involves the training of an acoustic (stochastic) model called the hidden Markov model (HMM) using the acoustic features of several hundred ...