Acoustic Markov models used in the Tangora speech recognition system | IEEE Conference Publication | IEEE Xplore