This work, first published in 1995, is primarily addressed to phonologists interested in speech and to speech engineers interested in phonology. Autosegmental theory is at the core of contemporary phonology and Markov models are the main tool of speech recognition.