Modeling coarticulation in continuous speech

Coarticulation is the articulatory influence of one phonetic element (e. g., vowel or consonant) on another. Our primary motivation to modeling coarticulation is to explain how a speech signal is influenced by phonetic context. Other straight forward applications of modeling coarticulation might include compact parametric speech synthesis techniques. Thus far, modeling coarticulation in speech has been largely limited to short sequences and/or limited phonetic context. This dissertation describes a coarticulation model for continuous speech.