"Eric A. Litman" <elitman at yahoo.com> writes: > Acoustic models do not work on senteces per se; rather, they characterize > speech sounds (phonemes) and how those sounds relate to other sounds or Gack. That should be "speech sounds (phones) and"