We present a study on acoustic modeling of Spanish phonetic units. Bootstrap with a set of English phonetic models, we first obtain context-independent unit models for Spanish. We then compare context-dependent modeling techniques based on the conventional maximum likelihood (ML) and the maximum a posteriori (MAP) criteria. We found the MAP-based context adaptation approach produces a better result than the ML approach when a large number of units need to be modeled but the amount of training data is limited.
Bibliographic reference. Alvarez-Cercadillo, J. / Lee, Chin-Hui / Hernandez-Gomez, Luis (1995): "Acoustic modeling of context dependent units, for large vocabulary speech recognition in Spanish", In EUROSPEECH-1995, 787-790.