Transitional cues and redundancy in speech recognition
Brian L. Scott, Lloyd A. Smith, Lisan Lin, J. M. Newell · The Journal of the Acoustical Society of America · 1989
Perhaps the most difficult features for speech recognition systems to properly capture and align are the transitional cues that exist between phones. There are advantages and disadvantages associated with both linear time normalization methods and dynamic time warping (DTW). Linear methods preserve the relative durations of transitional and steady-state portions of the signal but tend to smear features and lose resolution. DTW retains resolution but loses durational relations. The present algorithm attempts to compensate for the smearing associated with linear normalization by copying transitional regions of the utterance to additional locations within the representation. This process involves locating the onset and offset of the syllable nucleus based on the normalized representation and copying two frames immediately surrounding the onset and offset to another location in the representation. The process enhances performance of the recognizer in two ways. First of all, the redundancy serves to weight more heavily the perceptually important transitional region that exists between consonantal cues, such as burst frequencies, frication, nasal murmurs, etc., and the vowel nucleus of the syllable. And second, copying these regions to another location within the representation serves to enhance the alignment of these cues during training and matching. The system is being evaluated over standard telephone lines using the alphabet as the test vocabulary. The telephone environment, and the inherent confusibility of the alphabet as a test vocabulary, places significant demands on the recognizer to accurately capture and identify transitional cues. The pilot study, using only a single speaker, provided encouraging results. The recognizer was 96.7% accurate over 8320 live trials. Currently, an alphabet database is being collected and speaker-independent test results will be presented.