processUtterance
public void processUtterance(Utterance utterance)
throws ProcessException
Convert phoneme timing and f0-time targets
into MBROLA format. The results are saved in the Segment relations,
as features "mbr_dur" (int-valued) and
"mbr_targets" (String-valued). MBROLA-converted targets
are saved in the Segment relation because in MBROLA input, targets
are represented as properties of the segments during which they occur.
- Specified by:
processUtterance in interface UtteranceProcessor
- Parameters:
utterance - the utterance to process
- Throws:
ProcessException - if an error occurs while
processing of the utterance