Computer-generated Speech based on Algorithmic Prosodic Modification


Purpose of the Study

In the spoken language, the same utterance may convey different meanings to a hearer. Such ambiguities can be resolved by emphasizing accents at different positions in a sentence. The main objective of the study is to determine whether the prosody, generated by speech synthesis with our focus to emphasize tone (FET) system, can convey the focus to the listeners.
 
 
(1) No Focus: There is no prosodic modification at any focus parts
I    want to  leave   Boston around     5       p.m.         
                 
Tone
 
(2) Speaker focuses at "who wants to leave Boston": Focus part "I" is emphasized by high accent tone
I                   want to   leave   Boston around    5       p.m.        
L+H* L-                
Tone
 
(3) Speaker focuses at "what the speaker wants to do": Focus part of the sentence is "want to leave" which is emphazied by high tone
I    want to leave          Boston around      5       p.m.       
  H*   L+H* L-          
Tone
 
(4) Speaker focuses at "where and when the speaker wants to leave": Focuses are "Boston" and "5 p.m."
I   want to leave Boston        around    5        p.m.       
        L+H* L-   H* L-L%  
Tone
 
© 2007 Lalita Narupiyakul, last update: 9-Feb-07