F0 control characterization by perceptual impressions on speaking attitudes using multiple dimensional scaling analysis

Aiming at prosody control for speech synthesis expressing speaking attitudes, F0 shapes were characterized by their perceptual impressions. To directly correlate F0 shapes with perceptual impressions, single word utterances "n" extracted from daily conversations were employed. The analysis showed that speaking attitudes were manifested in the global F0 control of "n" as the differences of their average height (high-low) and dynamic patterns (rise, flat, fall and rise & fall). Next, controlled utterances of "n" were perceptually examined through multiple dimensional scaling analysis to confirm F0 control freedoms found in the analysis. The results showed the 3D structure of a perceptual impression space and factor dependent F0 control characteristics. The positive-negative attitude can be controlled by average F0 height while those of confident-doubtful or allowable-unacceptable are manifested through dynamic F0 patterns. These findings provide new possibilities of systematic F0 control for conversational speech synthesis with speaking attitudes using a corpus-based approach.