Speech synthesis with various Emotional Expressions and
Speaking Styles by Style Interpolation and Morphing

Style Interpolation More Samples(require Javascript)

Interpolation(Neutral,Joyful)
Interpolation ratio(a1,a2)=
(1,0)
(0.75,0.25)
(0.5,0.5)
(0.25,0.75)
(0,1)
 
Neutral
Slightly JoyfulH
Joyful

 

Interpolation(Neutral,Sad)
Interpolation ratio(a1,a2)=
(1,0)
(0.75,0.25)
(0.5,0.5)
(0.25,0.75)
(0,1)
 
Neutral
Slightly SadH
Sad

 

Interpolation(Neutral,Rough)
Interpolation ratio(a1,a2)=
(1,0)
(0.75,0.25)
(0.5,0.5)
(0.25,0.75)
(0,1)
 
Neutral
Slightly RoughH
Rough

/heikiN bairitsu o sageta keiseki ga aru/
("There are evidences that someone lowered average magni cation." in English)


Style Morphing

Joyful
¨ ¨ ¨ ¨ ¨ ¨ ¨ ¨ ¨ ¨ ¨ ¨  Sad  

  Sad 
¨ ¨ ¨ ¨ ¨ ¨ ¨ ¨ ¨ ¨ ¨ ¨ Joyful

/mawari no utsukushii yukigeshiki o miru yoyuu mo deta/
("I recovered my composure for looking at a surrounding beautiful snowcovered landscape." in English)

Reference:
Makoto Tachibana, Junichi Yamagishi, Takashi Masuko, Takao Kobayashi
HMM-based speech synthesis with various speaking styles using model interpolation
Proc. 2nd International Conference on Speech Prosody, SP2004, pp.413-416, Nara, Japan (2004.03)