Word-level style modeling and control for
expressive non-attentive speech synthesis

Konstantinos Klapsas¹, Nikos Ellinas¹, June Sig Sung², Hyoungmin Park², Spyros Raptis^1
1Innoetics, Samsung Electronics, Greece
²Mobile Communications Business, Samsung Electronics, Republic of Korea

{n.ellinas, js6.sung, hm94.park, s.raptis}@samsung.com