ISCA Archive Interspeech 2006
ISCA Archive Interspeech 2006

Training of coarticulation models using dominance functions and visual unit selection methods for audio-visual speech synthesis

Zdenek Krnoul, Milos Zelezný, Ludek Müller, Jakub Kanis

This paper presents results of training of coarticulation models for Czech audio-visual speech synthesis. Two approaches for solution of coarticulation in audio-visual speech synthesis were used, coarticulation based on dominance functions and visual unit selection. For both approaches, coarticulation models were trained. Models for unit selection approach were trained by visually clustered data. These data were obtained using decision tree algorithm. Outputs of audio-visual speech synthesis for both approaches were assessed and compared objectively.