Applying SFC Model for Chinese Expressive Speech Synthesis