A new Japanese text-to-speech synthesizer based on COC synthesis method.
Kazuo HakodaShin'ya NakajimaTomohisa HirokawaHideyuki MizunoPublished in: ICSLP (1990)
Keyphrases
- text to speech
- preprocessing
- synthetic data
- computational cost
- segmentation method
- detection method
- dynamic programming
- significant improvement
- prior knowledge
- cost function
- experimental evaluation
- high accuracy
- multiresolution
- pairwise
- neural network
- support vector machine svm
- optimization algorithm
- classification method
- optimization method
- feature selection