A method for estimating prosodic symbol from text for Japanese text-to-speech synthesis.
Ken-ichi MagataTomoki HamagamiMitsuo KomuraPublished in: ICSLP (1996)
Keyphrases
- text to speech synthesis
- significant improvement
- detection method
- support vector machine
- high precision
- synthetic data
- high accuracy
- experimental evaluation
- prior knowledge
- dynamic programming
- clustering method
- cost function
- computational cost
- probabilistic model
- similarity measure
- text categorization
- detection algorithm
- pairwise
- latent semantic analysis
- objective function