Speaker and language adaptive training for HMM-based polyglot speech synthesis.
Heiga ZenPublished in: INTERSPEECH (2010)
Keyphrases
- speech synthesis
- speech recognition
- prosodic features
- text to speech
- hidden markov models
- vocal tract
- acoustic models
- automatic speech recognition
- programming language
- speech signal
- natural language
- training set
- language model
- training process
- supervised learning
- object oriented
- pattern recognition
- feature extraction
- discriminative training
- data mining
- speech corpus
- neural network
- speaker recognition
- noisy environments
- target language
- language processing
- language learning
- image coding
- test set
- information retrieval