Phonetic alignment: speech synthesis based vs. hybrid HMM/ANN.
Fabrice MalfrèreOlivier DerooThierry DutoitPublished in: ICSLP (1998)
Keyphrases
- speech recognition
- speech synthesis
- hidden markov models
- artificial neural networks
- speaker independent
- language model
- automatic speech recognition
- using artificial neural networks
- pattern recognition
- speech recognizer
- speech signal
- vocal tract
- noisy environments
- prosodic features
- profile hidden markov models
- speaker identification
- image alignment
- neural network
- speech corpus
- dynamic time warping
- recurrent neural networks
- speaker adaptation
- gesture recognition
- neural nets
- multilayer perceptron
- neural network model
- conditional random fields
- genetic algorithm ga
- considerable increase
- back propagation
- bayesian networks