High-pitched excitation generation for glottal vocoding in statistical parametric speech synthesis using a deep neural network.
Lauri JuvelaBajibabu BollepalliManu AiraksinenPaavo AlkuPublished in: ICASSP (2016)
Keyphrases
- speech synthesis
- neural network
- speech recognition
- speech signal
- text to speech
- wide range
- vocal tract
- back propagation
- neural network model
- pattern recognition
- artificial neural networks
- prosodic features
- statistical analysis
- self organizing maps
- network model
- multi layer
- recurrent neural networks
- data driven
- fuzzy logic
- image reconstruction from projections
- neural network is trained
- prediction model
- backpropagation neural network
- learning vector quantization
- statistical information
- statistical tests
- bp neural network
- knn
- feature extraction