Non-Autoregressive Fully Parallel Deep Convolutional Neural Speech Synthesis.
Moa LeeJunmo LeeJoon-Hyuk ChangPublished in: IEEE ACM Trans. Audio Speech Lang. Process. (2022)
Keyphrases
- autoregressive
- speech synthesis
- network architecture
- speech recognition
- text to speech
- moving average
- deep learning
- non stationary
- gaussian markov random field
- random fields
- prosodic features
- neural network
- spectrum analysis
- vocal tract
- linear prediction
- sar images
- pattern recognition
- computer vision
- autoregressive model
- hidden markov models
- speech signal
- higher order
- random field models
- prior knowledge
- autoregressive moving average
- face recognition