DNN-Based Full-Band Speech Synthesis Using GMM Approximation of Spectral Envelope.
Junya KoguchiShinnosuke TakamichiMasanori MoriseHiroshi SaruwatariShigeki SagayamaPublished in: IEICE Trans. Inf. Syst. (2020)
Keyphrases
- speech synthesis
- speech recognition
- gaussian mixture model
- text to speech
- spectral bands
- speaker identification
- vocal tract
- spectral resolution
- prosodic features
- narrow band
- multispectral
- speaker recognition
- multispectral images
- em algorithm
- mixture model
- hyperspectral imagery
- spectral analysis
- closed form
- hyperspectral images
- hyperspectral
- speech corpus
- neural network
- computer vision
- speech signal
- hidden markov models
- image sequences
- spectral features
- noisy environments
- approximation algorithms
- pattern recognition
- feature extraction
- gaussian mixture modeling