GELP: GAN-Excited Linear Prediction for Speech Synthesis from Mel-spectrogram.
Lauri JuvelaBajibabu BollepalliJunichi YamagishiPaavo AlkuPublished in: CoRR (2019)
Keyphrases
- linear prediction
- speech signal
- speech synthesis
- speech recognition
- cepstral coefficients
- vocal tract
- automatic speech recognition
- speaker identification
- hidden markov models
- structuring elements
- noisy environments
- spectral analysis
- prosodic features
- language model
- pattern recognition
- acoustic features
- information retrieval
- non stationary
- multiresolution
- image segmentation
- computer vision