Improving GAN-based vocoder for fast and high-quality speech synthesis.
Mengnan HeTingwei GuoZhenxing LuRuixiong ZhangCaixia GongPublished in: INTERSPEECH (2022)
Keyphrases
- speech synthesis
- high quality
- speech recognition
- text to speech
- prosodic features
- vocal tract
- speech corpus
- low quality
- ground truth
- image quality
- higher quality
- depth map
- word processing
- highly accurate
- structuring elements
- neural network
- multiscale
- similarity measure
- three dimensional
- feature selection
- computer vision
- artificial intelligence
- learning algorithm