DiffGAN-TTS: High-Fidelity and Efficient Text-to-Speech with Denoising Diffusion GANs.
Songxiang LiuDan SuDong YuPublished in: CoRR (2022)
Keyphrases
- text to speech
- high fidelity
- denoising
- speech synthesis
- programming tool
- text to speech synthesis
- real time
- prosodic features
- medical image compression
- image denoising
- computer vision
- word processing
- multiresolution
- multiscale
- video conferencing
- total variation
- high quality
- diffusion processes
- human computer interaction
- natural images