MixGAN-TTS: Efficient and Stable Speech Synthesis Based on Diffusion Model | IEEE Journals & Magazine | IEEE Xplore