CM-TTS: Enhancing Real Time Text-to-Speech Synthesis Efficiency through Weighted Samplers and Consistency Models

arXiv:2404.00569v1 Announce Type: cross
Abstract: Neural Text-to-Speech (TTS) systems find broad applications in voice assistants, e-learning, and audiobook creation. The pursuit of modern models, like Diffusion Models (DMs), holds promise for achieving high-fidelity, real-time speech synthesis. Yet, the efficiency of multi-step sampling in Diffusion Models presents challenges. Efforts have been made to integrate GANs with DMs, speeding up inference by approximating denoising distributions, but this introduces issues with model convergence due to adversarial training. To overcome this, we introduce CM-TTS, …

abstract applications arxiv assistants audiobook challenges cs.cl cs.sd diffusion diffusion models eess.as efficiency e-learning fidelity modern real-time sampling speech synthesis systems text text-to-speech through tts type voice voice assistants