CSDL Bài trích Báo - Tạp chí
ADAPT-TTS: high-quality zero-shot multi-speaker text-to-speech adaptive-based for Vietnamese
Tác giả: Phuong Pham Ngoc, Chung Tran Quang, Mai Luong Chi
Số trang:
P. 159-173
Số phát hành:
Tập 39 - Số 2
Kiểu tài liệu:
Tạp chí trong nước
Nơi lưu trữ:
03 Quang Trung
Mã phân loại:
621
Ngôn ngữ:
English
Từ khóa:
Zero-shot TTS, multi-speaker, text-to-speech, diffusion models, mel-spectrogram denoiser, Extracting Mel-vector, EMV, Adapt-TTS
Chủ đề:
Engineering
Tóm tắt:
In this paper introduce the Adapt-TTS model that allows high-quality audio synthesis from a small adaptive sample without training to solve these problems. Key recommendations: 1. The Extracting Mel-vector (EMV) architecture allows for a better representation of speaker characteristics and speech style; 2. An improved zero-shot model with a denoising diffusion model (Mel-spectrogram denoiser) component allows for new voice synthesis without training with better quality (less noise).