Bailing-TTS: Chinese Dialectal Speech Synthesis Towards Human-like Spontaneous Representation

TL;DR

Bailing-TTS is a large-scale Chinese dialectal speech synthesis model that uses semi-supervised learning and a specialized transformer architecture to generate high-quality, human-like spontaneous speech.

Contribution

It introduces a novel semi-supervised training framework and a transformer-based architecture specifically designed for Chinese dialectal speech synthesis.

Findings

Achieves high-quality dialectal speech generation

Demonstrates effective alignment of text and speech tokens

Produces human-like spontaneous speech results

Abstract

Large-scale text-to-speech (TTS) models have made significant progress recently.However, they still fall short in the generation of Chinese dialectal speech. Toaddress this, we propose Bailing-TTS, a family of large-scale TTS models capable of generating high-quality Chinese dialectal speech. Bailing-TTS serves as a foundation model for Chinese dialectal speech generation. First, continual semi-supervised learning is proposed to facilitate the alignment of text tokens and speech tokens. Second, the Chinese dialectal representation learning is developed using a specific transformer architecture and multi-stage training processes. With the proposed design of novel network architecture and corresponding strategy, Bailing-TTS is able to generate Chinese dialectal speech from text effectively and efficiently. Experiments demonstrate that Bailing-TTS generates Chinese dialectal speech towards human-like spontaneous representation. Readers are encouraged to listen to demos at \url{https://c9412600.github.io/bltts_tech_report/index.html}.