ATRIE: Adaptive Tuning for Robust Inference and Emotion in Persona-Driven Speech Synthesis

TL;DR

ATRIE is a novel speech synthesis framework that enhances persona consistency and emotional expressiveness using a dual-track architecture distilled from a large language model.

Contribution

It introduces a unified P2-DT architecture with disentangled static and dynamic features, enabling robust identity preservation and emotional richness in synthesized speech.

Findings

Zero-Shot Speaker Verification EER: 0.04

Achieves state-of-the-art in generation and retrieval on AnimeTTS-Bench

Establishes a new paradigm for persona-driven multimedia synthesis

Abstract

High-fidelity character voice synthesis is a cornerstone of immersive multimedia applications, particularly for interacting with anime avatars and digital humans. However, existing systems struggle to maintain consistent persona traits across diverse emotional contexts. To bridge this gap, we present ATRIE, a unified framework utilizing a Persona-Prosody Dual-Track (P2-DT) architecture. Our system disentangles generation into a static Timbre Track (via Scalar Quantization) and a dynamic Prosody Track (via Hierarchical Flow-Matching), distilled from a 14B LLM teacher. This design enables robust identity preservation (Zero-Shot Speaker Verification EER: 0.04) and rich emotional expression. Evaluated on our extended AnimeTTS-Bench (50 characters), ATRIE achieves state-of-the-art performance in both generation and cross-modal retrieval (mAP: 0.75), establishing a new paradigm for persona-driven multimedia content creation.