Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

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

About

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.

Aoduo Li, Haoran Lv, Hongjian Xu, Shengmin Li, Sihao Qin, Zimeng Li, Chi Man Pun, Xuhang Chen• 2026

Related benchmarks

TaskDatasetResultRank
Cross-modal retrievalUnseen Characters (Strict OOD Split)
mAP75
10
Speech SynthesisAnimeTTS-Bench Zero-Shot Extended (test)
UTMOS4.28
5
Showing 2 of 2 rows

Other info

Follow for update